5 Changes between 0.9.6h and 0.9.7 [XX xxx 2002]
7 *) Change the SSL kerb5 codes to match RFC 2712.
10 *) Make -nameopt work fully for req and add -reqopt switch.
11 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
13 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
14 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
16 *) Make sure tests can be performed even if the corresponding algorithms
17 have been removed entirely. This was also the last step to make
18 OpenSSL compilable with DJGPP under all reasonable conditions.
19 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
21 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
22 to allow version independent disabling of normally unselected ciphers,
23 which may be activated as a side-effect of selecting a single cipher.
25 (E.g., cipher list string "RSA" enables ciphersuites that are left
26 out of "ALL" because they do not provide symmetric encryption.
27 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
28 [Lutz Jaenicke, Bodo Moeller]
30 *) Add appropriate support for separate platform-dependent build
31 directories. The recommended way to make a platform-dependent
32 build directory is the following (tested on Linux), maybe with
35 # Place yourself outside of the OpenSSL source tree. In
36 # this example, the environment variable OPENSSL_SOURCE
37 # is assumed to contain the absolute OpenSSL source directory.
38 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
39 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
40 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
42 ln -s $OPENSSL_SOURCE/$F $F
45 To be absolutely sure not to disturb the source tree, a "make clean"
46 is a good thing. If it isn't successfull, don't worry about it,
47 it probably means the source directory is very clean.
50 *) Make sure any ENGINE control commands make local copies of string
51 pointers passed to them whenever necessary. Otherwise it is possible
52 the caller may have overwritten (or deallocated) the original string
53 data when a later ENGINE operation tries to use the stored values.
54 [Götz Babin-Ebell <babinebell@trustcenter.de>]
56 *) Improve diagnostics in file reading and command-line digests.
57 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
59 *) Add AES modes CFB and OFB to the object database. Correct an
60 error in AES-CFB decryption.
63 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
64 allows existing EVP_CIPHER_CTX structures to be reused after
65 calling EVP_*Final(). This behaviour is used by encryption
66 BIOs and some applications. This has the side effect that
67 applications must explicitly clean up cipher contexts with
68 EVP_CIPHER_CTX_cleanup() or they will leak memory.
71 *) Check the values of dna and dnb in bn_mul_recursive before calling
72 bn_mul_comba (a non zero value means the a or b arrays do not contain
73 n2 elements) and fallback to bn_mul_normal if either is not zero.
76 *) Fix escaping of non-ASCII characters when using the -subj option
77 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
80 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
81 form for "surname", serialNumber has no short form.
82 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
83 therefore remove "mail" short name for "internet 7".
84 The OID for unique identifiers in X509 certificates is
85 x500UniqueIdentifier, not uniqueIdentifier.
86 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
89 *) Add an "init" command to the ENGINE config module and auto initialize
90 ENGINEs. Without any "init" command the ENGINE will be initialized
91 after all ctrl commands have been executed on it. If init=1 the
92 ENGINE is initailized at that point (ctrls before that point are run
93 on the uninitialized ENGINE and after on the initialized one). If
94 init=0 then the ENGINE will not be iniatialized at all.
97 *) Fix the 'app_verify_callback' interface so that the user-defined
98 argument is actually passed to the callback: In the
99 SSL_CTX_set_cert_verify_callback() prototype, the callback
100 declaration has been changed from
103 int (*cb)(X509_STORE_CTX *,void *);
104 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
105 i=s->ctx->app_verify_callback(&ctx)
106 has been changed into
107 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
109 To update applications using SSL_CTX_set_cert_verify_callback(),
110 a dummy argument can be added to their callback functions.
111 [D. K. Smetters <smetters@parc.xerox.com>]
113 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
114 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
116 *) Add and OPENSSL_LOAD_CONF define which will cause
117 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
118 This allows older applications to transparently support certain
119 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
120 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
121 load the config file and OPENSSL_add_all_algorithms_conf() which will
122 always load it have also been added.
125 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
126 Adjust NIDs and EVP layer.
127 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
129 *) Config modules support in openssl utility.
131 Most commands now load modules from the config file,
132 though in a few (such as version) this isn't done
133 because it couldn't be used for anything.
135 In the case of ca and req the config file used is
136 the same as the utility itself: that is the -config
137 command line option can be used to specify an
141 *) Move default behaviour from OPENSSL_config(). If appname is NULL
142 use "openssl_conf" if filename is NULL use default openssl config file.
145 *) Add an argument to OPENSSL_config() to allow the use of an alternative
146 config section name. Add a new flag to tolerate a missing config file
147 and move code to CONF_modules_load_file().
150 *) Support for crypto accelerator cards from Accelerated Encryption
151 Processing, www.aep.ie. (Use engine 'aep')
152 The support was copied from 0.9.6c [engine] and adapted/corrected
153 to work with the new engine framework.
154 [AEP Inc. and Richard Levitte]
156 *) Support for SureWare crypto accelerator cards from Baltimore
157 Technologies. (Use engine 'sureware')
158 The support was copied from 0.9.6c [engine] and adapted
159 to work with the new engine framework.
162 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
163 make the newer ENGINE framework commands for the CHIL engine work.
164 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
166 *) Make it possible to produce shared libraries on ReliantUNIX.
167 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
169 *) Add the configuration target debug-linux-ppro.
170 Make 'openssl rsa' use the general key loading routines
171 implemented in apps.c, and make those routines able to
172 handle the key format FORMAT_NETSCAPE and the variant
174 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
176 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
177 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
179 *) Add -keyform to rsautl, and document -engine.
180 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
182 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
183 BIO_R_NO_SUCH_FILE error code rather than the generic
184 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
189 ERR_peek_last_error_line
190 ERR_peek_last_error_line_data.
194 ERR_peek_error_line_data,
195 but report on the latest error recorded rather than the first one
196 still in the error queue.
197 [Ben Laurie, Bodo Moeller]
199 *) default_algorithms option in ENGINE config module. This allows things
201 default_algorithms = ALL
202 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
205 *) Prelminary ENGINE config module.
208 *) New experimental application configuration code.
211 *) Change the AES code to follow the same name structure as all other
212 symmetric ciphers, and behave the same way. Move everything to
213 the directory crypto/aes, thereby obsoleting crypto/rijndael.
214 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
216 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
217 [Ben Laurie and Theo de Raadt]
219 *) Add option to output public keys in req command.
220 [Massimiliano Pala madwolf@openca.org]
222 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
223 (up to about 10% better than before for P-192 and P-224).
226 *) New functions/macros
228 SSL_CTX_set_msg_callback(ctx, cb)
229 SSL_CTX_set_msg_callback_arg(ctx, arg)
230 SSL_set_msg_callback(ssl, cb)
231 SSL_set_msg_callback_arg(ssl, arg)
233 to request calling a callback function
235 void cb(int write_p, int version, int content_type,
236 const void *buf, size_t len, SSL *ssl, void *arg)
238 whenever a protocol message has been completely received
239 (write_p == 0) or sent (write_p == 1). Here 'version' is the
240 protocol version according to which the SSL library interprets
241 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
242 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
243 the content type as defined in the SSL 3.0/TLS 1.0 protocol
244 specification (change_cipher_spec(20), alert(21), handshake(22)).
245 'buf' and 'len' point to the actual message, 'ssl' to the
246 SSL object, and 'arg' is the application-defined value set by
247 SSL[_CTX]_set_msg_callback_arg().
249 'openssl s_client' and 'openssl s_server' have new '-msg' options
250 to enable a callback that displays all protocol messages.
253 *) Change the shared library support so shared libraries are built as
254 soon as the corresponding static library is finished, and thereby get
255 openssl and the test programs linked against the shared library.
256 This still only happens when the keyword "shard" has been given to
257 the configuration scripts.
259 NOTE: shared library support is still an experimental thing, and
260 backward binary compatibility is still not guaranteed.
261 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
263 *) Add support for Subject Information Access extension.
264 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
266 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
267 additional bytes when new memory had to be allocated, not just
268 when reusing an existing buffer.
271 *) New command line and configuration option 'utf8' for the req command.
272 This allows field values to be specified as UTF8 strings.
275 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
276 runs for the former and machine-readable output for the latter.
279 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
280 of the e-mail address in the DN (i.e., it will go into a certificate
281 extension only). The new configuration file option 'email_in_dn = no'
283 [Massimiliano Pala madwolf@openca.org]
285 *) Change all functions with names starting with des_ to be starting
286 with DES_ instead. Add wrappers that are compatible with libdes,
287 but are named _ossl_old_des_*. Finally, add macros that map the
288 des_* symbols to the corresponding _ossl_old_des_* if libdes
289 compatibility is desired. If OpenSSL 0.9.6c compatibility is
290 desired, the des_* symbols will be mapped to DES_*, with one
293 Since we provide two compatibility mappings, the user needs to
294 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
295 compatibility is desired. The default (i.e., when that macro
296 isn't defined) is OpenSSL 0.9.6c compatibility.
298 There are also macros that enable and disable the support of old
299 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
300 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
301 are defined, the default will apply: to support the old des routines.
303 In either case, one must include openssl/des.h to get the correct
304 definitions. Do not try to just include openssl/des_old.h, that
307 NOTE: This is a major break of an old API into a new one. Software
308 authors are encouraged to switch to the DES_ style functions. Some
309 time in the future, des_old.h and the libdes compatibility functions
310 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
311 default), and then completely removed.
314 *) Test for certificates which contain unsupported critical extensions.
315 If such a certificate is found during a verify operation it is
316 rejected by default: this behaviour can be overridden by either
317 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
318 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
319 X509_supported_extension() has also been added which returns 1 if a
320 particular extension is supported.
323 *) Modify the behaviour of EVP cipher functions in similar way to digests
324 to retain compatibility with existing code.
327 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
328 compatibility with existing code. In particular the 'ctx' parameter does
329 not have to be to be initialized before the call to EVP_DigestInit() and
330 it is tidied up after a call to EVP_DigestFinal(). New function
331 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
332 EVP_MD_CTX_copy() changed to not require the destination to be
333 initialized valid and new function EVP_MD_CTX_copy_ex() added which
334 requires the destination to be valid.
336 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
337 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
340 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
341 so that complete 'Handshake' protocol structures are kept in memory
342 instead of overwriting 'msg_type' and 'length' with 'body' data.
345 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
346 [Massimo Santin via Richard Levitte]
348 *) Major restructuring to the underlying ENGINE code. This includes
349 reduction of linker bloat, separation of pure "ENGINE" manipulation
350 (initialisation, etc) from functionality dealing with implementations
351 of specific crypto iterfaces. This change also introduces integrated
352 support for symmetric ciphers and digest implementations - so ENGINEs
353 can now accelerate these by providing EVP_CIPHER and EVP_MD
354 implementations of their own. This is detailed in crypto/engine/README
355 as it couldn't be adequately described here. However, there are a few
356 API changes worth noting - some RSA, DSA, DH, and RAND functions that
357 were changed in the original introduction of ENGINE code have now
358 reverted back - the hooking from this code to ENGINE is now a good
359 deal more passive and at run-time, operations deal directly with
360 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
361 dereferencing through an ENGINE pointer any more. Also, the ENGINE
362 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
363 they were not being used by the framework as there is no concept of a
364 BIGNUM_METHOD and they could not be generalised to the new
365 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
366 ENGINE_cpy() has been removed as it cannot be consistently defined in
370 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
373 *) Change mkdef.pl to sort symbols that get the same entry number,
374 and make sure the automatically generated functions ERR_load_*
375 become part of libeay.num as well.
378 *) New function SSL_renegotiate_pending(). This returns true once
379 renegotiation has been requested (either SSL_renegotiate() call
380 or HelloRequest/ClientHello receveived from the peer) and becomes
381 false once a handshake has been completed.
382 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
383 sends a HelloRequest, but does not ensure that a handshake takes
384 place. SSL_renegotiate_pending() is useful for checking if the
385 client has followed the request.)
388 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
389 By default, clients may request session resumption even during
390 renegotiation (if session ID contexts permit); with this option,
391 session resumption is possible only in the first handshake.
393 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
394 more bits available for options that should not be part of
395 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
398 *) Add some demos for certificate and certificate request creation.
401 *) Make maximum certificate chain size accepted from the peer application
402 settable (SSL*_get/set_max_cert_list()), as proposed by
403 "Douglas E. Engert" <deengert@anl.gov>.
406 *) Add support for shared libraries for Unixware-7
407 (Boyd Lynn Gerber <gerberb@zenez.com>).
410 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
411 be done prior to destruction. Use this to unload error strings from
412 ENGINEs that load their own error strings. NB: This adds two new API
413 functions to "get" and "set" this destroy handler in an ENGINE.
416 *) Alter all existing ENGINE implementations (except "openssl" and
417 "openbsd") to dynamically instantiate their own error strings. This
418 makes them more flexible to be built both as statically-linked ENGINEs
419 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
420 Also, add stub code to each that makes building them as self-contained
421 shared-libraries easier (see README.ENGINE).
424 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
425 implementations into applications that are completely implemented in
426 self-contained shared-libraries. The "dynamic" ENGINE exposes control
427 commands that can be used to configure what shared-library to load and
428 to control aspects of the way it is handled. Also, made an update to
429 the README.ENGINE file that brings its information up-to-date and
430 provides some information and instructions on the "dynamic" ENGINE
431 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
434 *) Make it possible to unload ranges of ERR strings with a new
435 "ERR_unload_strings" function.
438 *) Add a copy() function to EVP_MD.
441 *) Make EVP_MD routines take a context pointer instead of just the
442 md_data void pointer.
445 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
446 that the digest can only process a single chunk of data
447 (typically because it is provided by a piece of
448 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
449 is only going to provide a single chunk of data, and hence the
450 framework needn't accumulate the data for oneshot drivers.
453 *) As with "ERR", make it possible to replace the underlying "ex_data"
454 functions. This change also alters the storage and management of global
455 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
456 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
457 index counters. The API functions that use this state have been changed
458 to take a "class_index" rather than pointers to the class's local STACK
459 and counter, and there is now an API function to dynamically create new
460 classes. This centralisation allows us to (a) plug a lot of the
461 thread-safety problems that existed, and (b) makes it possible to clean
462 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
463 such data would previously have always leaked in application code and
464 workarounds were in place to make the memory debugging turn a blind eye
465 to it. Application code that doesn't use this new function will still
466 leak as before, but their memory debugging output will announce it now
467 rather than letting it slide.
469 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
470 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
471 has a return value to indicate success or failure.
474 *) Make it possible to replace the underlying "ERR" functions such that the
475 global state (2 LHASH tables and 2 locks) is only used by the "default"
476 implementation. This change also adds two functions to "get" and "set"
477 the implementation prior to it being automatically set the first time
478 any other ERR function takes place. Ie. an application can call "get",
479 pass the return value to a module it has just loaded, and that module
480 can call its own "set" function using that value. This means the
481 module's "ERR" operations will use (and modify) the error state in the
482 application and not in its own statically linked copy of OpenSSL code.
485 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
486 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
487 the operation, and provides a more encapsulated way for external code
488 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
489 to use these functions rather than manually incrementing the counts.
491 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
494 *) Add EVP test program.
497 *) Add symmetric cipher support to ENGINE. Expect the API to change!
500 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
501 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
502 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
503 These allow a CRL to be built without having to access X509_CRL fields
504 directly. Modify 'ca' application to use new functions.
507 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
508 bug workarounds. Rollback attack detection is a security feature.
509 The problem will only arise on OpenSSL servers when TLSv1 is not
510 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
511 Software authors not wanting to support TLSv1 will have special reasons
512 for their choice and can explicitly enable this option.
513 [Bodo Moeller, Lutz Jaenicke]
515 *) Rationalise EVP so it can be extended: don't include a union of
516 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
517 (similar to those existing for EVP_CIPHER_CTX).
522 EVP_MD_CTX_init(&md); /* new function call */
523 EVP_DigestInit(&md, EVP_sha1());
524 EVP_DigestUpdate(&md, in, len);
525 EVP_DigestFinal(&md, out, NULL);
526 EVP_MD_CTX_cleanup(&md); /* new function call */
530 *) Make DES key schedule conform to the usual scheme, as well as
531 correcting its structure. This means that calls to DES functions
532 now have to pass a pointer to a des_key_schedule instead of a
533 plain des_key_schedule (which was actually always a pointer
538 des_set_key_checked(..., &ks);
539 des_ncbc_encrypt(..., &ks, ...);
541 (Note that a later change renames 'des_...' into 'DES_...'.)
544 *) Initial reduction of linker bloat: the use of some functions, such as
545 PEM causes large amounts of unused functions to be linked in due to
546 poor organisation. For example pem_all.c contains every PEM function
547 which has a knock on effect of linking in large amounts of (unused)
548 ASN1 code. Grouping together similar functions and splitting unrelated
549 functions prevents this.
552 *) Cleanup of EVP macros.
555 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
559 *) Add initial OCSP responder support to ocsp application. The
560 revocation information is handled using the text based index
561 use by the ca application. The responder can either handle
562 requests generated internally, supplied in files (for example
563 via a CGI script) or using an internal minimal server.
566 *) Add configuration choices to get zlib compression for TLS.
569 *) Changes to Kerberos SSL for RFC 2712 compliance:
570 1. Implemented real KerberosWrapper, instead of just using
571 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
572 2. Implemented optional authenticator field of KerberosWrapper.
574 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
575 and authenticator structs; see crypto/krb5/.
577 Generalized Kerberos calls to support multiple Kerberos libraries.
578 [Vern Staats <staatsvr@asc.hpc.mil>,
579 Jeffrey Altman <jaltman@columbia.edu>
582 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
583 already does with RSA. testdsa.h now has 'priv_key/pub_key'
584 values for each of the key sizes rather than having just
585 parameters (and 'speed' generating keys each time).
588 *) Speed up EVP routines.
591 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
592 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
593 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
594 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
596 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
597 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
598 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
601 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
603 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
606 *) Added the OS2-EMX target.
607 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
609 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
610 to support NCONF routines in extension code. New function CONF_set_nconf()
611 to allow functions which take an NCONF to also handle the old LHASH
612 structure: this means that the old CONF compatible routines can be
613 retained (in particular wrt extensions) without having to duplicate the
614 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
617 *) Enhance the general user interface with mechanisms for inner control
618 and with possibilities to have yes/no kind of prompts.
621 *) Change all calls to low level digest routines in the library and
622 applications to use EVP. Add missing calls to HMAC_cleanup() and
623 don't assume HMAC_CTX can be copied using memcpy().
624 [Verdon Walker <VWalker@novell.com>, Steve Henson]
626 *) Add the possibility to control engines through control names but with
627 arbitrary arguments instead of just a string.
628 Change the key loaders to take a UI_METHOD instead of a callback
629 function pointer. NOTE: this breaks binary compatibility with earlier
630 versions of OpenSSL [engine].
631 Adapt the nCipher code for these new conditions and add a card insertion
635 *) Enhance the general user interface with mechanisms to better support
636 dialog box interfaces, application-defined prompts, the possibility
637 to use defaults (for example default passwords from somewhere else)
638 and interrupts/cancellations.
641 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
642 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
645 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
646 tidy up some unnecessarily weird code in 'sk_new()').
647 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
649 *) Change the key loading routines for ENGINEs to use the same kind
650 callback (pem_password_cb) as all other routines that need this
654 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
655 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
656 than this minimum value is recommended.
659 *) New random seeder for OpenVMS, using the system process statistics
660 that are easily reachable.
663 *) Windows apparently can't transparently handle global
664 variables defined in DLLs. Initialisations such as:
666 const ASN1_ITEM *it = &ASN1_INTEGER_it;
668 wont compile. This is used by the any applications that need to
669 declare their own ASN1 modules. This was fixed by adding the option
670 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
671 needed for static libraries under Win32.
674 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
675 setting of purpose and trust fields. New X509_STORE trust and
676 purpose functions and tidy up setting in other SSL functions.
679 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
680 structure. These are inherited by X509_STORE_CTX when it is
681 initialised. This allows various defaults to be set in the
682 X509_STORE structure (such as flags for CRL checking and custom
683 purpose or trust settings) for functions which only use X509_STORE_CTX
684 internally such as S/MIME.
686 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
687 trust settings if they are not set in X509_STORE. This allows X509_STORE
688 purposes and trust (in S/MIME for example) to override any set by default.
690 Add command line options for CRL checking to smime, s_client and s_server
694 *) Initial CRL based revocation checking. If the CRL checking flag(s)
695 are set then the CRL is looked up in the X509_STORE structure and
696 its validity and signature checked, then if the certificate is found
697 in the CRL the verify fails with a revoked error.
699 Various new CRL related callbacks added to X509_STORE_CTX structure.
701 Command line options added to 'verify' application to support this.
703 This needs some additional work, such as being able to handle multiple
704 CRLs with different times, extension based lookup (rather than just
705 by subject name) and ultimately more complete V2 CRL extension
709 *) Add a general user interface API (crypto/ui/). This is designed
710 to replace things like des_read_password and friends (backward
711 compatibility functions using this new API are provided).
712 The purpose is to remove prompting functions from the DES code
713 section as well as provide for prompting through dialog boxes in
714 a window system and the like.
717 *) Add "ex_data" support to ENGINE so implementations can add state at a
718 per-structure level rather than having to store it globally.
721 *) Make it possible for ENGINE structures to be copied when retrieved by
722 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
723 This causes the "original" ENGINE structure to act like a template,
724 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
725 operational state can be localised to each ENGINE structure, despite the
726 fact they all share the same "methods". New ENGINE structures returned in
727 this case have no functional references and the return value is the single
728 structural reference. This matches the single structural reference returned
729 by ENGINE_by_id() normally, when it is incremented on the pre-existing
733 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
734 needs to match any other type at all we need to manually clear the
738 *) Changes to the "openssl engine" utility to include;
739 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
740 about an ENGINE's available control commands.
741 - executing control commands from command line arguments using the
742 '-pre' and '-post' switches. '-post' is only used if '-t' is
743 specified and the ENGINE is successfully initialised. The syntax for
744 the individual commands are colon-separated, for example;
745 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
748 *) New dynamic control command support for ENGINEs. ENGINEs can now
749 declare their own commands (numbers), names (strings), descriptions,
750 and input types for run-time discovery by calling applications. A
751 subset of these commands are implicitly classed as "executable"
752 depending on their input type, and only these can be invoked through
753 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
754 can be based on user input, config files, etc). The distinction is
755 that "executable" commands cannot return anything other than a boolean
756 result and can only support numeric or string input, whereas some
757 discoverable commands may only be for direct use through
758 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
759 pointers, or other custom uses. The "executable" commands are to
760 support parameterisations of ENGINE behaviour that can be
761 unambiguously defined by ENGINEs and used consistently across any
762 OpenSSL-based application. Commands have been added to all the
763 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
764 control over shared-library paths without source code alterations.
767 *) Changed all ENGINE implementations to dynamically allocate their
768 ENGINEs rather than declaring them statically. Apart from this being
769 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
770 this also allows the implementations to compile without using the
771 internal engine_int.h header.
774 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
775 'const' value. Any code that should be able to modify a RAND_METHOD
776 should already have non-const pointers to it (ie. they should only
777 modify their own ones).
780 *) Made a variety of little tweaks to the ENGINE code.
781 - "atalla" and "ubsec" string definitions were moved from header files
782 to C code. "nuron" string definitions were placed in variables
783 rather than hard-coded - allowing parameterisation of these values
784 later on via ctrl() commands.
785 - Removed unused "#if 0"'d code.
786 - Fixed engine list iteration code so it uses ENGINE_free() to release
787 structural references.
788 - Constified the RAND_METHOD element of ENGINE structures.
789 - Constified various get/set functions as appropriate and added
790 missing functions (including a catch-all ENGINE_cpy that duplicates
791 all ENGINE values onto a new ENGINE except reference counts/state).
792 - Removed NULL parameter checks in get/set functions. Setting a method
793 or function to NULL is a way of cancelling out a previously set
794 value. Passing a NULL ENGINE parameter is just plain stupid anyway
795 and doesn't justify the extra error symbols and code.
796 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
797 flags from engine_int.h to engine.h.
798 - Changed prototypes for ENGINE handler functions (init(), finish(),
799 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
802 *) Implement binary inversion algorithm for BN_mod_inverse in addition
803 to the algorithm using long division. The binary algorithm can be
804 used only if the modulus is odd. On 32-bit systems, it is faster
805 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
806 roughly 5-15% for 256-bit moduli), so we use it only for moduli
807 up to 450 bits. In 64-bit environments, the binary algorithm
808 appears to be advantageous for much longer moduli; here we use it
809 for moduli up to 2048 bits.
812 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
813 could not support the combine flag in choice fields.
816 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
817 extensions from a certificate request to the certificate.
820 *) Allow multiple 'certopt' and 'nameopt' options to be separated
821 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
822 file: this allows the display of the certificate about to be
823 signed to be customised, to allow certain fields to be included
824 or excluded and extension details. The old system didn't display
825 multicharacter strings properly, omitted fields not in the policy
826 and couldn't display additional details such as extensions.
829 *) Function EC_POINTs_mul for multiple scalar multiplication
830 of an arbitrary number of elliptic curve points
831 \sum scalars[i]*points[i],
832 optionally including the generator defined for the EC_GROUP:
833 scalar*generator + \sum scalars[i]*points[i].
835 EC_POINT_mul is a simple wrapper function for the typical case
836 that the point list has just one item (besides the optional
840 *) First EC_METHODs for curves over GF(p):
842 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
843 operations and provides various method functions that can also
844 operate with faster implementations of modular arithmetic.
846 EC_GFp_mont_method() reuses most functions that are part of
847 EC_GFp_simple_method, but uses Montgomery arithmetic.
849 [Bodo Moeller; point addition and point doubling
850 implementation directly derived from source code provided by
851 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
853 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
856 Curves are EC_GROUP objects (with an optional group generator)
857 based on EC_METHODs that are built into the library.
859 Points are EC_POINT objects based on EC_GROUP objects.
861 Most of the framework would be able to handle curves over arbitrary
862 finite fields, but as there are no obvious types for fields other
863 than GF(p), some functions are limited to that for now.
866 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
867 that the file contains a complete HTTP response.
870 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
871 change the def and num file printf format specifier from "%-40sXXX"
872 to "%-39s XXX". The latter will always guarantee a space after the
873 field while the former will cause them to run together if the field
874 is 40 of more characters long.
877 *) Constify the cipher and digest 'method' functions and structures
878 and modify related functions to take constant EVP_MD and EVP_CIPHER
882 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
883 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
886 *) Modify EVP_Digest*() routines so they now return values. Although the
887 internal software routines can never fail additional hardware versions
891 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
893 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
894 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
897 ERR_R_NESTED_ASN1_ERROR
899 ERR_R_MISSING_ASN1_EOS
900 were 4 .. 9, conflicting with
901 ERR_LIB_RSA (= ERR_R_RSA_LIB)
903 ERR_LIB_PEM (= ERR_R_PEM_LIB).
904 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
906 Add new error code 'ERR_R_INTERNAL_ERROR'.
909 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
913 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
914 sets the subject name for a new request or supersedes the
915 subject name in a given request. Formats that can be parsed are
916 'CN=Some Name, OU=myOU, C=IT'
918 'CN=Some Name/OU=myOU/C=IT'.
920 Add options '-batch' and '-verbose' to 'openssl req'.
921 [Massimiliano Pala <madwolf@hackmasters.net>]
923 *) Introduce the possibility to access global variables through
924 functions on platform were that's the best way to handle exporting
925 global variables in shared libraries. To enable this functionality,
926 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
927 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
928 is normally done by Configure or something similar).
930 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
931 in the source file (foo.c) like this:
933 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
934 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
936 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
937 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
939 OPENSSL_DECLARE_GLOBAL(int,foo);
940 #define foo OPENSSL_GLOBAL_REF(foo)
941 OPENSSL_DECLARE_GLOBAL(double,bar);
942 #define bar OPENSSL_GLOBAL_REF(bar)
944 The #defines are very important, and therefore so is including the
945 header file everywhere where the defined globals are used.
947 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
948 of ASN.1 items, but that structure is a bit different.
950 The largest change is in util/mkdef.pl which has been enhanced with
951 better and easier to understand logic to choose which symbols should
952 go into the Windows .def files as well as a number of fixes and code
953 cleanup (among others, algorithm keywords are now sorted
954 lexicographically to avoid constant rewrites).
957 *) In BN_div() keep a copy of the sign of 'num' before writing the
958 result to 'rm' because if rm==num the value will be overwritten
959 and produce the wrong result if 'num' is negative: this caused
960 problems with BN_mod() and BN_nnmod().
963 *) Function OCSP_request_verify(). This checks the signature on an
964 OCSP request and verifies the signer certificate. The signer
965 certificate is just checked for a generic purpose and OCSP request
969 *) Add OCSP_check_validity() function to check the validity of OCSP
970 responses. OCSP responses are prepared in real time and may only
971 be a few seconds old. Simply checking that the current time lies
972 between thisUpdate and nextUpdate max reject otherwise valid responses
973 caused by either OCSP responder or client clock inaccuracy. Instead
974 we allow thisUpdate and nextUpdate to fall within a certain period of
975 the current time. The age of the response can also optionally be
976 checked. Two new options -validity_period and -status_age added to
980 *) If signature or public key algorithm is unrecognized print out its
981 OID rather that just UNKNOWN.
984 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
985 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
986 ID to be generated from the issuer certificate alone which can then be
987 passed to OCSP_id_issuer_cmp().
990 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
991 ASN1 modules to export functions returning ASN1_ITEM pointers
992 instead of the ASN1_ITEM structures themselves. This adds several
993 new macros which allow the underlying ASN1 function/structure to
994 be accessed transparently. As a result code should not use ASN1_ITEM
995 references directly (such as &X509_it) but instead use the relevant
996 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
997 use of the new ASN1 code on platforms where exporting structures
998 is problematical (for example in shared libraries) but exporting
999 functions returning pointers to structures is not.
1002 *) Add support for overriding the generation of SSL/TLS session IDs.
1003 These callbacks can be registered either in an SSL_CTX or per SSL.
1004 The purpose of this is to allow applications to control, if they wish,
1005 the arbitrary values chosen for use as session IDs, particularly as it
1006 can be useful for session caching in multiple-server environments. A
1007 command-line switch for testing this (and any client code that wishes
1008 to use such a feature) has been added to "s_server".
1009 [Geoff Thorpe, Lutz Jaenicke]
1011 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1012 of the form '#if defined(...) || defined(...) || ...' and
1013 '#if !defined(...) && !defined(...) && ...'. This also avoids
1014 the growing number of special cases it was previously handling.
1017 *) Make all configuration macros available for application by making
1018 sure they are available in opensslconf.h, by giving them names starting
1019 with "OPENSSL_" to avoid conflicts with other packages and by making
1020 sure e_os2.h will cover all platform-specific cases together with
1022 Additionally, it is now possible to define configuration/platform-
1023 specific names (called "system identities"). In the C code, these
1024 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1025 macro with the name beginning with "OPENSSL_SYS_", which is determined
1026 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1030 *) New option -set_serial to 'req' and 'x509' this allows the serial
1031 number to use to be specified on the command line. Previously self
1032 signed certificates were hard coded with serial number 0 and the
1033 CA options of 'x509' had to use a serial number in a file which was
1037 *) New options to 'ca' utility to support V2 CRL entry extensions.
1038 Currently CRL reason, invalidity date and hold instruction are
1039 supported. Add new CRL extensions to V3 code and some new objects.
1042 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1043 disable standard block padding (aka PKCS#5 padding) in the EVP
1044 API, which was previously mandatory. This means that the data is
1045 not padded in any way and so the total length much be a multiple
1046 of the block size, otherwise an error occurs.
1049 *) Initial (incomplete) OCSP SSL support.
1052 *) New function OCSP_parse_url(). This splits up a URL into its host,
1053 port and path components: primarily to parse OCSP URLs. New -url
1054 option to ocsp utility.
1057 *) New nonce behavior. The return value of OCSP_check_nonce() now
1058 reflects the various checks performed. Applications can decide
1059 whether to tolerate certain situations such as an absent nonce
1060 in a response when one was present in a request: the ocsp application
1061 just prints out a warning. New function OCSP_add1_basic_nonce()
1062 this is to allow responders to include a nonce in a response even if
1063 the request is nonce-less.
1066 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1067 skipped when using openssl x509 multiple times on a single input file,
1068 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1071 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1072 set string type: to handle setting ASN1_TIME structures. Fix ca
1073 utility to correctly initialize revocation date of CRLs.
1076 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1077 the clients preferred ciphersuites and rather use its own preferences.
1078 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1079 Internet Explorer by ensuring unchanged hash method during stepup.
1080 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1083 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1084 to aes and add a new 'exist' option to print out symbols that don't
1088 *) Additional options to ocsp utility to allow flags to be set and
1089 additional certificates supplied.
1092 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1093 OCSP client a number of certificate to only verify the response
1097 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1098 handle the new API. Currently only ECB, CBC modes supported. Add new
1101 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1102 Encryption Standard (AES) Ciphersuites for Transport Layer
1103 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1104 not enabled by default and were not part of the "ALL" ciphersuite
1105 alias because they were not yet official; they could be
1106 explicitly requested by specifying the "AESdraft" ciphersuite
1107 group alias. In the final release of OpenSSL 0.9.7, the group
1108 alias is called "AES" and is part of "ALL".)
1109 [Ben Laurie, Steve Henson, Bodo Moeller]
1111 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1112 request to response.
1115 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1116 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1117 extract information from a certificate request. OCSP_response_create()
1118 creates a response and optionally adds a basic response structure.
1119 OCSP_basic_add1_status() adds a complete single response to a basic
1120 response and returns the OCSP_SINGLERESP structure just added (to allow
1121 extensions to be included for example). OCSP_basic_add1_cert() adds a
1122 certificate to a basic response and OCSP_basic_sign() signs a basic
1123 response with various flags. New helper functions ASN1_TIME_check()
1124 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1125 (converts ASN1_TIME to GeneralizedTime).
1128 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1129 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1130 structure from a certificate. X509_pubkey_digest() digests the public_key
1131 contents: this is used in various key identifiers.
1134 *) Make sk_sort() tolerate a NULL argument.
1135 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1137 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1138 passed by the function are trusted implicitly. If any of them signed the
1139 response then it is assumed to be valid and is not verified.
1142 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1143 to data. This was previously part of the PKCS7 ASN1 code. This
1144 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1145 [Steve Henson, reported by Kenneth R. Robinette
1146 <support@securenetterm.com>]
1148 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1149 routines: without these tracing memory leaks is very painful.
1150 Fix leaks in PKCS12 and PKCS7 routines.
1153 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1154 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1155 effectively meant GeneralizedTime would never be used. Now it
1156 is initialised to -1 but X509_time_adj() now has to check the value
1157 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1158 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1159 [Steve Henson, reported by Kenneth R. Robinette
1160 <support@securenetterm.com>]
1162 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1163 result in a zero length in the ASN1_INTEGER structure which was
1164 not consistent with the structure when d2i_ASN1_INTEGER() was used
1165 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1166 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1167 where it did not print out a minus for negative ASN1_INTEGER.
1170 *) Add summary printout to ocsp utility. The various functions which
1171 convert status values to strings have been renamed to:
1172 OCSP_response_status_str(), OCSP_cert_status_str() and
1173 OCSP_crl_reason_str() and are no longer static. New options
1174 to verify nonce values and to disable verification. OCSP response
1175 printout format cleaned up.
1178 *) Add additional OCSP certificate checks. These are those specified
1179 in RFC2560. This consists of two separate checks: the CA of the
1180 certificate being checked must either be the OCSP signer certificate
1181 or the issuer of the OCSP signer certificate. In the latter case the
1182 OCSP signer certificate must contain the OCSP signing extended key
1183 usage. This check is performed by attempting to match the OCSP
1184 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1185 in the OCSP_CERTID structures of the response.
1188 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1189 and related routines. This uses the standard OpenSSL certificate
1190 verify routines to perform initial checks (just CA validity) and
1191 to obtain the certificate chain. Then additional checks will be
1192 performed on the chain. Currently the root CA is checked to see
1193 if it is explicitly trusted for OCSP signing. This is used to set
1194 a root CA as a global signing root: that is any certificate that
1195 chains to that CA is an acceptable OCSP signing certificate.
1198 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1199 extensions from a separate configuration file.
1200 As when reading extensions from the main configuration file,
1201 the '-extensions ...' option may be used for specifying the
1203 [Massimiliano Pala <madwolf@comune.modena.it>]
1205 *) New OCSP utility. Allows OCSP requests to be generated or
1206 read. The request can be sent to a responder and the output
1207 parsed, outputed or printed in text form. Not complete yet:
1208 still needs to check the OCSP response validity.
1211 *) New subcommands for 'openssl ca':
1212 'openssl ca -status <serial>' prints the status of the cert with
1213 the given serial number (according to the index file).
1214 'openssl ca -updatedb' updates the expiry status of certificates
1216 [Massimiliano Pala <madwolf@comune.modena.it>]
1218 *) New '-newreq-nodes' command option to CA.pl. This is like
1219 '-newreq', but calls 'openssl req' with the '-nodes' option
1220 so that the resulting key is not encrypted.
1221 [Damien Miller <djm@mindrot.org>]
1223 *) New configuration for the GNU Hurd.
1224 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1226 *) Initial code to implement OCSP basic response verify. This
1227 is currently incomplete. Currently just finds the signer's
1228 certificate and verifies the signature on the response.
1231 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1232 value of OPENSSLDIR. This is available via the new '-d' option
1233 to 'openssl version', and is also included in 'openssl version -a'.
1236 *) Allowing defining memory allocation callbacks that will be given
1237 file name and line number information in additional arguments
1238 (a const char* and an int). The basic functionality remains, as
1239 well as the original possibility to just replace malloc(),
1240 realloc() and free() by functions that do not know about these
1241 additional arguments. To register and find out the current
1242 settings for extended allocation functions, the following
1243 functions are provided:
1245 CRYPTO_set_mem_ex_functions
1246 CRYPTO_set_locked_mem_ex_functions
1247 CRYPTO_get_mem_ex_functions
1248 CRYPTO_get_locked_mem_ex_functions
1250 These work the same way as CRYPTO_set_mem_functions and friends.
1251 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1252 extended allocation function is enabled.
1253 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1254 a conventional allocation function is enabled.
1255 [Richard Levitte, Bodo Moeller]
1257 *) Finish off removing the remaining LHASH function pointer casts.
1258 There should no longer be any prototype-casting required when using
1259 the LHASH abstraction, and any casts that remain are "bugs". See
1260 the callback types and macros at the head of lhash.h for details
1261 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1264 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1265 If /dev/[u]random devices are not available or do not return enough
1266 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1268 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1269 /etc/entropy will be queried once each in this sequence, quering stops
1270 when enough entropy was collected without querying more sockets.
1273 *) Change the Unix RAND_poll() variant to be able to poll several
1274 random devices, as specified by DEVRANDOM, until a sufficient amount
1275 of data has been collected. We spend at most 10 ms on each file
1276 (select timeout) and read in non-blocking mode. DEVRANDOM now
1277 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1278 (previously it was just the string "/dev/urandom"), so on typical
1279 platforms the 10 ms delay will never occur.
1280 Also separate out the Unix variant to its own file, rand_unix.c.
1281 For VMS, there's a currently-empty rand_vms.c.
1284 *) Move OCSP client related routines to ocsp_cl.c. These
1285 provide utility functions which an application needing
1286 to issue a request to an OCSP responder and analyse the
1287 response will typically need: as opposed to those which an
1288 OCSP responder itself would need which will be added later.
1290 OCSP_request_sign() signs an OCSP request with an API similar
1291 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1292 response. OCSP_response_get1_basic() extracts basic response
1293 from response. OCSP_resp_find_status(): finds and extracts status
1294 information from an OCSP_CERTID structure (which will be created
1295 when the request structure is built). These are built from lower
1296 level functions which work on OCSP_SINGLERESP structures but
1297 wont normally be used unless the application wishes to examine
1298 extensions in the OCSP response for example.
1300 Replace nonce routines with a pair of functions.
1301 OCSP_request_add1_nonce() adds a nonce value and optionally
1302 generates a random value. OCSP_check_nonce() checks the
1303 validity of the nonce in an OCSP response.
1306 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1307 This doesn't copy the supplied OCSP_CERTID and avoids the
1308 need to free up the newly created id. Change return type
1309 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1310 This can then be used to add extensions to the request.
1311 Deleted OCSP_request_new(), since most of its functionality
1312 is now in OCSP_REQUEST_new() (and the case insensitive name
1313 clash) apart from the ability to set the request name which
1314 will be added elsewhere.
1317 *) Update OCSP API. Remove obsolete extensions argument from
1318 various functions. Extensions are now handled using the new
1319 OCSP extension code. New simple OCSP HTTP function which
1320 can be used to send requests and parse the response.
1323 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1324 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1325 uses the special reorder version of SET OF to sort the attributes
1326 and reorder them to match the encoded order. This resolves a long
1327 standing problem: a verify on a PKCS7 structure just after signing
1328 it used to fail because the attribute order did not match the
1329 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1330 it uses the received order. This is necessary to tolerate some broken
1331 software that does not order SET OF. This is handled by encoding
1332 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1333 to produce the required SET OF.
1336 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1337 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1338 files to get correct declarations of the ASN.1 item variables.
1341 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1342 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1343 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1344 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1345 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1346 ASN1_ITEM and no wrapper functions.
1349 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1350 replace the old function pointer based I/O routines. Change most of
1351 the *_d2i_bio() and *_d2i_fp() functions to use these.
1354 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1355 lines, recognice more "algorithms" that can be deselected, and make
1356 it complain about algorithm deselection that isn't recognised.
1359 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1360 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1361 to use new functions. Add NO_ASN1_OLD which can be set to remove
1362 some old style ASN1 functions: this can be used to determine if old
1363 code will still work when these eventually go away.
1366 *) New extension functions for OCSP structures, these follow the
1367 same conventions as certificates and CRLs.
1370 *) New function X509V3_add1_i2d(). This automatically encodes and
1371 adds an extension. Its behaviour can be customised with various
1372 flags to append, replace or delete. Various wrappers added for
1373 certifcates and CRLs.
1376 *) Fix to avoid calling the underlying ASN1 print routine when
1377 an extension cannot be parsed. Correct a typo in the
1378 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1381 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1382 entries for variables.
1385 *) Add functionality to apps/openssl.c for detecting locking
1386 problems: As the program is single-threaded, all we have
1387 to do is register a locking callback using an array for
1388 storing which locks are currently held by the program.
1391 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1392 SSL_get_ex_data_X509_STORE_idx(), which is used in
1393 ssl_verify_cert_chain() and thus can be called at any time
1394 during TLS/SSL handshakes so that thread-safety is essential.
1395 Unfortunately, the ex_data design is not at all suited
1396 for multi-threaded use, so it probably should be abolished.
1399 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1400 [Broadcom, tweaked and integrated by Geoff Thorpe]
1402 *) Move common extension printing code to new function
1403 X509V3_print_extensions(). Reorganise OCSP print routines and
1404 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1407 *) New function X509_signature_print() to remove duplication in some
1411 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1412 set (this was treated exactly the same as SET OF previously). This
1413 is used to reorder the STACK representing the structure to match the
1414 encoding. This will be used to get round a problem where a PKCS7
1415 structure which was signed could not be verified because the STACK
1416 order did not reflect the encoded order.
1419 *) Reimplement the OCSP ASN1 module using the new code.
1422 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1423 for its ASN1 operations. The old style function pointers still exist
1424 for now but they will eventually go away.
1427 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1428 completely replaces the old ASN1 functionality with a table driven
1429 encoder and decoder which interprets an ASN1_ITEM structure describing
1430 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1431 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1432 has also been converted to the new form.
1435 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1436 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1437 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1438 for negative moduli.
1441 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1442 of not touching the result's sign bit.
1445 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1449 *) Changed the LHASH code to use prototypes for callbacks, and created
1450 macros to declare and implement thin (optionally static) functions
1451 that provide type-safety and avoid function pointer casting for the
1452 type-specific callbacks.
1455 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1457 [Veers Staats <staatsvr@asc.hpc.mil>,
1458 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1460 *) Reformat the FAQ so the different questions and answers can be divided
1461 in sections depending on the subject.
1464 *) Have the zlib compression code load ZLIB.DLL dynamically under
1468 *) New function BN_mod_sqrt for computing square roots modulo a prime
1469 (using the probabilistic Tonelli-Shanks algorithm unless
1470 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1471 be handled deterministically).
1472 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1474 *) Make BN_mod_inverse faster by explicitly handling small quotients
1475 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1476 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1479 *) New function BN_kronecker.
1482 *) Fix BN_gcd so that it works on negative inputs; the result is
1483 positive unless both parameters are zero.
1484 Previously something reasonably close to an infinite loop was
1485 possible because numbers could be growing instead of shrinking
1486 in the implementation of Euclid's algorithm.
1489 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1490 sign of the number in question.
1492 Fix BN_is_word(a,w) to work correctly for w == 0.
1494 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1495 because its test if the absolute value of 'a' equals 'w'.
1496 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1497 it exists mostly for use in the implementations of BN_is_zero(),
1498 BN_is_one(), and BN_is_word().
1501 *) New function BN_swap.
1504 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1505 the exponentiation functions are more likely to produce reasonable
1506 results on negative inputs.
1509 *) Change BN_mod_mul so that the result is always non-negative.
1510 Previously, it could be negative if one of the factors was negative;
1511 I don't think anyone really wanted that behaviour.
1514 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1515 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1516 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1517 and add new functions:
1526 BN_mod_lshift1_quick
1530 These functions always generate non-negative results.
1532 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1533 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1535 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1536 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1537 be reduced modulo m.
1538 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1540 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1541 was actually never needed) and in BN_mul(). The removal in BN_mul()
1542 required a small change in bn_mul_part_recursive() and the addition
1543 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1544 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1545 bn_sub_words() and bn_add_words() except they take arrays with
1549 *) In 'openssl passwd', verify passwords read from the terminal
1550 unless the '-salt' option is used (which usually means that
1551 verification would just waste user's time since the resulting
1552 hash is going to be compared with some given password hash)
1553 or the new '-noverify' option is used.
1555 This is an incompatible change, but it does not affect
1556 non-interactive use of 'openssl passwd' (passwords on the command
1557 line, '-stdin' option, '-in ...' option) and thus should not
1561 *) Remove all references to RSAref, since there's no more need for it.
1564 *) Make DSO load along a path given through an environment variable
1565 (SHLIB_PATH) with shl_load().
1568 *) Constify the ENGINE code as a result of BIGNUM constification.
1569 Also constify the RSA code and most things related to it. In a
1570 few places, most notable in the depth of the ASN.1 code, ugly
1571 casts back to non-const were required (to be solved at a later
1575 *) Make it so the openssl application has all engines loaded by default.
1578 *) Constify the BIGNUM routines a little more.
1581 *) Add the following functions:
1583 ENGINE_load_cswift()
1585 ENGINE_load_atalla()
1587 ENGINE_load_builtin_engines()
1589 That way, an application can itself choose if external engines that
1590 are built-in in OpenSSL shall ever be used or not. The benefit is
1591 that applications won't have to be linked with libdl or other dso
1592 libraries unless it's really needed.
1594 Changed 'openssl engine' to load all engines on demand.
1595 Changed the engine header files to avoid the duplication of some
1596 declarations (they differed!).
1599 *) 'openssl engine' can now list capabilities.
1602 *) Better error reporting in 'openssl engine'.
1605 *) Never call load_dh_param(NULL) in s_server.
1608 *) Add engine application. It can currently list engines by name and
1609 identity, and test if they are actually available.
1612 *) Improve RPM specification file by forcing symbolic linking and making
1613 sure the installed documentation is also owned by root.root.
1614 [Damien Miller <djm@mindrot.org>]
1616 *) Give the OpenSSL applications more possibilities to make use of
1617 keys (public as well as private) handled by engines.
1620 *) Add OCSP code that comes from CertCo.
1623 *) Add VMS support for the Rijndael code.
1626 *) Added untested support for Nuron crypto accelerator.
1629 *) Add support for external cryptographic devices. This code was
1630 previously distributed separately as the "engine" branch.
1631 [Geoff Thorpe, Richard Levitte]
1633 *) Rework the filename-translation in the DSO code. It is now possible to
1634 have far greater control over how a "name" is turned into a filename
1635 depending on the operating environment and any oddities about the
1636 different shared library filenames on each system.
1639 *) Support threads on FreeBSD-elf in Configure.
1642 *) Fix for SHA1 assembly problem with MASM: it produces
1643 warnings about corrupt line number information when assembling
1644 with debugging information. This is caused by the overlapping
1646 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1649 NCONF_get_number() has no error checking at all. As a replacement,
1650 NCONF_get_number_e() is defined (_e for "error checking") and is
1651 promoted strongly. The old NCONF_get_number is kept around for
1652 binary backward compatibility.
1653 Make it possible for methods to load from something other than a BIO,
1654 by providing a function pointer that is given a name instead of a BIO.
1655 For example, this could be used to load configuration data from an
1659 *) Fix for non blocking accept BIOs. Added new I/O special reason
1660 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1661 with non blocking I/O was not possible because no retry code was
1662 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1666 *) Added the beginnings of Rijndael support.
1669 *) Fix for bug in DirectoryString mask setting. Add support for
1670 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1671 to allow certificate printing to more controllable, additional
1672 'certopt' option to 'x509' to allow new printing options to be
1676 *) Clean old EAY MD5 hack from e_os.h.
1679 Changes between 0.9.6g and 0.9.6h [xx XXX xxxx]
1681 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
1682 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
1685 *) Fix initialization code race conditions in
1686 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
1687 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
1688 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
1689 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
1690 ssl2_get_cipher_by_char(),
1691 ssl3_get_cipher_by_char().
1692 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
1694 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
1695 the cached sessions are flushed, as the remove_cb() might use ex_data
1696 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
1697 (see [openssl.org #212]).
1698 [Geoff Thorpe, Lutz Jaenicke]
1700 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
1701 length, instead of the encoding length to d2i_ASN1_OBJECT.
1704 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
1706 *) [In 0.9.6g-engine release:]
1707 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
1708 [Lynn Gazis <lgazis@rainbow.com>]
1710 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
1712 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
1713 and get fix the header length calculation.
1714 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
1715 Alon Kantor <alonk@checkpoint.com> (and others),
1718 *) Use proper error handling instead of 'assertions' in buffer
1719 overflow checks added in 0.9.6e. This prevents DoS (the
1720 assertions could call abort()).
1721 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
1723 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
1725 *) Add various sanity checks to asn1_get_length() to reject
1726 the ASN1 length bytes if they exceed sizeof(long), will appear
1727 negative or the content length exceeds the length of the
1729 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
1731 *) Fix cipher selection routines: ciphers without encryption had no flags
1732 for the cipher strength set and where therefore not handled correctly
1733 by the selection routines (PR #130).
1736 *) Fix EVP_dsa_sha macro.
1740 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
1741 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
1742 that was added in OpenSSL 0.9.6d.
1744 As the countermeasure turned out to be incompatible with some
1745 broken SSL implementations, the new option is part of SSL_OP_ALL.
1746 SSL_OP_ALL is usually employed when compatibility with weird SSL
1747 implementations is desired (e.g. '-bugs' option to 's_client' and
1748 's_server'), so the new option is automatically set in many
1752 *) Changes in security patch:
1754 Changes marked "(CHATS)" were sponsored by the Defense Advanced
1755 Research Projects Agency (DARPA) and Air Force Research Laboratory,
1756 Air Force Materiel Command, USAF, under agreement number
1759 *) Add various sanity checks to asn1_get_length() to reject
1760 the ASN1 length bytes if they exceed sizeof(long), will appear
1761 negative or the content length exceeds the length of the
1762 supplied buffer. (CAN-2002-0659)
1763 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
1765 *) Assertions for various potential buffer overflows, not known to
1767 [Ben Laurie (CHATS)]
1769 *) Various temporary buffers to hold ASCII versions of integers were
1770 too small for 64 bit platforms. (CAN-2002-0655)
1771 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
1773 *) Remote buffer overflow in SSL3 protocol - an attacker could
1774 supply an oversized master key in Kerberos-enabled versions.
1776 [Ben Laurie (CHATS)]
1778 *) Remote buffer overflow in SSL3 protocol - an attacker could
1779 supply an oversized session ID to a client. (CAN-2002-0656)
1780 [Ben Laurie (CHATS)]
1782 *) Remote buffer overflow in SSL2 protocol - an attacker could
1783 supply an oversized client master key. (CAN-2002-0656)
1784 [Ben Laurie (CHATS)]
1787 Changes between 0.9.6c and 0.9.6d [9 May 2002]
1789 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
1790 encoded as NULL) with id-dsa-with-sha1.
1791 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
1793 *) Check various X509_...() return values in apps/req.c.
1794 [Nils Larsch <nla@trustcenter.de>]
1796 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
1797 an end-of-file condition would erronously be flagged, when the CRLF
1798 was just at the end of a processed block. The bug was discovered when
1799 processing data through a buffering memory BIO handing the data to a
1800 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
1801 <ptsekov@syntrex.com> and Nedelcho Stanev.
1804 *) Implement a countermeasure against a vulnerability recently found
1805 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
1806 before application data chunks to avoid the use of known IVs
1807 with data potentially chosen by the attacker.
1810 *) Fix length checks in ssl3_get_client_hello().
1813 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
1814 to prevent ssl3_read_internal() from incorrectly assuming that
1815 ssl3_read_bytes() found application data while handshake
1816 processing was enabled when in fact s->s3->in_read_app_data was
1817 merely automatically cleared during the initial handshake.
1818 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
1820 *) Fix object definitions for Private and Enterprise: they were not
1821 recognized in their shortname (=lowercase) representation. Extend
1822 obj_dat.pl to issue an error when using undefined keywords instead
1823 of silently ignoring the problem (Svenning Sorensen
1824 <sss@sss.dnsalias.net>).
1827 *) Fix DH_generate_parameters() so that it works for 'non-standard'
1828 generators, i.e. generators other than 2 and 5. (Previously, the
1829 code did not properly initialise the 'add' and 'rem' values to
1830 BN_generate_prime().)
1832 In the new general case, we do not insist that 'generator' is
1833 actually a primitive root: This requirement is rather pointless;
1834 a generator of the order-q subgroup is just as good, if not
1838 *) Map new X509 verification errors to alerts. Discovered and submitted by
1839 Tom Wu <tom@arcot.com>.
1842 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
1843 returning non-zero before the data has been completely received
1844 when using non-blocking I/O.
1845 [Bodo Moeller; problem pointed out by John Hughes]
1847 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
1848 [Ben Laurie, Lutz Jaenicke]
1850 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
1851 Yoram Zahavi <YoramZ@gilian.com>).
1854 *) Add information about CygWin 1.3 and on, and preserve proper
1855 configuration for the versions before that.
1856 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1858 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
1859 check whether we deal with a copy of a session and do not delete from
1860 the cache in this case. Problem reported by "Izhar Shoshani Levi"
1861 <izhar@checkpoint.com>.
1864 *) Do not store session data into the internal session cache, if it
1865 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
1866 flag is set). Proposed by Aslam <aslam@funk.com>.
1869 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
1873 *) Add the configuration target linux-s390x.
1874 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
1876 *) [In 0.9.6d-engine release:]
1877 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1878 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1880 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
1881 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
1882 variable as an indication that a ClientHello message has been
1883 received. As the flag value will be lost between multiple
1884 invocations of ssl3_accept when using non-blocking I/O, the
1885 function may not be aware that a handshake has actually taken
1886 place, thus preventing a new session from being added to the
1889 To avoid this problem, we now set s->new_session to 2 instead of
1890 using a local variable.
1891 [Lutz Jaenicke, Bodo Moeller]
1893 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
1894 if the SSL_R_LENGTH_MISMATCH error is detected.
1895 [Geoff Thorpe, Bodo Moeller]
1897 *) New 'shared_ldflag' column in Configure platform table.
1900 *) Fix EVP_CIPHER_mode macro.
1901 ["Dan S. Camper" <dan@bti.net>]
1903 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
1904 type, we must throw them away by setting rr->length to 0.
1905 [D P Chang <dpc@qualys.com>]
1907 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
1909 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
1910 <Dominikus.Scherkl@biodata.com>. (The previous implementation
1911 worked incorrectly for those cases where range = 10..._2 and
1912 3*range is two bits longer than range.)
1915 *) Only add signing time to PKCS7 structures if it is not already
1919 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
1920 OBJ_ld_ce should be OBJ_id_ce.
1921 Also some ip-pda OIDs in crypto/objects/objects.txt were
1922 incorrect (cf. RFC 3039).
1923 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
1925 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
1926 returns early because it has nothing to do.
1927 [Andy Schneider <andy.schneider@bjss.co.uk>]
1929 *) [In 0.9.6c-engine release:]
1930 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
1931 [Andy Schneider <andy.schneider@bjss.co.uk>]
1933 *) [In 0.9.6c-engine release:]
1934 Add support for Cryptographic Appliance's keyserver technology.
1935 (Use engine 'keyclient')
1936 [Cryptographic Appliances and Geoff Thorpe]
1938 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
1939 is called via tools/c89.sh because arguments have to be
1940 rearranged (all '-L' options must appear before the first object
1942 [Richard Shapiro <rshapiro@abinitio.com>]
1944 *) [In 0.9.6c-engine release:]
1945 Add support for Broadcom crypto accelerator cards, backported
1947 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
1949 *) [In 0.9.6c-engine release:]
1950 Add support for SureWare crypto accelerator cards from
1951 Baltimore Technologies. (Use engine 'sureware')
1952 [Baltimore Technologies and Mark Cox]
1954 *) [In 0.9.6c-engine release:]
1955 Add support for crypto accelerator cards from Accelerated
1956 Encryption Processing, www.aep.ie. (Use engine 'aep')
1957 [AEP Inc. and Mark Cox]
1959 *) Add a configuration entry for gcc on UnixWare.
1960 [Gary Benson <gbenson@redhat.com>]
1962 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
1963 messages are stored in a single piece (fixed-length part and
1964 variable-length part combined) and fix various bugs found on the way.
1967 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
1968 instead. BIO_gethostbyname() does not know what timeouts are
1969 appropriate, so entries would stay in cache even when they have
1971 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
1973 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
1974 faced with a pathologically small ClientHello fragment that does
1975 not contain client_version: Instead of aborting with an error,
1976 simply choose the highest available protocol version (i.e.,
1977 TLS 1.0 unless it is disabled). In practice, ClientHello
1978 messages are never sent like this, but this change gives us
1979 strictly correct behaviour at least for TLS.
1982 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
1983 never resets s->method to s->ctx->method when called from within
1984 one of the SSL handshake functions.
1985 [Bodo Moeller; problem pointed out by Niko Baric]
1987 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
1988 (sent using the client's version number) if client_version is
1989 smaller than the protocol version in use. Also change
1990 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
1991 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
1992 the client will at least see that alert.
1995 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
1999 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2000 client receives HelloRequest while in a handshake.
2001 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2003 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2004 should end in 'break', not 'goto end' which circuments various
2005 cleanups done in state SSL_ST_OK. But session related stuff
2006 must be disabled for SSL_ST_OK in the case that we just sent a
2009 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2010 before just sending a HelloRequest.
2011 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2013 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2014 reveal whether illegal block cipher padding was found or a MAC
2015 verification error occured. (Neither SSLerr() codes nor alerts
2016 are directly visible to potential attackers, but the information
2017 may leak via logfiles.)
2019 Similar changes are not required for the SSL 2.0 implementation
2020 because the number of padding bytes is sent in clear for SSL 2.0,
2021 and the extra bytes are just ignored. However ssl/s2_pkt.c
2022 failed to verify that the purported number of padding bytes is in
2026 *) Add OpenUNIX-8 support including shared libraries
2027 (Boyd Lynn Gerber <gerberb@zenez.com>).
2030 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2031 'wristwatch attack' using huge encoding parameters (cf.
2032 James H. Manger's CRYPTO 2001 paper). Note that the
2033 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2034 encoding parameters and hence was not vulnerable.
2037 *) BN_sqr() bug fix.
2038 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2040 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2041 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2042 followed by modular reduction.
2043 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2045 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2046 equivalent based on BN_pseudo_rand() instead of BN_rand().
2049 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2050 This function was broken, as the check for a new client hello message
2051 to handle SGC did not allow these large messages.
2052 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2055 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2058 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2059 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2062 *) Rework the configuration and shared library support for Tru64 Unix.
2063 The configuration part makes use of modern compiler features and
2064 still retains old compiler behavior for those that run older versions
2065 of the OS. The shared library support part includes a variant that
2066 uses the RPATH feature, and is available through the special
2067 configuration target "alpha-cc-rpath", which will never be selected
2069 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2071 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2072 with the same message size as in ssl3_get_certificate_request().
2073 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2074 messages might inadvertently be reject as too long.
2075 [Petr Lampa <lampa@fee.vutbr.cz>]
2077 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2080 *) Modified SSL library such that the verify_callback that has been set
2081 specificly for an SSL object with SSL_set_verify() is actually being
2082 used. Before the change, a verify_callback set with this function was
2083 ignored and the verify_callback() set in the SSL_CTX at the time of
2084 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2085 to allow the necessary settings.
2088 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2089 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2090 done automatically (in contradiction to the requirements of the C
2091 standard). This made problems when used from OpenSSH.
2094 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2095 dh->length and always used
2097 BN_rand_range(priv_key, dh->p).
2099 BN_rand_range() is not necessary for Diffie-Hellman, and this
2100 specific range makes Diffie-Hellman unnecessarily inefficient if
2101 dh->length (recommended exponent length) is much smaller than the
2102 length of dh->p. We could use BN_rand_range() if the order of
2103 the subgroup was stored in the DH structure, but we only have
2108 BN_rand(priv_key, l, ...)
2110 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2116 RSA_eay_public_encrypt
2117 RSA_eay_private_decrypt
2118 RSA_eay_private_encrypt (signing)
2119 RSA_eay_public_decrypt (signature verification)
2121 (default implementations for RSA_public_encrypt,
2122 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2123 always reject numbers >= n.
2126 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2127 to synchronize access to 'locking_thread'. This is necessary on
2128 systems where access to 'locking_thread' (an 'unsigned long'
2129 variable) is not atomic.
2132 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2133 *before* setting the 'crypto_lock_rand' flag. The previous code had
2134 a race condition if 0 is a valid thread ID.
2135 [Travis Vitek <vitek@roguewave.com>]
2137 *) Add support for shared libraries under Irix.
2138 [Albert Chin-A-Young <china@thewrittenword.com>]
2140 *) Add configuration option to build on Linux on both big-endian and
2142 [Ralf Baechle <ralf@uni-koblenz.de>]
2144 *) Add the possibility to create shared libraries on HP-UX.
2147 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2149 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2150 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2151 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2152 PRNG state recovery was possible based on the output of
2153 one PRNG request appropriately sized to gain knowledge on
2154 'md' followed by enough consecutive 1-byte PRNG requests
2155 to traverse all of 'state'.
2157 1. When updating 'md_local' (the current thread's copy of 'md')
2158 during PRNG output generation, hash all of the previous
2159 'md_local' value, not just the half used for PRNG output.
2161 2. Make the number of bytes from 'state' included into the hash
2162 independent from the number of PRNG bytes requested.
2164 The first measure alone would be sufficient to avoid
2165 Markku-Juhani's attack. (Actually it had never occurred
2166 to me that the half of 'md_local' used for chaining was the
2167 half from which PRNG output bytes were taken -- I had always
2168 assumed that the secret half would be used.) The second
2169 measure makes sure that additional data from 'state' is never
2170 mixed into 'md_local' in small portions; this heuristically
2171 further strengthens the PRNG.
2174 *) Fix crypto/bn/asm/mips3.s.
2177 *) When only the key is given to "enc", the IV is undefined. Print out
2178 an error message in this case.
2181 *) Handle special case when X509_NAME is empty in X509 printing routines.
2184 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2185 positive and less than q.
2188 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2189 used: it isn't thread safe and the add_lock_callback should handle
2191 [Paul Rose <Paul.Rose@bridge.com>]
2193 *) Verify that incoming data obeys the block size in
2194 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2198 [Ulf Möller, Bodo Möller]
2200 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2201 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2202 when fixing the server behaviour for backwards-compatible 'client
2203 hello' messages. (Note that the attack is impractical against
2204 SSL 3.0 and TLS 1.0 anyway because length and version checking
2205 means that the probability of guessing a valid ciphertext is
2206 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2209 Before 0.9.5, the countermeasure (hide the error by generating a
2210 random 'decryption result') did not work properly because
2211 ERR_clear_error() was missing, meaning that SSL_get_error() would
2212 detect the supposedly ignored error.
2214 Both problems are now fixed.
2217 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2218 (previously it was 1024).
2221 *) Fix for compatibility mode trust settings: ignore trust settings
2222 unless some valid trust or reject settings are present.
2225 *) Fix for blowfish EVP: its a variable length cipher.
2228 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2229 parameters in DSA public key structures and return an error in the
2230 DSA routines if parameters are absent.
2233 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2234 in the current directory if neither $RANDFILE nor $HOME was set.
2235 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2236 caused some confusion to Windows users who haven't defined $HOME.
2237 Thus RAND_file_name() is changed again: e_os.h can define a
2238 DEFAULT_HOME, which will be used if $HOME is not set.
2239 For Windows, we use "C:"; on other platforms, we still require
2240 environment variables.
2242 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2243 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2244 having multiple threads call RAND_poll() concurrently.
2247 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2248 combination of a flag and a thread ID variable.
2249 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2250 flag), *other* threads can enter ssleay_add_bytes without obeying
2251 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2252 that they do not hold after the first thread unsets add_do_not_lock).
2255 *) Change bctest again: '-x' expressions are not available in all
2259 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2261 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2262 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2264 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2265 the default extension for executables, if any. Also, make the perl
2266 scripts that use symlink() to test if it really exists and use "cp"
2267 if it doesn't. All this made OpenSSL compilable and installable in
2271 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2272 If SEQUENCE is length is indefinite just set c->slen to the total
2273 amount of data available.
2274 [Steve Henson, reported by shige@FreeBSD.org]
2275 [This change does not apply to 0.9.7.]
2277 *) Change bctest to avoid here-documents inside command substitution
2278 (workaround for FreeBSD /bin/sh bug).
2279 For compatibility with Ultrix, avoid shell functions (introduced
2280 in the bctest version that searches along $PATH).
2283 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2284 with des_encrypt() defined on some operating systems, like Solaris
2288 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2289 On the Importance of Eliminating Errors in Cryptographic
2290 Computations, J. Cryptology 14 (2001) 2, 101-119,
2291 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2294 *) MIPS assembler BIGNUM division bug fix.
2297 *) Disabled incorrect Alpha assembler code.
2300 *) Fix PKCS#7 decode routines so they correctly update the length
2301 after reading an EOC for the EXPLICIT tag.
2303 [This change does not apply to 0.9.7.]
2305 *) Fix bug in PKCS#12 key generation routines. This was triggered
2306 if a 3DES key was generated with a 0 initial byte. Include
2307 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2308 (but broken) behaviour.
2311 *) Enhance bctest to search for a working bc along $PATH and print
2313 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2315 *) Fix memory leaks in err.c: free err_data string if necessary;
2316 don't write to the wrong index in ERR_set_error_data.
2319 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2323 *) Replace rdtsc with _emit statements for VC++ version 5.
2324 [Jeremy Cooper <jeremy@baymoo.org>]
2326 *) Make it possible to reuse SSLv2 sessions.
2329 *) In copy_email() check for >= 0 as a return value for
2330 X509_NAME_get_index_by_NID() since 0 is a valid index.
2331 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2333 *) Avoid coredump with unsupported or invalid public keys by checking if
2334 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2335 PKCS7_verify() fails with non detached data.
2338 *) Don't use getenv in library functions when run as setuid/setgid.
2339 New function OPENSSL_issetugid().
2342 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2343 due to incorrect handling of multi-threading:
2345 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2347 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2349 3. Count how many times MemCheck_off() has been called so that
2350 nested use can be treated correctly. This also avoids
2351 inband-signalling in the previous code (which relied on the
2352 assumption that thread ID 0 is impossible).
2355 *) Add "-rand" option also to s_client and s_server.
2358 *) Fix CPU detection on Irix 6.x.
2359 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2360 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2362 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2365 [This change does not apply to 0.9.7.]
2367 *) Use the cached encoding of an X509_NAME structure rather than
2368 copying it. This is apparently the reason for the libsafe "errors"
2369 but the code is actually correct.
2372 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2373 Bleichenbacher's DSA attack.
2374 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2375 to be set and top=0 forces the highest bit to be set; top=-1 is new
2376 and leaves the highest bit random.
2377 [Ulf Moeller, Bodo Moeller]
2379 *) In the NCONF_...-based implementations for CONF_... queries
2380 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2381 a temporary CONF structure with the data component set to NULL
2382 (which gives segmentation faults in lh_retrieve).
2383 Instead, use NULL for the CONF pointer in CONF_get_string and
2384 CONF_get_number (which may use environment variables) and directly
2385 return NULL from CONF_get_section.
2388 *) Fix potential buffer overrun for EBCDIC.
2391 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2392 keyUsage if basicConstraints absent for a CA.
2395 *) Make SMIME_write_PKCS7() write mail header values with a format that
2396 is more generally accepted (no spaces before the semicolon), since
2397 some programs can't parse those values properly otherwise. Also make
2398 sure BIO's that break lines after each write do not create invalid
2402 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2403 macros previously used would not encode an empty SEQUENCE OF
2404 and break the signature.
2406 [This change does not apply to 0.9.7.]
2408 *) Zero the premaster secret after deriving the master secret in
2412 *) Add some EVP_add_digest_alias registrations (as found in
2413 OpenSSL_add_all_digests()) to SSL_library_init()
2414 aka OpenSSL_add_ssl_algorithms(). This provides improved
2415 compatibility with peers using X.509 certificates
2416 with unconventional AlgorithmIdentifier OIDs.
2419 *) Fix for Irix with NO_ASM.
2420 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2422 *) ./config script fixes.
2423 [Ulf Moeller, Richard Levitte]
2425 *) Fix 'openssl passwd -1'.
2428 *) Change PKCS12_key_gen_asc() so it can cope with non null
2429 terminated strings whose length is passed in the passlen
2430 parameter, for example from PEM callbacks. This was done
2431 by adding an extra length parameter to asc2uni().
2432 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2434 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2435 call failed, free the DSA structure.
2438 *) Fix to uni2asc() to cope with zero length Unicode strings.
2439 These are present in some PKCS#12 files.
2442 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2443 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2444 when writing a 32767 byte record.
2445 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2447 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2448 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2450 (RSA objects have a reference count access to which is protected
2451 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2452 so they are meant to be shared between threads.)
2453 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2454 "Reddie, Steven" <Steven.Reddie@ca.com>]
2456 *) Fix a deadlock in CRYPTO_mem_leaks().
2459 *) Use better test patterns in bntest.
2462 *) rand_win.c fix for Borland C.
2465 *) BN_rshift bugfix for n == 0.
2468 *) Add a 'bctest' script that checks for some known 'bc' bugs
2469 so that 'make test' does not abort just because 'bc' is broken.
2472 *) Store verify_result within SSL_SESSION also for client side to
2473 avoid potential security hole. (Re-used sessions on the client side
2474 always resulted in verify_result==X509_V_OK, not using the original
2475 result of the server certificate verification.)
2478 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2479 SSL3_RT_APPLICATION_DATA, return 0.
2480 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2484 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2485 releases, have been re-implemented by renaming the previous
2486 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2487 and ssl3_read_internal, respectively, and adding 'peek' parameters
2488 to them. The new ssl[23]_{read,peek} functions are calls to
2489 ssl[23]_read_internal with the 'peek' flag set appropriately.
2490 A 'peek' parameter has also been added to ssl3_read_bytes, which
2491 does the actual work for ssl3_read_internal.
2494 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2495 the method-specific "init()" handler. Also clean up ex_data after
2496 calling the method-specific "finish()" handler. Previously, this was
2497 happening the other way round.
2500 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2501 The previous value, 12, was not always sufficient for BN_mod_exp().
2504 *) Make sure that shared libraries get the internal name engine with
2505 the full version number and not just 0. This should mark the
2506 shared libraries as not backward compatible. Of course, this should
2507 be changed again when we can guarantee backward binary compatibility.
2510 *) Fix typo in get_cert_by_subject() in by_dir.c
2511 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2513 *) Rework the system to generate shared libraries:
2515 - Make note of the expected extension for the shared libraries and
2516 if there is a need for symbolic links from for example libcrypto.so.0
2517 to libcrypto.so.0.9.7. There is extended info in Configure for
2520 - Make as few rebuilds of the shared libraries as possible.
2522 - Still avoid linking the OpenSSL programs with the shared libraries.
2524 - When installing, install the shared libraries separately from the
2528 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2530 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2531 and not in SSL_clear because the latter is also used by the
2532 accept/connect functions; previously, the settings made by
2533 SSL_set_read_ahead would be lost during the handshake.
2534 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2536 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2537 Previously, it would create entries for disableed algorithms no
2541 *) Added several new manual pages for SSL_* function.
2544 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2546 *) In ssl23_get_client_hello, generate an error message when faced
2547 with an initial SSL 3.0/TLS record that is too small to contain the
2548 first two bytes of the ClientHello message, i.e. client_version.
2549 (Note that this is a pathologic case that probably has never happened
2550 in real life.) The previous approach was to use the version number
2551 from the record header as a substitute; but our protocol choice
2552 should not depend on that one because it is not authenticated
2553 by the Finished messages.
2556 *) More robust randomness gathering functions for Windows.
2557 [Jeffrey Altman <jaltman@columbia.edu>]
2559 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2560 not set then we don't setup the error code for issuer check errors
2561 to avoid possibly overwriting other errors which the callback does
2562 handle. If an application does set the flag then we assume it knows
2563 what it is doing and can handle the new informational codes
2567 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2568 a general "ANY" type, as such it should be able to decode anything
2569 including tagged types. However it didn't check the class so it would
2570 wrongly interpret tagged types in the same way as their universal
2571 counterpart and unknown types were just rejected. Changed so that the
2572 tagged and unknown types are handled in the same way as a SEQUENCE:
2573 that is the encoding is stored intact. There is also a new type
2574 "V_ASN1_OTHER" which is used when the class is not universal, in this
2575 case we have no idea what the actual type is so we just lump them all
2579 *) On VMS, stdout may very well lead to a file that is written to
2580 in a record-oriented fashion. That means that every write() will
2581 write a separate record, which will be read separately by the
2582 programs trying to read from it. This can be very confusing.
2584 The solution is to put a BIO filter in the way that will buffer
2585 text until a linefeed is reached, and then write everything a
2586 line at a time, so every record written will be an actual line,
2587 not chunks of lines and not (usually doesn't happen, but I've
2588 seen it once) several lines in one record. BIO_f_linebuffer() is
2591 Currently, it's a VMS-only method, because that's where it has
2592 been tested well enough.
2595 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2596 it can return incorrect results.
2597 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2598 but it was in 0.9.6-beta[12].)
2601 *) Disable the check for content being present when verifying detached
2602 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2603 include zero length content when signing messages.
2606 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2607 BIO_ctrl (for BIO pairs).
2610 *) Add DSO method for VMS.
2613 *) Bug fix: Montgomery multiplication could produce results with the
2617 *) Add RPM specification openssl.spec and modify it to build three
2618 packages. The default package contains applications, application
2619 documentation and run-time libraries. The devel package contains
2620 include files, static libraries and function documentation. The
2621 doc package contains the contents of the doc directory. The original
2622 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2625 *) Add a large number of documentation files for many SSL routines.
2626 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2628 *) Add a configuration entry for Sony News 4.
2629 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2631 *) Don't set the two most significant bits to one when generating a
2632 random number < q in the DSA library.
2635 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2636 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2637 the underlying transport is blocking) if a handshake took place.
2638 (The default behaviour is needed by applications such as s_client
2639 and s_server that use select() to determine when to use SSL_read;
2640 but for applications that know in advance when to expect data, it
2641 just makes things more complicated.)
2644 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2648 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2649 work better on such systems.
2650 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2652 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2653 Update PKCS12_parse() so it copies the friendlyName and the
2654 keyid to the certificates aux info.
2657 *) Fix bug in PKCS7_verify() which caused an infinite loop
2658 if there was more than one signature.
2659 [Sven Uszpelkat <su@celocom.de>]
2661 *) Major change in util/mkdef.pl to include extra information
2662 about each symbol, as well as presentig variables as well
2663 as functions. This change means that there's n more need
2664 to rebuild the .num files when some algorithms are excluded.
2667 *) Allow the verify time to be set by an application,
2668 rather than always using the current time.
2671 *) Phase 2 verify code reorganisation. The certificate
2672 verify code now looks up an issuer certificate by a
2673 number of criteria: subject name, authority key id
2674 and key usage. It also verifies self signed certificates
2675 by the same criteria. The main comparison function is
2676 X509_check_issued() which performs these checks.
2678 Lot of changes were necessary in order to support this
2679 without completely rewriting the lookup code.
2681 Authority and subject key identifier are now cached.
2683 The LHASH 'certs' is X509_STORE has now been replaced
2684 by a STACK_OF(X509_OBJECT). This is mainly because an
2685 LHASH can't store or retrieve multiple objects with
2686 the same hash value.
2688 As a result various functions (which were all internal
2689 use only) have changed to handle the new X509_STORE
2690 structure. This will break anything that messed round
2691 with X509_STORE internally.
2693 The functions X509_STORE_add_cert() now checks for an
2694 exact match, rather than just subject name.
2696 The X509_STORE API doesn't directly support the retrieval
2697 of multiple certificates matching a given criteria, however
2698 this can be worked round by performing a lookup first
2699 (which will fill the cache with candidate certificates)
2700 and then examining the cache for matches. This is probably
2701 the best we can do without throwing out X509_LOOKUP
2702 entirely (maybe later...).
2704 The X509_VERIFY_CTX structure has been enhanced considerably.
2706 All certificate lookup operations now go via a get_issuer()
2707 callback. Although this currently uses an X509_STORE it
2708 can be replaced by custom lookups. This is a simple way
2709 to bypass the X509_STORE hackery necessary to make this
2710 work and makes it possible to use more efficient techniques
2711 in future. A very simple version which uses a simple
2712 STACK for its trusted certificate store is also provided
2713 using X509_STORE_CTX_trusted_stack().
2715 The verify_cb() and verify() callbacks now have equivalents
2716 in the X509_STORE_CTX structure.
2718 X509_STORE_CTX also has a 'flags' field which can be used
2719 to customise the verify behaviour.
2722 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2723 excludes S/MIME capabilities.
2726 *) When a certificate request is read in keep a copy of the
2727 original encoding of the signed data and use it when outputing
2728 again. Signatures then use the original encoding rather than
2729 a decoded, encoded version which may cause problems if the
2730 request is improperly encoded.
2733 *) For consistency with other BIO_puts implementations, call
2734 buffer_write(b, ...) directly in buffer_puts instead of calling
2737 In BIO_puts, increment b->num_write as in BIO_write.
2738 [Peter.Sylvester@EdelWeb.fr]
2740 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2741 BN_zero, we may not return a BIGNUM with an array consisting of
2745 *) Avoid calling abort() from within the library when problems are
2746 detected, except if preprocessor symbols have been defined
2747 (such as REF_CHECK, BN_DEBUG etc.).
2750 *) New openssl application 'rsautl'. This utility can be
2751 used for low level RSA operations. DER public key
2752 BIO/fp routines also added.
2755 *) New Configure entry and patches for compiling on QNX 4.
2756 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2758 *) A demo state-machine implementation was sponsored by
2759 Nuron (http://www.nuron.com/) and is now available in
2760 demos/state_machine.
2763 *) New options added to the 'dgst' utility for signature
2764 generation and verification.
2767 *) Unrecognized PKCS#7 content types are now handled via a
2768 catch all ASN1_TYPE structure. This allows unsupported
2769 types to be stored as a "blob" and an application can
2770 encode and decode it manually.
2773 *) Fix various signed/unsigned issues to make a_strex.c
2775 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2777 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2778 length if passed a buffer. ASN1_INTEGER_to_BN failed
2779 if passed a NULL BN and its argument was negative.
2780 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2782 *) Modification to PKCS#7 encoding routines to output definite
2783 length encoding. Since currently the whole structures are in
2784 memory there's not real point in using indefinite length
2785 constructed encoding. However if OpenSSL is compiled with
2786 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2789 *) Added BIO_vprintf() and BIO_vsnprintf().
2792 *) Added more prefixes to parse for in the the strings written
2793 through a logging bio, to cover all the levels that are available
2794 through syslog. The prefixes are now:
2796 PANIC, EMERG, EMR => LOG_EMERG
2797 ALERT, ALR => LOG_ALERT
2798 CRIT, CRI => LOG_CRIT
2799 ERROR, ERR => LOG_ERR
2800 WARNING, WARN, WAR => LOG_WARNING
2801 NOTICE, NOTE, NOT => LOG_NOTICE
2802 INFO, INF => LOG_INFO
2803 DEBUG, DBG => LOG_DEBUG
2805 and as before, if none of those prefixes are present at the
2806 beginning of the string, LOG_ERR is chosen.
2808 On Win32, the LOG_* levels are mapped according to this:
2810 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2811 LOG_WARNING => EVENTLOG_WARNING_TYPE
2812 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2816 *) Made it possible to reconfigure with just the configuration
2817 argument "reconf" or "reconfigure". The command line arguments
2818 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2819 and are retrieved from there when reconfiguring.
2823 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2825 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2828 *) The obj_dat.pl script was messing up the sorting of object
2829 names. The reason was that it compared the quoted version
2830 of strings as a result "OCSP" > "OCSP Signing" because
2831 " > SPACE. Changed script to store unquoted versions of
2832 names and add quotes on output. It was also omitting some
2833 names from the lookup table if they were given a default
2834 value (that is if SN is missing it is given the same
2835 value as LN and vice versa), these are now added on the
2836 grounds that if an object has a name we should be able to
2837 look it up. Finally added warning output when duplicate
2838 short or long names are found.
2841 *) Changes needed for Tandem NSK.
2842 [Scott Uroff <scott@xypro.com>]
2844 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2845 RSA_padding_check_SSLv23(), special padding was never detected
2846 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2847 version rollback attacks was not effective.
2849 In s23_clnt.c, don't use special rollback-attack detection padding
2850 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2851 client; similarly, in s23_srvr.c, don't do the rollback check if
2852 SSL 2.0 is the only protocol enabled in the server.
2855 *) Make it possible to get hexdumps of unprintable data with 'openssl
2856 asn1parse'. By implication, the functions ASN1_parse_dump() and
2857 BIO_dump_indent() are added.
2860 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2861 these print out strings and name structures based on various
2862 flags including RFC2253 support and proper handling of
2863 multibyte characters. Added options to the 'x509' utility
2864 to allow the various flags to be set.
2867 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2868 Also change the functions X509_cmp_current_time() and
2869 X509_gmtime_adj() work with an ASN1_TIME structure,
2870 this will enable certificates using GeneralizedTime in validity
2871 dates to be checked.
2874 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2875 negative public key encodings) on by default,
2876 NO_NEG_PUBKEY_BUG can be set to disable it.
2879 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2880 content octets. An i2c_ASN1_OBJECT is unnecessary because
2881 the encoding can be trivially obtained from the structure.
2884 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2885 not read locks (CRYPTO_r_[un]lock).
2888 *) A first attempt at creating official support for shared
2889 libraries through configuration. I've kept it so the
2890 default is static libraries only, and the OpenSSL programs
2891 are always statically linked for now, but there are
2892 preparations for dynamic linking in place.
2893 This has been tested on Linux and Tru64.
2896 *) Randomness polling function for Win9x, as described in:
2897 Peter Gutmann, Software Generation of Practically Strong
2901 *) Fix so PRNG is seeded in req if using an already existing
2905 *) New options to smime application. -inform and -outform
2906 allow alternative formats for the S/MIME message including
2907 PEM and DER. The -content option allows the content to be
2908 specified separately. This should allow things like Netscape
2909 form signing output easier to verify.
2912 *) Fix the ASN1 encoding of tags using the 'long form'.
2915 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2916 STRING types. These convert content octets to and from the
2917 underlying type. The actual tag and length octets are
2918 already assumed to have been read in and checked. These
2919 are needed because all other string types have virtually
2920 identical handling apart from the tag. By having versions
2921 of the ASN1 functions that just operate on content octets
2922 IMPLICIT tagging can be handled properly. It also allows
2923 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2924 and ASN1_INTEGER are identical apart from the tag.
2927 *) Change the handling of OID objects as follows:
2929 - New object identifiers are inserted in objects.txt, following
2930 the syntax given in objects.README.
2931 - objects.pl is used to process obj_mac.num and create a new
2933 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2936 This is currently kind of a hack, and the perl code in objects.pl
2937 isn't very elegant, but it works as I intended. The simplest way
2938 to check that it worked correctly is to look in obj_dat.h and
2939 check the array nid_objs and make sure the objects haven't moved
2940 around (this is important!). Additions are OK, as well as
2941 consistent name changes.
2944 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2947 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2948 The given file adds to whatever has already been seeded into the
2949 random pool through the RANDFILE configuration file option or
2950 environment variable, or the default random state file.
2953 *) mkstack.pl now sorts each macro group into lexical order.
2954 Previously the output order depended on the order the files
2955 appeared in the directory, resulting in needless rewriting
2959 *) Patches to make OpenSSL compile under Win32 again. Mostly
2960 work arounds for the VC++ problem that it treats func() as
2961 func(void). Also stripped out the parts of mkdef.pl that
2962 added extra typesafe functions: these no longer exist.
2965 *) Reorganisation of the stack code. The macros are now all
2966 collected in safestack.h . Each macro is defined in terms of
2967 a "stack macro" of the form SKM_<name>(type, a, b). The
2968 DEBUG_SAFESTACK is now handled in terms of function casts,
2969 this has the advantage of retaining type safety without the
2970 use of additional functions. If DEBUG_SAFESTACK is not defined
2971 then the non typesafe macros are used instead. Also modified the
2972 mkstack.pl script to handle the new form. Needs testing to see
2973 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2974 the default if no major problems. Similar behaviour for ASN1_SET_OF
2975 and PKCS12_STACK_OF.
2978 *) When some versions of IIS use the 'NET' form of private key the
2979 key derivation algorithm is different. Normally MD5(password) is
2980 used as a 128 bit RC4 key. In the modified case
2981 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2982 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2983 as the old Netscape_RSA functions except they have an additional
2984 'sgckey' parameter which uses the modified algorithm. Also added
2985 an -sgckey command line option to the rsa utility. Thanks to
2986 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2987 algorithm to openssl-dev.
2990 *) The evp_local.h macros were using 'c.##kname' which resulted in
2991 invalid expansion on some systems (SCO 5.0.5 for example).
2992 Corrected to 'c.kname'.
2993 [Phillip Porch <root@theporch.com>]
2995 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2996 a STACK of email addresses from a certificate or request, these look
2997 in the subject name and the subject alternative name extensions and
2998 omit any duplicate addresses.
3001 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3002 This makes DSA verification about 2 % faster.
3005 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3006 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3007 plus overhead for 1024 bit moduli).
3008 This makes exponentiations about 0.5 % faster for 1024 bit
3009 exponents (as measured by "openssl speed rsa2048").
3012 *) Rename memory handling macros to avoid conflicts with other
3014 Malloc => OPENSSL_malloc
3015 Malloc_locked => OPENSSL_malloc_locked
3016 Realloc => OPENSSL_realloc
3017 Free => OPENSSL_free
3020 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3021 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3024 *) CygWin32 support.
3025 [John Jarvie <jjarvie@newsguy.com>]
3027 *) The type-safe stack code has been rejigged. It is now only compiled
3028 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3029 by default all type-specific stack functions are "#define"d back to
3030 standard stack functions. This results in more streamlined output
3031 but retains the type-safety checking possibilities of the original
3035 *) The STACK code has been cleaned up, and certain type declarations
3036 that didn't make a lot of sense have been brought in line. This has
3037 also involved a cleanup of sorts in safestack.h to more correctly
3038 map type-safe stack functions onto their plain stack counterparts.
3039 This work has also resulted in a variety of "const"ifications of
3040 lots of the code, especially "_cmp" operations which should normally
3041 be prototyped with "const" parameters anyway.
3044 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3045 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3046 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3047 where all of 'md' is used each time the PRNG is used, but 'state'
3048 is used only indexed by a cyclic counter. As entropy may not be
3049 well distributed from the beginning, 'md' is important as a
3050 chaining variable. However, the output function chains only half
3051 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3052 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3053 in all of 'state' being rewritten, with the new values depending
3054 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3057 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3058 the handshake is continued after ssl_verify_cert_chain();
3059 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3060 can lead to 'unexplainable' connection aborts later.
3061 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3063 *) Major EVP API cipher revision.
3064 Add hooks for extra EVP features. This allows various cipher
3065 parameters to be set in the EVP interface. Support added for variable
3066 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3067 setting of RC2 and RC5 parameters.
3069 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3072 Remove lots of duplicated code from the EVP library. For example *every*
3073 cipher init() function handles the 'iv' in the same way according to the
3074 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3075 for CFB and OFB modes they zero ctx->num.
3077 New functionality allows removal of S/MIME code RC2 hack.
3079 Most of the routines have the same form and so can be declared in terms
3082 By shifting this to the top level EVP_CipherInit() it can be removed from
3083 all individual ciphers. If the cipher wants to handle IVs or keys
3084 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3087 Change lots of functions like EVP_EncryptUpdate() to now return a
3088 value: although software versions of the algorithms cannot fail
3089 any installed hardware versions can.
3092 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3093 this option is set, tolerate broken clients that send the negotiated
3094 protocol version number instead of the requested protocol version
3098 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3099 i.e. non-zero for export ciphersuites, zero otherwise.
3100 Previous versions had this flag inverted, inconsistent with
3101 rsa_tmp_cb (..._TMP_RSA_CB).
3102 [Bodo Moeller; problem reported by Amit Chopra]
3104 *) Add missing DSA library text string. Work around for some IIS
3105 key files with invalid SEQUENCE encoding.
3108 *) Add a document (doc/standards.txt) that list all kinds of standards
3109 and so on that are implemented in OpenSSL.
3112 *) Enhance c_rehash script. Old version would mishandle certificates
3113 with the same subject name hash and wouldn't handle CRLs at all.
3114 Added -fingerprint option to crl utility, to support new c_rehash
3118 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3121 *) Fix for SSL server purpose checking. Server checking was
3122 rejecting certificates which had extended key usage present
3123 but no ssl client purpose.
3124 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3126 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3127 is a little unclear about how a blank password is handled.
3128 Since the password in encoded as a BMPString with terminating
3129 double NULL a zero length password would end up as just the
3130 double NULL. However no password at all is different and is
3131 handled differently in the PKCS#12 key generation code. NS
3132 treats a blank password as zero length. MSIE treats it as no
3133 password on export: but it will try both on import. We now do
3134 the same: PKCS12_parse() tries zero length and no password if
3135 the password is set to "" or NULL (NULL is now a valid password:
3136 it wasn't before) as does the pkcs12 application.
3139 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3140 perror when PEM_read_bio_X509_REQ fails, the error message must
3141 be obtained from the error queue.
3144 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3145 it in ERR_remove_state if appropriate, and change ERR_get_state
3146 accordingly to avoid race conditions (this is necessary because
3147 thread_hash is no longer constant once set).
3150 *) Bugfix for linux-elf makefile.one.
3153 *) RSA_get_default_method() will now cause a default
3154 RSA_METHOD to be chosen if one doesn't exist already.
3155 Previously this was only set during a call to RSA_new()
3156 or RSA_new_method(NULL) meaning it was possible for
3157 RSA_get_default_method() to return NULL.
3160 *) Added native name translation to the existing DSO code
3161 that will convert (if the flag to do so is set) filenames
3162 that are sufficiently small and have no path information
3163 into a canonical native form. Eg. "blah" converted to
3164 "libblah.so" or "blah.dll" etc.
3167 *) New function ERR_error_string_n(e, buf, len) which is like
3168 ERR_error_string(e, buf), but writes at most 'len' bytes
3169 including the 0 terminator. For ERR_error_string_n, 'buf'
3171 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3173 *) CONF library reworked to become more general. A new CONF
3174 configuration file reader "class" is implemented as well as a
3175 new functions (NCONF_*, for "New CONF") to handle it. The now
3176 old CONF_* functions are still there, but are reimplemented to
3177 work in terms of the new functions. Also, a set of functions
3178 to handle the internal storage of the configuration data is
3179 provided to make it easier to write new configuration file
3180 reader "classes" (I can definitely see something reading a
3181 configuration file in XML format, for example), called _CONF_*,
3182 or "the configuration storage API"...
3184 The new configuration file reading functions are:
3186 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3187 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3189 NCONF_default, NCONF_WIN32
3191 NCONF_dump_fp, NCONF_dump_bio
3193 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3194 NCONF_new creates a new CONF object. This works in the same way
3195 as other interfaces in OpenSSL, like the BIO interface.
3196 NCONF_dump_* dump the internal storage of the configuration file,
3197 which is useful for debugging. All other functions take the same
3198 arguments as the old CONF_* functions wth the exception of the
3199 first that must be a `CONF *' instead of a `LHASH *'.
3201 To make it easer to use the new classes with the old CONF_* functions,
3202 the function CONF_set_default_method is provided.
3205 *) Add '-tls1' option to 'openssl ciphers', which was already
3206 mentioned in the documentation but had not been implemented.
3207 (This option is not yet really useful because even the additional
3208 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3211 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3212 OpenSSL-based applications) load shared libraries and bind to
3213 them in a portable way.
3214 [Geoff Thorpe, with contributions from Richard Levitte]
3216 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3218 *) Make sure _lrotl and _lrotr are only used with MSVC.
3220 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3221 (the default implementation of RAND_status).
3223 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3224 to '-clrext' (= clear extensions), as intended and documented.
3225 [Bodo Moeller; inconsistency pointed out by Michael Attili
3228 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3229 was larger than the MD block size.
3230 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3232 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3233 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3234 using the passed key: if the passed key was a private key the result
3235 of X509_print(), for example, would be to print out all the private key
3239 *) des_quad_cksum() byte order bug fix.
3240 [Ulf Möller, using the problem description in krb4-0.9.7, where
3241 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3243 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3245 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3247 *) For easily testing in shell scripts whether some command
3248 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3249 returns with exit code 0 iff no command of the given name is available.
3250 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3251 the output goes to stdout and nothing is printed to stderr.
3252 Additional arguments are always ignored.
3254 Since for each cipher there is a command of the same name,
3255 the 'no-cipher' compilation switches can be tested this way.
3257 ('openssl no-XXX' is not able to detect pseudo-commands such
3258 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3261 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3264 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3265 is set; it will be thrown away anyway because each handshake creates
3267 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3268 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3269 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3270 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3273 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3274 'Q' and 'R' lose their special meanings (quit/renegotiate).
3275 This is part of what -quiet does; unlike -quiet, -ign_eof
3276 does not suppress any output.
3279 *) Add compatibility options to the purpose and trust code. The
3280 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3281 accepts a certificate or CA, this was the previous behaviour,
3282 with all the associated security issues.
3284 X509_TRUST_COMPAT is the old trust behaviour: only and
3285 automatically trust self signed roots in certificate store. A
3286 new trust setting X509_TRUST_DEFAULT is used to specify that
3287 a purpose has no associated trust setting and it should instead
3288 use the value in the default purpose.
3291 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3292 and fix a memory leak.
3295 *) In util/mkerr.pl (which implements 'make errors'), preserve
3296 reason strings from the previous version of the .c file, as
3297 the default to have only downcase letters (and digits) in
3298 automatically generated reasons codes is not always appropriate.
3301 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3302 using strerror. Previously, ERR_reason_error_string() returned
3303 library names as reason strings for SYSerr; but SYSerr is a special
3304 case where small numbers are errno values, not library numbers.
3307 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3308 converts DSA parameters into DH parameters. (When creating parameters,
3309 DSA_generate_parameters is used.)
3312 *) Include 'length' (recommended exponent length) in C code generated
3313 by 'openssl dhparam -C'.
3316 *) The second argument to set_label in perlasm was already being used
3317 so couldn't be used as a "file scope" flag. Moved to third argument
3321 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3322 instead of RAND_bytes for encryption IVs and salts.
3325 *) Include RAND_status() into RAND_METHOD instead of implementing
3326 it only for md_rand.c Otherwise replacing the PRNG by calling
3327 RAND_set_rand_method would be impossible.
3330 *) Don't let DSA_generate_key() enter an infinite loop if the random
3331 number generation fails.
3334 *) New 'rand' application for creating pseudo-random output.
3337 *) Added configuration support for Linux/IA64
3338 [Rolf Haberrecker <rolf@suse.de>]
3340 *) Assembler module support for Mingw32.
3343 *) Shared library support for HPUX (in shlib/).
3344 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3346 *) Shared library support for Solaris gcc.
3347 [Lutz Behnke <behnke@trustcenter.de>]
3349 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3351 *) PKCS7_encrypt() was adding text MIME headers twice because they
3352 were added manually and by SMIME_crlf_copy().
3355 *) In bntest.c don't call BN_rand with zero bits argument.
3356 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3358 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3359 case was implemented. This caused BN_div_recp() to fail occasionally.
3362 *) Add an optional second argument to the set_label() in the perl
3363 assembly language builder. If this argument exists and is set
3364 to 1 it signals that the assembler should use a symbol whose
3365 scope is the entire file, not just the current function. This
3366 is needed with MASM which uses the format label:: for this scope.
3367 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3369 *) Change the ASN1 types so they are typedefs by default. Before
3370 almost all types were #define'd to ASN1_STRING which was causing
3371 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3375 *) Change names of new functions to the new get1/get0 naming
3376 convention: After 'get1', the caller owns a reference count
3377 and has to call ..._free; 'get0' returns a pointer to some
3378 data structure without incrementing reference counters.
3379 (Some of the existing 'get' functions increment a reference
3380 counter, some don't.)
3381 Similarly, 'set1' and 'add1' functions increase reference
3382 counters or duplicate objects.
3385 *) Allow for the possibility of temp RSA key generation failure:
3386 the code used to assume it always worked and crashed on failure.
3389 *) Fix potential buffer overrun problem in BIO_printf().
3390 [Ulf Möller, using public domain code by Patrick Powell; problem
3391 pointed out by David Sacerdote <das33@cornell.edu>]
3393 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3394 RAND_egd() and RAND_status(). In the command line application,
3395 the EGD socket can be specified like a seed file using RANDFILE
3399 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3400 Some CAs (e.g. Verisign) distribute certificates in this form.
3403 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3404 list to exclude them. This means that no special compilation option
3405 is needed to use anonymous DH: it just needs to be included in the
3409 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3410 EVP_MD_type. The old functionality is available in a new macro called
3411 EVP_MD_md(). Change code that uses it and update docs.
3414 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3415 where the 'void *' argument is replaced by a function pointer argument.
3416 Previously 'void *' was abused to point to functions, which works on
3417 many platforms, but is not correct. As these functions are usually
3418 called by macros defined in OpenSSL header files, most source code
3419 should work without changes.
3422 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3423 sections with information on -D... compiler switches used for
3424 compiling the library so that applications can see them. To enable
3425 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3426 must be defined. E.g.,
3427 #define OPENSSL_ALGORITHM_DEFINES
3428 #include <openssl/opensslconf.h>
3429 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3430 [Richard Levitte, Ulf and Bodo Möller]
3432 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3436 *) Change the 'other' type in certificate aux info to a STACK_OF
3437 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3438 the required ASN1 format: arbitrary types determined by an OID.
3441 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3442 argument to 'req'. This is not because the function is newer or
3443 better than others it just uses the work 'NEW' in the certificate
3444 request header lines. Some software needs this.
3447 *) Reorganise password command line arguments: now passwords can be
3448 obtained from various sources. Delete the PEM_cb function and make
3449 it the default behaviour: i.e. if the callback is NULL and the
3450 usrdata argument is not NULL interpret it as a null terminated pass
3451 phrase. If usrdata and the callback are NULL then the pass phrase
3452 is prompted for as usual.
3455 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3456 the support is automatically enabled. The resulting binaries will
3457 autodetect the card and use it if present.
3458 [Ben Laurie and Compaq Inc.]
3460 *) Work around for Netscape hang bug. This sends certificate request
3461 and server done in one record. Since this is perfectly legal in the
3462 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3463 the bugs/SSLv3 entry for more info.
3466 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3469 *) Add -rand argument to smime and pkcs12 applications and read/write
3473 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3476 *) Add command line password options to the remaining applications.
3479 *) Bug fix for BN_div_recp() for numerators with an even number of
3483 *) More tests in bntest.c, and changed test_bn output.
3486 *) ./config recognizes MacOS X now.
3489 *) Bug fix for BN_div() when the first words of num and divsor are
3490 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3493 *) Add support for various broken PKCS#8 formats, and command line
3494 options to produce them.
3497 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3498 get temporary BIGNUMs from a BN_CTX.
3501 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3505 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3506 include a #define from the old name to the new. The original intent
3507 was that statically linked binaries could for example just call
3508 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3509 link with digests. This never worked becayse SSLeay_add_all_digests()
3510 and SSLeay_add_all_ciphers() were in the same source file so calling
3511 one would link with the other. They are now in separate source files.
3514 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3517 *) Use a less unusual form of the Miller-Rabin primality test (it used
3518 a binary algorithm for exponentiation integrated into the Miller-Rabin
3519 loop, our standard modexp algorithms are faster).
3522 *) Support for the EBCDIC character set completed.
3523 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3525 *) Source code cleanups: use const where appropriate, eliminate casts,
3526 use void * instead of char * in lhash.
3529 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3530 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3531 this the server could overwrite ephemeral keys that the client
3535 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3536 using 50 iterations of the Rabin-Miller test.
3538 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3539 iterations of the Rabin-Miller test as required by the appendix
3540 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3541 As BN_is_prime_fasttest includes trial division, DSA parameter
3542 generation becomes much faster.
3544 This implies a change for the callback functions in DSA_is_prime
3545 and DSA_generate_parameters: The callback function is called once
3546 for each positive witness in the Rabin-Miller test, not just
3547 occasionally in the inner loop; and the parameters to the
3548 callback function now provide an iteration count for the outer
3549 loop rather than for the current invocation of the inner loop.
3550 DSA_generate_parameters additionally can call the callback
3551 function with an 'iteration count' of -1, meaning that a
3552 candidate has passed the trial division test (when q is generated
3553 from an application-provided seed, trial division is skipped).
3556 *) New function BN_is_prime_fasttest that optionally does trial
3557 division before starting the Rabin-Miller test and has
3558 an additional BN_CTX * argument (whereas BN_is_prime always
3559 has to allocate at least one BN_CTX).
3560 'callback(1, -1, cb_arg)' is called when a number has passed the
3561 trial division stage.
3564 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3568 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3571 *) New function BN_pseudo_rand().
3574 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3575 bignum version of BN_from_montgomery() with the working code from
3576 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3580 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3581 made it impossible to use the same SSL_SESSION data structure in
3582 SSL2 clients in multiple threads.
3585 *) The return value of RAND_load_file() no longer counts bytes obtained
3586 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3587 to seed the PRNG (previously an explicit byte count was required).
3588 [Ulf Möller, Bodo Möller]
3590 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3591 used (char *) instead of (void *) and had casts all over the place.
3594 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3597 *) Retain source code compatibility for BN_prime_checks macro:
3598 BN_is_prime(..., BN_prime_checks, ...) now uses
3599 BN_prime_checks_for_size to determine the appropriate number of
3600 Rabin-Miller iterations.
3603 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3604 DH_CHECK_P_NOT_SAFE_PRIME.
3605 (Check if this is true? OpenPGP calls them "strong".)
3608 *) Merge the functionality of "dh" and "gendh" programs into a new program
3609 "dhparam". The old programs are retained for now but will handle DH keys
3610 (instead of parameters) in future.
3613 *) Make the ciphers, s_server and s_client programs check the return values
3614 when a new cipher list is set.
3617 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3618 ciphers. Before when the 56bit ciphers were enabled the sorting was
3621 The syntax for the cipher sorting has been extended to support sorting by
3622 cipher-strength (using the strength_bits hard coded in the tables).
3623 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3625 Fix a bug in the cipher-command parser: when supplying a cipher command
3626 string with an "undefined" symbol (neither command nor alphanumeric
3627 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3628 an error is flagged.
3630 Due to the strength-sorting extension, the code of the
3631 ssl_create_cipher_list() function was completely rearranged. I hope that
3632 the readability was also increased :-)
3633 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3635 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3636 for the first serial number and places 2 in the serial number file. This
3637 avoids problems when the root CA is created with serial number zero and
3638 the first user certificate has the same issuer name and serial number
3642 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3643 the new code. Add documentation for this stuff.
3646 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3647 X509_*() to X509at_*() on the grounds that they don't handle X509
3648 structures and behave in an analagous way to the X509v3 functions:
3649 they shouldn't be called directly but wrapper functions should be used
3652 So we also now have some wrapper functions that call the X509at functions
3653 when passed certificate requests. (TO DO: similar things can be done with
3654 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3655 things. Some of these need some d2i or i2d and print functionality
3656 because they handle more complex structures.)
3659 *) Add missing #ifndefs that caused missing symbols when building libssl
3660 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3661 NO_RSA in ssl/s2*.c.
3662 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3664 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3665 has a return value which indicates the quality of the random data
3666 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3667 error queue. New function RAND_pseudo_bytes() generates output that is
3668 guaranteed to be unique but not unpredictable. RAND_add is like
3669 RAND_seed, but takes an extra argument for an entropy estimate
3670 (RAND_seed always assumes full entropy).
3673 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3674 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3675 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3676 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3677 false-positive rate of at most 2^-80 for random input.
3680 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3683 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3684 in the 0.9.5 release), this returns the chain
3685 from an X509_CTX structure with a dup of the stack and all
3686 the X509 reference counts upped: so the stack will exist
3687 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3690 Also make SSL_SESSION_print() print out the verify return
3694 *) Add manpage for the pkcs12 command. Also change the default
3695 behaviour so MAC iteration counts are used unless the new
3696 -nomaciter option is used. This improves file security and
3697 only older versions of MSIE (4.0 for example) need it.
3700 *) Honor the no-xxx Configure options when creating .DEF files.
3703 *) Add PKCS#10 attributes to field table: challengePassword,
3704 unstructuredName and unstructuredAddress. These are taken from
3705 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3706 international characters are used.
3708 More changes to X509_ATTRIBUTE code: allow the setting of types
3709 based on strings. Remove the 'loc' parameter when adding
3710 attributes because these will be a SET OF encoding which is sorted
3714 *) Initial changes to the 'req' utility to allow request generation
3715 automation. This will allow an application to just generate a template
3716 file containing all the field values and have req construct the
3719 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3720 used all over the place including certificate requests and PKCS#7
3721 structures. They are currently handled manually where necessary with
3722 some primitive wrappers for PKCS#7. The new functions behave in a
3723 manner analogous to the X509 extension functions: they allow
3724 attributes to be looked up by NID and added.
3726 Later something similar to the X509V3 code would be desirable to
3727 automatically handle the encoding, decoding and printing of the
3728 more complex types. The string types like challengePassword can
3729 be handled by the string table functions.
3731 Also modified the multi byte string table handling. Now there is
3732 a 'global mask' which masks out certain types. The table itself
3733 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3734 is useful when for example there is only one permissible type
3735 (as in countryName) and using the mask might result in no valid
3739 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3740 SSL_get_peer_finished to allow applications to obtain the latest
3741 Finished messages sent to the peer or expected from the peer,
3742 respectively. (SSL_get_peer_finished is usually the Finished message
3743 actually received from the peer, otherwise the protocol will be aborted.)
3745 As the Finished message are message digests of the complete handshake
3746 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3747 be used for external authentication procedures when the authentication
3748 provided by SSL/TLS is not desired or is not enough.
3751 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3752 the host supports BWX extension and if Compaq C is present on the
3753 $PATH. Just exploiting of the BWX extension results in 20-30%
3754 performance kick for some algorithms, e.g. DES and RC4 to mention
3755 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3759 *) Add support for MS "fast SGC". This is arguably a violation of the
3760 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3761 weak crypto and after checking the certificate is SGC a second one
3762 with strong crypto. MS SGC stops the first handshake after receiving
3763 the server certificate message and sends a second client hello. Since
3764 a server will typically do all the time consuming operations before
3765 expecting any further messages from the client (server key exchange
3766 is the most expensive) there is little difference between the two.
3768 To get OpenSSL to support MS SGC we have to permit a second client
3769 hello message after we have sent server done. In addition we have to
3770 reset the MAC if we do get this second client hello.
3773 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3774 if a DER encoded private key is RSA or DSA traditional format. Changed
3775 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3776 format DER encoded private key. Newer code should use PKCS#8 format which
3777 has the key type encoded in the ASN1 structure. Added DER private key
3778 support to pkcs8 application.
3781 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3782 ciphersuites has been selected (as required by the SSL 3/TLS 1
3783 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3784 is set, we interpret this as a request to violate the specification
3785 (the worst that can happen is a handshake failure, and 'correct'
3786 behaviour would result in a handshake failure anyway).
3789 *) In SSL_CTX_add_session, take into account that there might be multiple
3790 SSL_SESSION structures with the same session ID (e.g. when two threads
3791 concurrently obtain them from an external cache).
3792 The internal cache can handle only one SSL_SESSION with a given ID,
3793 so if there's a conflict, we now throw out the old one to achieve
3797 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3798 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3799 some routines that use cipher OIDs: some ciphers do not have OIDs
3800 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3804 *) Simplify the trust setting structure and code. Now we just have
3805 two sequences of OIDs for trusted and rejected settings. These will
3806 typically have values the same as the extended key usage extension
3807 and any application specific purposes.
3809 The trust checking code now has a default behaviour: it will just
3810 check for an object with the same NID as the passed id. Functions can
3811 be provided to override either the default behaviour or the behaviour
3812 for a given id. SSL client, server and email already have functions
3813 in place for compatibility: they check the NID and also return "trusted"
3814 if the certificate is self signed.
3817 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3818 traditional format into an EVP_PKEY structure.
3821 *) Add a password callback function PEM_cb() which either prompts for
3822 a password if usr_data is NULL or otherwise assumes it is a null
3823 terminated password. Allow passwords to be passed on command line
3824 environment or config files in a few more utilities.
3827 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3828 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3829 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3830 Update documentation.
3833 *) Support for ASN1 "NULL" type. This could be handled before by using
3834 ASN1_TYPE but there wasn't any function that would try to read a NULL
3835 and produce an error if it couldn't. For compatibility we also have
3836 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3837 don't allocate anything because they don't need to.
3840 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3842 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3844 *) Rebuild of the memory allocation routines used by OpenSSL code and
3845 possibly others as well. The purpose is to make an interface that
3846 provide hooks so anyone can build a separate set of allocation and
3847 deallocation routines to be used by OpenSSL, for example memory
3848 pool implementations, or something else, which was previously hard
3849 since Malloc(), Realloc() and Free() were defined as macros having
3850 the values malloc, realloc and free, respectively (except for Win32
3851 compilations). The same is provided for memory debugging code.
3852 OpenSSL already comes with functionality to find memory leaks, but
3853 this gives people a chance to debug other memory problems.
3855 With these changes, a new set of functions and macros have appeared:
3857 CRYPTO_set_mem_debug_functions() [F]
3858 CRYPTO_get_mem_debug_functions() [F]
3859 CRYPTO_dbg_set_options() [F]
3860 CRYPTO_dbg_get_options() [F]
3861 CRYPTO_malloc_debug_init() [M]
3863 The memory debug functions are NULL by default, unless the library
3864 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3865 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3866 gives the standard debugging functions that come with OpenSSL) or
3867 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3868 provided by the library user) must be used. When the standard
3869 debugging functions are used, CRYPTO_dbg_set_options can be used to
3870 request additional information:
3871 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3872 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3874 Also, things like CRYPTO_set_mem_functions will always give the
3875 expected result (the new set of functions is used for allocation
3876 and deallocation) at all times, regardless of platform and compiler
3879 To finish it up, some functions that were never use in any other
3880 way than through macros have a new API and new semantic:
3883 CRYPTO_dbg_realloc()
3886 All macros of value have retained their old syntax.
3887 [Richard Levitte and Bodo Moeller]
3889 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3890 ordering of SMIMECapabilities wasn't in "strength order" and there
3891 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3895 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3896 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3897 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3899 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3900 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3901 functionality to handle multipart/signed properly) and a utility
3902 called 'smime' to call all this stuff. This is based on code I
3903 originally wrote for Celo who have kindly allowed it to be
3904 included in OpenSSL.
3907 *) Add variants des_set_key_checked and des_set_key_unchecked of
3908 des_set_key (aka des_key_sched). Global variable des_check_key
3909 decides which of these is called by des_set_key; this way
3910 des_check_key behaves as it always did, but applications and
3911 the library itself, which was buggy for des_check_key == 1,
3912 have a cleaner way to pick the version they need.
3915 *) New function PKCS12_newpass() which changes the password of a
3919 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3920 dynamic mix. In both cases the ids can be used as an index into the
3921 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3922 functions so they accept a list of the field values and the
3923 application doesn't need to directly manipulate the X509_TRUST
3927 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3931 *) Modify the way the V3 extension code looks up extensions. This now
3932 works in a similar way to the object code: we have some "standard"
3933 extensions in a static table which is searched with OBJ_bsearch()
3934 and the application can add dynamic ones if needed. The file
3935 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3936 updated whenever a new extension is added to the core code and kept
3937 in ext_nid order. There is a simple program 'tabtest.c' which checks
3938 this. New extensions are not added too often so this file can readily
3939 be maintained manually.
3941 There are two big advantages in doing things this way. The extensions
3942 can be looked up immediately and no longer need to be "added" using
3943 X509V3_add_standard_extensions(): this function now does nothing.
3944 [Side note: I get *lots* of email saying the extension code doesn't
3945 work because people forget to call this function]
3946 Also no dynamic allocation is done unless new extensions are added:
3947 so if we don't add custom extensions there is no need to call
3948 X509V3_EXT_cleanup().
3951 *) Modify enc utility's salting as follows: make salting the default. Add a
3952 magic header, so unsalted files fail gracefully instead of just decrypting
3953 to garbage. This is because not salting is a big security hole, so people
3954 should be discouraged from doing it.
3957 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3958 digest to be passed on the command line but it only used this
3959 parameter when signing a certificate. Modified so all relevant
3960 operations are affected by the digest parameter including the
3961 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3962 DSA key was used because it didn't fix the digest.
3965 *) Initial certificate chain verify code. Currently tests the untrusted
3966 certificates for consistency with the verify purpose (which is set
3967 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3969 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3970 this is because it will reject chains with invalid extensions whereas
3971 every previous version of OpenSSL and SSLeay made no checks at all.
3973 Trust code: checks the root CA for the relevant trust settings. Trust
3974 settings have an initial value consistent with the verify purpose: e.g.
3975 if the verify purpose is for SSL client use it expects the CA to be
3976 trusted for SSL client use. However the default value can be changed to
3977 permit custom trust settings: one example of this would be to only trust
3978 certificates from a specific "secure" set of CAs.
3980 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3981 which should be used for version portability: especially since the
3982 verify structure is likely to change more often now.
3984 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3985 to set them. If not set then assume SSL clients will verify SSL servers
3988 Two new options to the verify program: -untrusted allows a set of
3989 untrusted certificates to be passed in and -purpose which sets the
3990 intended purpose of the certificate. If a purpose is set then the
3991 new chain verify code is used to check extension consistency.
3994 *) Support for the authority information access extension.
3997 *) Modify RSA and DSA PEM read routines to transparently handle
3998 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3999 public keys in a format compatible with certificate
4000 SubjectPublicKeyInfo structures. Unfortunately there were already
4001 functions called *_PublicKey_* which used various odd formats so
4002 these are retained for compatibility: however the DSA variants were
4003 never in a public release so they have been deleted. Changed dsa/rsa
4004 utilities to handle the new format: note no releases ever handled public
4005 keys so we should be OK.
4007 The primary motivation for this change is to avoid the same fiasco
4008 that dogs private keys: there are several incompatible private key
4009 formats some of which are standard and some OpenSSL specific and
4010 require various evil hacks to allow partial transparent handling and
4011 even then it doesn't work with DER formats. Given the option anything
4012 other than PKCS#8 should be dumped: but the other formats have to
4013 stay in the name of compatibility.
4015 With public keys and the benefit of hindsight one standard format
4016 is used which works with EVP_PKEY, RSA or DSA structures: though
4017 it clearly returns an error if you try to read the wrong kind of key.
4019 Added a -pubkey option to the 'x509' utility to output the public key.
4020 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4021 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4022 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4023 that do the same as the EVP_PKEY_assign_*() except they up the
4024 reference count of the added key (they don't "swallow" the
4028 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4029 CRLs would fail if the file contained no certificates or no CRLs:
4030 added a new function to read in both types and return the number
4031 read: this means that if none are read it will be an error. The
4032 DER versions of the certificate and CRL reader would always fail
4033 because it isn't possible to mix certificates and CRLs in DER format
4034 without choking one or the other routine. Changed this to just read
4035 a certificate: this is the best we can do. Also modified the code
4036 in apps/verify.c to take notice of return codes: it was previously
4037 attempting to read in certificates from NULL pointers and ignoring
4038 any errors: this is one reason why the cert and CRL reader seemed
4039 to work. It doesn't check return codes from the default certificate
4040 routines: these may well fail if the certificates aren't installed.
4043 *) Code to support otherName option in GeneralName.
4046 *) First update to verify code. Change the verify utility
4047 so it warns if it is passed a self signed certificate:
4048 for consistency with the normal behaviour. X509_verify
4049 has been modified to it will now verify a self signed
4050 certificate if *exactly* the same certificate appears
4051 in the store: it was previously impossible to trust a
4052 single self signed certificate. This means that:
4053 openssl verify ss.pem
4054 now gives a warning about a self signed certificate but
4055 openssl verify -CAfile ss.pem ss.pem
4059 *) For servers, store verify_result in SSL_SESSION data structure
4060 (and add it to external session representation).
4061 This is needed when client certificate verifications fails,
4062 but an application-provided verification callback (set by
4063 SSL_CTX_set_cert_verify_callback) allows accepting the session
4064 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4065 but returns 1): When the session is reused, we have to set
4066 ssl->verify_result to the appropriate error code to avoid
4068 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4070 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4071 case in PKCS7_dataInit() where the signed PKCS7 structure
4072 didn't contain any existing data because it was being created.
4073 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4075 *) Add a salt to the key derivation routines in enc.c. This
4076 forms the first 8 bytes of the encrypted file. Also add a
4077 -S option to allow a salt to be input on the command line.
4080 *) New function X509_cmp(). Oddly enough there wasn't a function
4081 to compare two certificates. We do this by working out the SHA1
4082 hash and comparing that. X509_cmp() will be needed by the trust
4086 *) SSL_get1_session() is like SSL_get_session(), but increments
4087 the reference count in the SSL_SESSION returned.
4088 [Geoff Thorpe <geoff@eu.c2.net>]
4090 *) Fix for 'req': it was adding a null to request attributes.
4091 Also change the X509_LOOKUP and X509_INFO code to handle
4092 certificate auxiliary information.
4095 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4099 *) Add the possibility to add extra information to the memory leak
4100 detecting output, to form tracebacks, showing from where each
4101 allocation was originated: CRYPTO_push_info("constant string") adds
4102 the string plus current file name and line number to a per-thread
4103 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4104 is like calling CYRPTO_pop_info() until the stack is empty.
4105 Also updated memory leak detection code to be multi-thread-safe.
4108 *) Add options -text and -noout to pkcs7 utility and delete the
4109 encryption options which never did anything. Update docs.
4112 *) Add options to some of the utilities to allow the pass phrase
4113 to be included on either the command line (not recommended on
4114 OSes like Unix) or read from the environment. Update the
4115 manpages and fix a few bugs.
4118 *) Add a few manpages for some of the openssl commands.
4121 *) Fix the -revoke option in ca. It was freeing up memory twice,
4122 leaking and not finding already revoked certificates.
4125 *) Extensive changes to support certificate auxiliary information.
4126 This involves the use of X509_CERT_AUX structure and X509_AUX
4127 functions. An X509_AUX function such as PEM_read_X509_AUX()
4128 can still read in a certificate file in the usual way but it
4129 will also read in any additional "auxiliary information". By
4130 doing things this way a fair degree of compatibility can be
4131 retained: existing certificates can have this information added
4132 using the new 'x509' options.
4134 Current auxiliary information includes an "alias" and some trust
4135 settings. The trust settings will ultimately be used in enhanced
4136 certificate chain verification routines: currently a certificate
4137 can only be trusted if it is self signed and then it is trusted
4141 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4142 The problem was that one of the replacement routines had not been working
4143 since SSLeay releases. For now the offending routine has been replaced
4144 with non-optimised assembler. Even so, this now gives around 95%
4145 performance improvement for 1024 bit RSA signs.
4148 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4149 handling. Most clients have the effective key size in bits equal to
4150 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4151 A few however don't do this and instead use the size of the decrypted key
4152 to determine the RC2 key length and the AlgorithmIdentifier to determine
4153 the effective key length. In this case the effective key length can still
4154 be 40 bits but the key length can be 168 bits for example. This is fixed
4155 by manually forcing an RC2 key into the EVP_PKEY structure because the
4156 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4157 the key length and effective key length are equal.
4160 *) Add a bunch of functions that should simplify the creation of
4161 X509_NAME structures. Now you should be able to do:
4162 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4163 and have it automatically work out the correct field type and fill in
4164 the structures. The more adventurous can try:
4165 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4166 and it will (hopefully) work out the correct multibyte encoding.
4169 *) Change the 'req' utility to use the new field handling and multibyte
4170 copy routines. Before the DN field creation was handled in an ad hoc
4171 way in req, ca, and x509 which was rather broken and didn't support
4172 BMPStrings or UTF8Strings. Since some software doesn't implement
4173 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4174 using the dirstring_type option. See the new comment in the default
4175 openssl.cnf for more info.
4178 *) Make crypto/rand/md_rand.c more robust:
4179 - Assure unique random numbers after fork().
4180 - Make sure that concurrent threads access the global counter and
4181 md serializably so that we never lose entropy in them
4182 or use exactly the same state in multiple threads.
4183 Access to the large state is not always serializable because
4184 the additional locking could be a performance killer, and
4185 md should be large enough anyway.
4188 *) New file apps/app_rand.c with commonly needed functionality
4189 for handling the random seed file.
4191 Use the random seed file in some applications that previously did not:
4193 dsaparam -genkey (which also ignored its '-rand' option),
4196 x509 (when signing).
4197 Except on systems with /dev/urandom, it is crucial to have a random
4198 seed file at least for key creation, DSA signing, and for DH exchanges;
4199 for RSA signatures we could do without one.
4201 gendh and gendsa (unlike genrsa) used to read only the first byte
4202 of each file listed in the '-rand' option. The function as previously
4203 found in genrsa is now in app_rand.c and is used by all programs
4204 that support '-rand'.
4207 *) In RAND_write_file, use mode 0600 for creating files;
4208 don't just chmod when it may be too late.
4211 *) Report an error from X509_STORE_load_locations
4212 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4215 *) New function ASN1_mbstring_copy() this copies a string in either
4216 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4217 into an ASN1_STRING type. A mask of permissible types is passed
4218 and it chooses the "minimal" type to use or an error if not type
4222 *) Add function equivalents to the various macros in asn1.h. The old
4223 macros are retained with an M_ prefix. Code inside the library can
4224 use the M_ macros. External code (including the openssl utility)
4225 should *NOT* in order to be "shared library friendly".
4228 *) Add various functions that can check a certificate's extensions
4229 to see if it usable for various purposes such as SSL client,
4230 server or S/MIME and CAs of these types. This is currently
4231 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4232 verification. Also added a -purpose flag to x509 utility to
4233 print out all the purposes.
4236 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4240 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4241 for, obtain and decode and extension and obtain its critical flag.
4242 This allows all the necessary extension code to be handled in a
4243 single function call.
4246 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4247 platforms. See crypto/rc4/rc4_enc.c for further details.
4250 *) New -noout option to asn1parse. This causes no output to be produced
4251 its main use is when combined with -strparse and -out to extract data
4252 from a file (which may not be in ASN.1 format).
4255 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4256 when producing the local key id.
4257 [Richard Levitte <levitte@stacken.kth.se>]
4259 *) New option -dhparam in s_server. This allows a DH parameter file to be
4260 stated explicitly. If it is not stated then it tries the first server
4261 certificate file. The previous behaviour hard coded the filename
4265 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4266 a public key to be input or output. For example:
4267 openssl rsa -in key.pem -pubout -out pubkey.pem
4268 Also added necessary DSA public key functions to handle this.
4271 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4272 in the message. This was handled by allowing
4273 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4274 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4276 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4277 to the end of the strings whereas this didn't. This would cause problems
4278 if strings read with d2i_ASN1_bytes() were later modified.
4279 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4281 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4282 data and it contains EOF it will end up returning an error. This is
4283 caused by input 46 bytes long. The cause is due to the way base64
4284 BIOs find the start of base64 encoded data. They do this by trying a
4285 trial decode on each line until they find one that works. When they
4286 do a flag is set and it starts again knowing it can pass all the
4287 data directly through the decoder. Unfortunately it doesn't reset
4288 the context it uses. This means that if EOF is reached an attempt
4289 is made to pass two EOFs through the context and this causes the
4290 resulting error. This can also cause other problems as well. As is
4291 usual with these problems it takes *ages* to find and the fix is
4292 trivial: move one line.
4293 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4295 *) Ugly workaround to get s_client and s_server working under Windows. The
4296 old code wouldn't work because it needed to select() on sockets and the
4297 tty (for keypresses and to see if data could be written). Win32 only
4298 supports select() on sockets so we select() with a 1s timeout on the
4299 sockets and then see if any characters are waiting to be read, if none
4300 are present then we retry, we also assume we can always write data to
4301 the tty. This isn't nice because the code then blocks until we've
4302 received a complete line of data and it is effectively polling the
4303 keyboard at 1s intervals: however it's quite a bit better than not
4304 working at all :-) A dedicated Windows application might handle this
4305 with an event loop for example.
4308 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4309 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4310 will be called when RSA_sign() and RSA_verify() are used. This is useful
4311 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4312 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4313 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4314 This necessitated the support of an extra signature type NID_md5_sha1
4315 for SSL signatures and modifications to the SSL library to use it instead
4316 of calling RSA_public_decrypt() and RSA_private_encrypt().
4319 *) Add new -verify -CAfile and -CApath options to the crl program, these
4320 will lookup a CRL issuers certificate and verify the signature in a
4321 similar way to the verify program. Tidy up the crl program so it
4322 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4323 less strict. It will now permit CRL extensions even if it is not
4324 a V2 CRL: this will allow it to tolerate some broken CRLs.
4327 *) Initialize all non-automatic variables each time one of the openssl
4328 sub-programs is started (this is necessary as they may be started
4329 multiple times from the "OpenSSL>" prompt).
4330 [Lennart Bang, Bodo Moeller]
4332 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4333 removing all other RSA functionality (this is what NO_RSA does). This
4334 is so (for example) those in the US can disable those operations covered
4335 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4339 *) Non-copying interface to BIO pairs.
4340 (still largely untested)
4343 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4344 ASCII string. This was handled independently in various places before.
4347 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4348 UTF8 strings a character at a time.
4351 *) Use client_version from client hello to select the protocol
4352 (s23_srvr.c) and for RSA client key exchange verification
4353 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4356 *) Add various utility functions to handle SPKACs, these were previously
4357 handled by poking round in the structure internals. Added new function
4358 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4359 print, verify and generate SPKACs. Based on an original idea from
4360 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4363 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4366 *) Allow the config file extension section to be overwritten on the
4367 command line. Based on an original idea from Massimiliano Pala
4368 <madwolf@comune.modena.it>. The new option is called -extensions
4369 and can be applied to ca, req and x509. Also -reqexts to override
4370 the request extensions in req and -crlexts to override the crl extensions
4374 *) Add new feature to the SPKAC handling in ca. Now you can include
4375 the same field multiple times by preceding it by "XXXX." for example:
4378 this is the same syntax as used in the req config file.
4381 *) Allow certificate extensions to be added to certificate requests. These
4382 are specified in a 'req_extensions' option of the req section of the
4383 config file. They can be printed out with the -text option to req but
4384 are otherwise ignored at present.
4387 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4388 data read consists of only the final block it would not decrypted because
4389 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4390 A misplaced 'break' also meant the decrypted final block might not be
4391 copied until the next read.
4394 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4395 a few extra parameters to the DH structure: these will be useful if
4396 for example we want the value of 'q' or implement X9.42 DH.
4399 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4400 provides hooks that allow the default DSA functions or functions on a
4401 "per key" basis to be replaced. This allows hardware acceleration and
4402 hardware key storage to be handled without major modification to the
4403 library. Also added low level modexp hooks and CRYPTO_EX structure and
4404 associated functions.
4407 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4408 as "read only": it can't be written to and the buffer it points to will
4409 not be freed. Reading from a read only BIO is much more efficient than
4410 a normal memory BIO. This was added because there are several times when
4411 an area of memory needs to be read from a BIO. The previous method was
4412 to create a memory BIO and write the data to it, this results in two
4413 copies of the data and an O(n^2) reading algorithm. There is a new
4414 function BIO_new_mem_buf() which creates a read only memory BIO from
4415 an area of memory. Also modified the PKCS#7 routines to use read only
4419 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4420 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4421 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4422 but a retry condition occured while trying to read the rest.
4425 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4426 NID_pkcs7_encrypted by default: this was wrong since this should almost
4427 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4428 the encrypted data type: this is a more sensible place to put it and it
4429 allows the PKCS#12 code to be tidied up that duplicated this
4433 *) Changed obj_dat.pl script so it takes its input and output files on
4434 the command line. This should avoid shell escape redirection problems
4438 *) Initial support for certificate extension requests, these are included
4439 in things like Xenroll certificate requests. Included functions to allow
4440 extensions to be obtained and added.
4443 *) -crlf option to s_client and s_server for sending newlines as
4444 CRLF (as required by many protocols).
4447 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4449 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4450 [Ralf S. Engelschall]
4452 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4453 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4455 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4459 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4460 DH parameters/keys (q is lost during that conversion, but the resulting
4461 DH parameters contain its length).
4463 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4464 much faster than DH_generate_parameters (which creates parameters
4465 where p = 2*q + 1), and also the smaller q makes DH computations
4466 much more efficient (160-bit exponentiation instead of 1024-bit
4467 exponentiation); so this provides a convenient way to support DHE
4468 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4469 utter importance to use
4470 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4472 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4473 when such DH parameters are used, because otherwise small subgroup
4474 attacks may become possible!
4477 *) Avoid memory leak in i2d_DHparams.
4480 *) Allow the -k option to be used more than once in the enc program:
4481 this allows the same encrypted message to be read by multiple recipients.
4484 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4485 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4486 it will always use the numerical form of the OID, even if it has a short
4490 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4491 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4492 otherwise bn_mod_exp was called. In the case of hardware keys for example
4493 no private key components need be present and it might store extra data
4494 in the RSA structure, which cannot be accessed from bn_mod_exp.
4495 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4496 private key operations.
4499 *) Added support for SPARC Linux.
4502 *) pem_password_cb function type incompatibly changed from
4503 typedef int pem_password_cb(char *buf, int size, int rwflag);
4505 ....(char *buf, int size, int rwflag, void *userdata);
4506 so that applications can pass data to their callbacks:
4507 The PEM[_ASN1]_{read,write}... functions and macros now take an
4508 additional void * argument, which is just handed through whenever
4509 the password callback is called.
4510 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4512 New function SSL_CTX_set_default_passwd_cb_userdata.
4514 Compatibility note: As many C implementations push function arguments
4515 onto the stack in reverse order, the new library version is likely to
4516 interoperate with programs that have been compiled with the old
4517 pem_password_cb definition (PEM_whatever takes some data that
4518 happens to be on the stack as its last argument, and the callback
4519 just ignores this garbage); but there is no guarantee whatsoever that
4522 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4523 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4524 problems not only on Windows, but also on some Unix platforms.
4525 To avoid problematic command lines, these definitions are now in an
4526 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4527 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4530 *) MIPS III/IV assembler module is reimplemented.
4533 *) More DES library cleanups: remove references to srand/rand and
4534 delete an unused file.
4537 *) Add support for the the free Netwide assembler (NASM) under Win32,
4538 since not many people have MASM (ml) and it can be hard to obtain.
4539 This is currently experimental but it seems to work OK and pass all
4540 the tests. Check out INSTALL.W32 for info.
4543 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4544 without temporary keys kept an extra copy of the server key,
4545 and connections with temporary keys did not free everything in case
4549 *) New function RSA_check_key and new openssl rsa option -check
4550 for verifying the consistency of RSA keys.
4551 [Ulf Moeller, Bodo Moeller]
4553 *) Various changes to make Win32 compile work:
4554 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4555 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4556 comparison" warnings.
4557 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4560 *) Add a debugging option to PKCS#5 v2 key generation function: when
4561 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4562 derived keys are printed to stderr.
4565 *) Copy the flags in ASN1_STRING_dup().
4566 [Roman E. Pavlov <pre@mo.msk.ru>]
4568 *) The x509 application mishandled signing requests containing DSA
4569 keys when the signing key was also DSA and the parameters didn't match.
4571 It was supposed to omit the parameters when they matched the signing key:
4572 the verifying software was then supposed to automatically use the CA's
4573 parameters if they were absent from the end user certificate.
4575 Omitting parameters is no longer recommended. The test was also
4576 the wrong way round! This was probably due to unusual behaviour in
4577 EVP_cmp_parameters() which returns 1 if the parameters match.
4578 This meant that parameters were omitted when they *didn't* match and
4579 the certificate was useless. Certificates signed with 'ca' didn't have
4581 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4583 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4584 The interface is as follows:
4585 Applications can use
4586 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4587 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4588 "off" is now the default.
4589 The library internally uses
4590 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4591 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4592 to disable memory-checking temporarily.
4594 Some inconsistent states that previously were possible (and were
4595 even the default) are now avoided.
4597 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4598 with each memory chunk allocated; this is occasionally more helpful
4599 than just having a counter.
4601 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4603 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4607 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4608 which largely parallels "options", but is for changing API behaviour,
4609 whereas "options" are about protocol behaviour.
4610 Initial "mode" flags are:
4612 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4613 a single record has been written.
4614 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4615 retries use the same buffer location.
4616 (But all of the contents must be
4620 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4623 *) Fix problems with no-hmac etc.
4624 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4626 *) New functions RSA_get_default_method(), RSA_set_method() and
4627 RSA_get_method(). These allows replacement of RSA_METHODs without having
4628 to mess around with the internals of an RSA structure.
4631 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4632 Also really enable memory leak checks in openssl.c and in some
4634 [Chad C. Mulligan, Bodo Moeller]
4636 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4637 up the length of negative integers. This has now been simplified to just
4638 store the length when it is first determined and use it later, rather
4639 than trying to keep track of where data is copied and updating it to
4641 [Steve Henson, reported by Brien Wheeler
4642 <bwheeler@authentica-security.com>]
4644 *) Add a new function PKCS7_signatureVerify. This allows the verification
4645 of a PKCS#7 signature but with the signing certificate passed to the
4646 function itself. This contrasts with PKCS7_dataVerify which assumes the
4647 certificate is present in the PKCS#7 structure. This isn't always the
4648 case: certificates can be omitted from a PKCS#7 structure and be
4649 distributed by "out of band" means (such as a certificate database).
4652 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4653 function prototypes in pem.h, also change util/mkdef.pl to add the
4654 necessary function names.
4657 *) mk1mf.pl (used by Windows builds) did not properly read the
4658 options set by Configure in the top level Makefile, and Configure
4659 was not even able to write more than one option correctly.
4660 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4663 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4664 file to be loaded from a BIO or FILE pointer. The BIO version will
4665 for example allow memory BIOs to contain config info.
4668 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4669 Whoever hopes to achieve shared-library compatibility across versions
4670 must use this, not the compile-time macro.
4671 (Exercise 0.9.4: Which is the minimum library version required by
4673 Note: All this applies only to multi-threaded programs, others don't
4677 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4678 through a BIO pair triggered the default case, i.e.
4679 SSLerr(...,SSL_R_UNKNOWN_STATE).
4682 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4683 can use the SSL library even if none of the specific BIOs is
4687 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4688 for the encoded length.
4689 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4691 *) Add initial documentation of the X509V3 functions.
4694 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4695 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4696 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4697 secure PKCS#8 private key format with a high iteration count.
4700 *) Fix determination of Perl interpreter: A perl or perl5
4701 _directory_ in $PATH was also accepted as the interpreter.
4702 [Ralf S. Engelschall]
4704 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4705 wrong with it but it was very old and did things like calling
4706 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4710 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4711 to use the new extension code.
4714 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4715 with macros. This should make it easier to change their form, add extra
4716 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4720 *) Add to configuration table a new entry that can specify an alternative
4721 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4722 according to Mark Crispin <MRC@Panda.COM>.
4726 *) DES CBC did not update the IV. Weird.
4729 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4730 Changing the behaviour of the former might break existing programs --
4731 where IV updating is needed, des_ncbc_encrypt can be used.
4734 *) When bntest is run from "make test" it drives bc to check its
4735 calculations, as well as internally checking them. If an internal check
4736 fails, it needs to cause bc to give a non-zero result or make test carries
4737 on without noticing the failure. Fixed.
4740 *) DES library cleanups.
4743 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4744 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4745 ciphers. NOTE: although the key derivation function has been verified
4746 against some published test vectors it has not been extensively tested
4747 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4751 *) Instead of "mkdir -p", which is not fully portable, use new
4752 Perl script "util/mkdir-p.pl".
4755 *) Rewrite the way password based encryption (PBE) is handled. It used to
4756 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4757 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4758 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4759 the 'parameter' field of the AlgorithmIdentifier is passed to the
4760 underlying key generation function so it must do its own ASN1 parsing.
4761 This has also changed the EVP_PBE_CipherInit() function which now has a
4762 'parameter' argument instead of literal salt and iteration count values
4763 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4766 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4767 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4768 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4769 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4770 value was just used as a "magic string" and not used directly its
4771 value doesn't matter.
4774 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4778 *) "linux-sparc64" configuration (ultrapenguin).
4779 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4780 "linux-sparc" configuration.
4781 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4783 *) config now generates no-xxx options for missing ciphers.
4786 *) Support the EBCDIC character set (work in progress).
4787 File ebcdic.c not yet included because it has a different license.
4788 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4790 *) Support BS2000/OSD-POSIX.
4791 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4793 *) Make callbacks for key generation use void * instead of char *.
4796 *) Make S/MIME samples compile (not yet tested).
4799 *) Additional typesafe stacks.
4802 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4806 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4808 *) New configuration variant "sco5-gcc".
4810 *) Updated some demos.
4811 [Sean O Riordain, Wade Scholine]
4813 *) Add missing BIO_free at exit of pkcs12 application.
4816 *) Fix memory leak in conf.c.
4819 *) Updates for Win32 to assembler version of MD5.
4822 *) Set #! path to perl in apps/der_chop to where we found it
4823 instead of using a fixed path.
4826 *) SHA library changes for irix64-mips4-cc.
4829 *) Improvements for VMS support.
4833 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4835 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4836 This also avoids the problems with SC4.2 and unpatched SC5.
4837 [Andy Polyakov <appro@fy.chalmers.se>]
4839 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4840 These are required because of the typesafe stack would otherwise break
4841 existing code. If old code used a structure member which used to be STACK
4842 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4843 sk_num or sk_value it would produce an error because the num, data members
4844 are not present in STACK_OF. Now it just produces a warning. sk_set
4845 replaces the old method of assigning a value to sk_value
4846 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4847 that does this will no longer work (and should use sk_set instead) but
4848 this could be regarded as a "questionable" behaviour anyway.
4851 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4852 correctly handle encrypted S/MIME data.
4855 *) Change type of various DES function arguments from des_cblock
4856 (which means, in function argument declarations, pointer to char)
4857 to des_cblock * (meaning pointer to array with 8 char elements),
4858 which allows the compiler to do more typechecking; it was like
4859 that back in SSLeay, but with lots of ugly casts.
4861 Introduce new type const_des_cblock.
4864 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4865 problems: find RecipientInfo structure that matches recipient certificate
4866 and initialise the ASN1 structures properly based on passed cipher.
4869 *) Belatedly make the BN tests actually check the results.
4872 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4873 to and from BNs: it was completely broken. New compilation option
4874 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4875 key elements as negative integers.
4878 *) Reorganize and speed up MD5.
4879 [Andy Polyakov <appro@fy.chalmers.se>]
4882 [Richard Levitte <richard@levitte.org>]
4884 *) New option -out to asn1parse to allow the parsed structure to be
4885 output to a file. This is most useful when combined with the -strparse
4886 option to examine the output of things like OCTET STRINGS.
4889 *) Make SSL library a little more fool-proof by not requiring any longer
4890 that SSL_set_{accept,connect}_state be called before
4891 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4892 in many applications because usually everything *appeared* to work as
4893 intended anyway -- now it really works as intended).
4896 *) Move openssl.cnf out of lib/.
4899 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4900 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4901 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4902 [Ralf S. Engelschall]
4904 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4905 handle PKCS#7 enveloped data properly.
4906 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4908 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4909 copying pointers. The cert_st handling is changed by this in
4910 various ways (and thus what used to be known as ctx->default_cert
4911 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4912 any longer when s->cert does not give us what we need).
4913 ssl_cert_instantiate becomes obsolete by this change.
4914 As soon as we've got the new code right (possibly it already is?),
4915 we have solved a couple of bugs of the earlier code where s->cert
4916 was used as if it could not have been shared with other SSL structures.
4918 Note that using the SSL API in certain dirty ways now will result
4919 in different behaviour than observed with earlier library versions:
4920 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4921 does not influence s as it used to.
4923 In order to clean up things more thoroughly, inside SSL_SESSION
4924 we don't use CERT any longer, but a new structure SESS_CERT
4925 that holds per-session data (if available); currently, this is
4926 the peer's certificate chain and, for clients, the server's certificate
4927 and temporary key. CERT holds only those values that can have
4928 meaningful defaults in an SSL_CTX.
4931 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4932 from the internal representation. Various PKCS#7 fixes: remove some
4933 evil casts and set the enc_dig_alg field properly based on the signing
4937 *) Allow PKCS#12 password to be set from the command line or the
4938 environment. Let 'ca' get its config file name from the environment
4939 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4943 *) Allow certificate policies extension to use an IA5STRING for the
4944 organization field. This is contrary to the PKIX definition but
4945 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4949 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4950 without disallowing inline assembler and the like for non-pedantic builds.
4953 *) Support Borland C++ builder.
4954 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4959 *) SHA-1 cleanups and performance enhancements.
4960 [Andy Polyakov <appro@fy.chalmers.se>]
4962 *) Sparc v8plus assembler for the bignum library.
4963 [Andy Polyakov <appro@fy.chalmers.se>]
4965 *) Accept any -xxx and +xxx compiler options in Configure.
4968 *) Update HPUX configuration.
4971 *) Add missing sk_<type>_unshift() function to safestack.h
4972 [Ralf S. Engelschall]
4974 *) New function SSL_CTX_use_certificate_chain_file that sets the
4975 "extra_cert"s in addition to the certificate. (This makes sense
4976 only for "PEM" format files, as chains as a whole are not
4980 *) Support verify_depth from the SSL API.
4981 x509_vfy.c had what can be considered an off-by-one-error:
4982 Its depth (which was not part of the external interface)
4983 was actually counting the number of certificates in a chain;
4984 now it really counts the depth.
4987 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4988 instead of X509err, which often resulted in confusing error
4989 messages since the error codes are not globally unique
4990 (e.g. an alleged error in ssl3_accept when a certificate
4991 didn't match the private key).
4993 *) New function SSL_CTX_set_session_id_context that allows to set a default
4994 value (so that you don't need SSL_set_session_id_context for each
4995 connection using the SSL_CTX).
4998 *) OAEP decoding bug fix.
5001 *) Support INSTALL_PREFIX for package builders, as proposed by
5005 *) New Configure options "threads" and "no-threads". For systems
5006 where the proper compiler options are known (currently Solaris
5007 and Linux), "threads" is the default.
5010 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5013 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5014 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5015 such as /usr/local/bin.
5018 *) "make linux-shared" to build shared libraries.
5019 [Niels Poppe <niels@netbox.org>]
5021 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5024 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5025 extension adding in x509 utility.
5028 *) Remove NOPROTO sections and error code comments.
5031 *) Partial rewrite of the DEF file generator to now parse the ANSI
5035 *) New Configure options --prefix=DIR and --openssldir=DIR.
5038 *) Complete rewrite of the error code script(s). It is all now handled
5039 by one script at the top level which handles error code gathering,
5040 header rewriting and C source file generation. It should be much better
5041 than the old method: it now uses a modified version of Ulf's parser to
5042 read the ANSI prototypes in all header files (thus the old K&R definitions
5043 aren't needed for error creation any more) and do a better job of
5044 translating function codes into names. The old 'ASN1 error code imbedded
5045 in a comment' is no longer necessary and it doesn't use .err files which
5046 have now been deleted. Also the error code call doesn't have to appear all
5047 on one line (which resulted in some large lines...).
5050 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5053 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5054 0 (which usually indicates a closed connection), but continue reading.
5057 *) Fix some race conditions.
5060 *) Add support for CRL distribution points extension. Add Certificate
5061 Policies and CRL distribution points documentation.
5064 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5067 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5068 8 of keying material. Merlin has also confirmed interop with this fix
5069 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5070 [Merlin Hughes <merlin@baltimore.ie>]
5072 *) Fix lots of warnings.
5073 [Richard Levitte <levitte@stacken.kth.se>]
5075 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5076 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5077 [Richard Levitte <levitte@stacken.kth.se>]
5079 *) Fix problems with sizeof(long) == 8.
5080 [Andy Polyakov <appro@fy.chalmers.se>]
5082 *) Change functions to ANSI C.
5085 *) Fix typos in error codes.
5086 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5088 *) Remove defunct assembler files from Configure.
5091 *) SPARC v8 assembler BIGNUM implementation.
5092 [Andy Polyakov <appro@fy.chalmers.se>]
5094 *) Support for Certificate Policies extension: both print and set.
5095 Various additions to support the r2i method this uses.
5098 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5099 return a const string when you are expecting an allocated buffer.
5102 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5103 types DirectoryString and DisplayText.
5106 *) Add code to allow r2i extensions to access the configuration database,
5107 add an LHASH database driver and add several ctx helper functions.
5110 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5111 fail when they extended the size of a BIGNUM.
5114 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5115 support typesafe stack.
5118 *) Fix typo in SSL_[gs]et_options().
5119 [Nils Frostberg <nils@medcom.se>]
5121 *) Delete various functions and files that belonged to the (now obsolete)
5122 old X509V3 handling code.
5125 *) New Configure option "rsaref".
5128 *) Don't auto-generate pem.h.
5131 *) Introduce type-safe ASN.1 SETs.
5134 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5135 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5137 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5138 that links with OpenSSL (well at least cause lots of warnings), but fear
5139 not: the conversion is trivial, and it eliminates loads of evil casts. A
5140 few STACKed things have been converted already. Feel free to convert more.
5141 In the fullness of time, I'll do away with the STACK type altogether.
5144 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5145 specified in <certfile> by updating the entry in the index.txt file.
5146 This way one no longer has to edit the index.txt file manually for
5147 revoking a certificate. The -revoke option does the gory details now.
5148 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5150 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5151 `-text' option at all and this way the `-noout -text' combination was
5152 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5153 [Ralf S. Engelschall]
5155 *) Make sure a corresponding plain text error message exists for the
5156 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5157 verify callback function determined that a certificate was revoked.
5158 [Ralf S. Engelschall]
5160 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5161 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5162 all available cipers including rc5, which was forgotten until now.
5163 In order to let the testing shell script know which algorithms
5164 are available, a new (up to now undocumented) command
5165 "openssl list-cipher-commands" is used.
5168 *) Bugfix: s_client occasionally would sleep in select() when
5169 it should have checked SSL_pending() first.
5172 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5173 the raw DSA values prior to ASN.1 encoding.
5176 *) Tweaks to Configure
5177 [Niels Poppe <niels@netbox.org>]
5179 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5183 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5186 *) New config option to avoid instructions that are illegal on the 80386.
5187 The default code is faster, but requires at least a 486.
5190 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5191 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5192 same as SSL2_VERSION anyway.
5195 *) New "-showcerts" option for s_client.
5198 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5199 application. Various cleanups and fixes.
5202 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5203 modify error routines to work internally. Add error codes and PBE init
5204 to library startup routines.
5207 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5208 packing functions to asn1 and evp. Changed function names and error
5209 codes along the way.
5212 *) PKCS12 integration: and so it begins... First of several patches to
5213 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5214 objects to objects.h
5217 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5218 and display support for Thawte strong extranet extension.
5221 *) Add LinuxPPC support.
5222 [Jeff Dubrule <igor@pobox.org>]
5224 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5225 bn_div_words in alpha.s.
5226 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5228 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5229 OAEP isn't supported when OpenSSL is built with RSAref.
5230 [Ulf Moeller <ulf@fitug.de>]
5232 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5233 so they no longer are missing under -DNOPROTO.
5234 [Soren S. Jorvang <soren@t.dk>]
5237 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5239 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5240 doesn't work when the session is reused. Coming soon!
5243 *) Fix a security hole, that allows sessions to be reused in the wrong
5244 context thus bypassing client cert protection! All software that uses
5245 client certs and session caches in multiple contexts NEEDS PATCHING to
5246 allow session reuse! A fuller solution is in the works.
5247 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5249 *) Some more source tree cleanups (removed obsolete files
5250 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5251 permission on "config" script to be executable) and a fix for the INSTALL
5253 [Ulf Moeller <ulf@fitug.de>]
5255 *) Remove some legacy and erroneous uses of malloc, free instead of
5257 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5259 *) Make rsa_oaep_test return non-zero on error.
5260 [Ulf Moeller <ulf@fitug.de>]
5262 *) Add support for native Solaris shared libraries. Configure
5263 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5264 if someone would make that last step automatic.
5265 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5267 *) ctx_size was not built with the right compiler during "make links". Fixed.
5270 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5271 except NULL ciphers". This means the default cipher list will no longer
5272 enable NULL ciphers. They need to be specifically enabled e.g. with
5273 the string "DEFAULT:eNULL".
5276 *) Fix to RSA private encryption routines: if p < q then it would
5277 occasionally produce an invalid result. This will only happen with
5278 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5281 *) Be less restrictive and allow also `perl util/perlpath.pl
5282 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5283 because this way one can also use an interpreter named `perl5' (which is
5284 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5285 installed as `perl').
5286 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5288 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5289 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5291 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5292 advapi32.lib to Win32 build and change the pem test comparision
5293 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5294 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5295 and crypto/des/ede_cbcm_enc.c.
5298 *) DES quad checksum was broken on big-endian architectures. Fixed.
5301 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5302 Win32 test batch file so it (might) work again. The Win32 test batch file
5303 is horrible: I feel ill....
5306 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5307 in e_os.h. Audit of header files to check ANSI and non ANSI
5308 sections: 10 functions were absent from non ANSI section and not exported
5309 from Windows DLLs. Fixed up libeay.num for new functions.
5312 *) Make `openssl version' output lines consistent.
5313 [Ralf S. Engelschall]
5315 *) Fix Win32 symbol export lists for BIO functions: Added
5316 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5317 to ms/libeay{16,32}.def.
5318 [Ralf S. Engelschall]
5320 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5321 fine under Unix and passes some trivial tests I've now added. But the
5322 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5323 added to make sure no one expects that this stuff really works in the
5324 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5325 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5327 [Ralf S. Engelschall]
5329 *) Fix the generation of two part addresses in perl.
5330 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5332 *) Add config entry for Linux on MIPS.
5333 [John Tobey <jtobey@channel1.com>]
5335 *) Make links whenever Configure is run, unless we are on Windoze.
5338 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5339 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5343 *) Add a useful kludge to allow package maintainers to specify compiler and
5344 other platforms details on the command line without having to patch the
5345 Configure script everytime: One now can use ``perl Configure
5346 <id>:<details>'', i.e. platform ids are allowed to have details appended
5347 to them (seperated by colons). This is treated as there would be a static
5348 pre-configured entry in Configure's %table under key <id> with value
5349 <details> and ``perl Configure <id>'' is called. So, when you want to
5350 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5351 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5352 now, which overrides the FreeBSD-elf entry on-the-fly.
5353 [Ralf S. Engelschall]
5355 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5358 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5359 on the `perl Configure ...' command line. This way one can compile
5360 OpenSSL libraries with Position Independent Code (PIC) which is needed
5361 for linking it into DSOs.
5362 [Ralf S. Engelschall]
5364 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5368 *) Cleaned up the LICENSE document: The official contact for any license
5369 questions now is the OpenSSL core team under openssl-core@openssl.org.
5370 And add a paragraph about the dual-license situation to make sure people
5371 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5372 to the OpenSSL toolkit.
5373 [Ralf S. Engelschall]
5375 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5376 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5377 Additonally cleaned up the `make links' target: Remove unnecessary
5378 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5379 to speed processing and no longer clutter the display with confusing
5380 stuff. Instead only the actually done links are displayed.
5381 [Ralf S. Engelschall]
5383 *) Permit null encryption ciphersuites, used for authentication only. It used
5384 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5385 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5389 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5390 signed attributes when verifying signatures (this would break them),
5391 the detached data encoding was wrong and public keys obtained using
5392 X509_get_pubkey() weren't freed.
5395 *) Add text documentation for the BUFFER functions. Also added a work around
5396 to a Win95 console bug. This was triggered by the password read stuff: the
5397 last character typed gets carried over to the next fread(). If you were
5398 generating a new cert request using 'req' for example then the last
5399 character of the passphrase would be CR which would then enter the first
5403 *) Added the new `Includes OpenSSL Cryptography Software' button as
5404 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5405 button and can be used by applications based on OpenSSL to show the
5406 relationship to the OpenSSL project.
5407 [Ralf S. Engelschall]
5409 *) Remove confusing variables in function signatures in files
5410 ssl/ssl_lib.c and ssl/ssl.h.
5411 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5413 *) Don't install bss_file.c under PREFIX/include/
5414 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5416 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5417 functions that return function pointers and has support for NT specific
5418 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5419 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5420 unsigned to signed types: this was killing the Win32 compile.
5423 *) Add new certificate file to stack functions,
5424 SSL_add_dir_cert_subjects_to_stack() and
5425 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5426 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5427 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5428 This means that Apache-SSL and similar packages don't have to mess around
5429 to add as many CAs as they want to the preferred list.
5432 *) Experiment with doxygen documentation. Currently only partially applied to
5434 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5435 openssl.doxy as the configuration file.
5438 *) Get rid of remaining C++-style comments which strict C compilers hate.
5439 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5441 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5442 compiled in by default: it has problems with large keys.
5445 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5446 DH private keys and/or callback functions which directly correspond to
5447 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5448 is needed for applications which have to configure certificates on a
5449 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5451 For the RSA certificate situation is makes no difference, but
5452 for the DSA certificate situation this fixes the "no shared cipher"
5453 problem where the OpenSSL cipher selection procedure failed because the
5454 temporary keys were not overtaken from the context and the API provided
5455 no way to reconfigure them.
5456 The new functions now let applications reconfigure the stuff and they
5457 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5458 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5459 non-public-API function ssl_cert_instantiate() is used as a helper
5460 function and also to reduce code redundancy inside ssl_rsa.c.
5461 [Ralf S. Engelschall]
5463 *) Move s_server -dcert and -dkey options out of the undocumented feature
5464 area because they are useful for the DSA situation and should be
5465 recognized by the users.
5466 [Ralf S. Engelschall]
5468 *) Fix the cipher decision scheme for export ciphers: the export bits are
5469 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5470 SSL_EXP_MASK. So, the original variable has to be used instead of the
5471 already masked variable.
5472 [Richard Levitte <levitte@stacken.kth.se>]
5474 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5475 [Richard Levitte <levitte@stacken.kth.se>]
5477 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5478 from `int' to `unsigned int' because it's a length and initialized by
5479 EVP_DigestFinal() which expects an `unsigned int *'.
5480 [Richard Levitte <levitte@stacken.kth.se>]
5482 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5483 script. Instead use the usual Shell->Perl transition trick.
5484 [Ralf S. Engelschall]
5486 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5487 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5488 -noout -modulus' as it's already the case for `openssl rsa -noout
5489 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5490 currently the public key is printed (a decision which was already done by
5491 `openssl dsa -modulus' in the past) which serves a similar purpose.
5492 Additionally the NO_RSA no longer completely removes the whole -modulus
5493 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5495 [Ralf S. Engelschall]
5497 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5498 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5499 [Arne Ansper <arne@ats.cyber.ee>]
5501 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5502 to be added. Now both 'req' and 'ca' can use new objects defined in the
5506 *) Add cool BIO that does syslog (or event log on NT).
5507 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5509 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5510 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5511 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5512 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5515 *) Add preliminary config info for new extension code.
5518 *) Make RSA_NO_PADDING really use no padding.
5519 [Ulf Moeller <ulf@fitug.de>]
5521 *) Generate errors when private/public key check is done.
5524 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5525 for some CRL extensions and new objects added.
5528 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5529 key usage extension and fuller support for authority key id.
5532 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5533 padding method for RSA, which is recommended for new applications in PKCS
5534 #1 v2.0 (RFC 2437, October 1998).
5535 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5536 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5537 against Bleichbacher's attack on RSA.
5538 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5541 *) Updates to the new SSL compression code
5542 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5544 *) Fix so that the version number in the master secret, when passed
5545 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5546 (because the server will not accept higher), that the version number
5547 is 0x03,0x01, not 0x03,0x00
5548 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5550 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5551 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5552 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5555 *) Support for RAW extensions where an arbitrary extension can be
5556 created by including its DER encoding. See apps/openssl.cnf for
5560 *) Make sure latest Perl versions don't interpret some generated C array
5561 code as Perl array code in the crypto/err/err_genc.pl script.
5562 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5564 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5565 not many people have the assembler. Various Win32 compilation fixes and
5566 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5570 *) Modify configure script 'Configure' to automatically create crypto/date.h
5571 file under Win32 and also build pem.h from pem.org. New script
5572 util/mkfiles.pl to create the MINFO file on environments that can't do a
5573 'make files': perl util/mkfiles.pl >MINFO should work.
5576 *) Major rework of DES function declarations, in the pursuit of correctness
5577 and purity. As a result, many evil casts evaporated, and some weirdness,
5578 too. You may find this causes warnings in your code. Zapping your evil
5579 casts will probably fix them. Mostly.
5582 *) Fix for a typo in asn1.h. Bug fix to object creation script
5583 obj_dat.pl. It considered a zero in an object definition to mean
5584 "end of object": none of the objects in objects.h have any zeros
5585 so it wasn't spotted.
5586 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5588 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5589 Masking (CBCM). In the absence of test vectors, the best I have been able
5590 to do is check that the decrypt undoes the encrypt, so far. Send me test
5591 vectors if you have them.
5594 *) Correct calculation of key length for export ciphers (too much space was
5595 allocated for null ciphers). This has not been tested!
5598 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5599 message is now correct (it understands "crypto" and "ssl" on its
5600 command line). There is also now an "update" option. This will update
5601 the util/ssleay.num and util/libeay.num files with any new functions.
5603 perl util/mkdef.pl crypto ssl update
5604 it will update them.
5607 *) Overhauled the Perl interface (perl/*):
5608 - ported BN stuff to OpenSSL's different BN library
5609 - made the perl/ source tree CVS-aware
5610 - renamed the package from SSLeay to OpenSSL (the files still contain
5611 their history because I've copied them in the repository)
5612 - removed obsolete files (the test scripts will be replaced
5613 by better Test::Harness variants in the future)
5614 [Ralf S. Engelschall]
5616 *) First cut for a very conservative source tree cleanup:
5617 1. merge various obsolete readme texts into doc/ssleay.txt
5618 where we collect the old documents and readme texts.
5619 2. remove the first part of files where I'm already sure that we no
5620 longer need them because of three reasons: either they are just temporary
5621 files which were left by Eric or they are preserved original files where
5622 I've verified that the diff is also available in the CVS via "cvs diff
5623 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5624 the crypto/md/ stuff).
5625 [Ralf S. Engelschall]
5627 *) More extension code. Incomplete support for subject and issuer alt
5628 name, issuer and authority key id. Change the i2v function parameters
5629 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5630 what that's for :-) Fix to ASN1 macro which messed up
5631 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5634 *) Preliminary support for ENUMERATED type. This is largely copied from the
5638 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5639 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5641 *) Make sure `make rehash' target really finds the `openssl' program.
5642 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5644 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5645 like to hear about it if this slows down other processors.
5648 *) Add CygWin32 platform information to Configure script.
5649 [Alan Batie <batie@aahz.jf.intel.com>]
5651 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5652 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5654 *) New program nseq to manipulate netscape certificate sequences
5657 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5661 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5662 but the BN code had some problems that would cause failures when
5663 doing certificate verification and some other functions.
5664 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5666 *) Add ASN1 and PEM code to support netscape certificate sequences.
5669 *) Add ASN1 and PEM code to support netscape certificate sequences.
5672 *) Add several PKIX and private extended key usage OIDs.
5675 *) Modify the 'ca' program to handle the new extension code. Modify
5676 openssl.cnf for new extension format, add comments.
5679 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5680 and add a sample to openssl.cnf so req -x509 now adds appropriate
5684 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5685 error code, add initial support to X509_print() and x509 application.
5688 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5689 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5690 stuff is currently isolated and isn't even compiled yet.
5693 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5694 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5695 Removed the versions check from X509 routines when loading extensions:
5696 this allows certain broken certificates that don't set the version
5697 properly to be processed.
5700 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5701 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5702 can still be regenerated with "make depend".
5705 *) Spelling mistake in C version of CAST-128.
5706 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5708 *) Changes to the error generation code. The perl script err-code.pl
5709 now reads in the old error codes and retains the old numbers, only
5710 adding new ones if necessary. It also only changes the .err files if new
5711 codes are added. The makefiles have been modified to only insert errors
5712 when needed (to avoid needlessly modifying header files). This is done
5713 by only inserting errors if the .err file is newer than the auto generated
5714 C file. To rebuild all the error codes from scratch (the old behaviour)
5715 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5716 or delete all the .err files.
5719 *) CAST-128 was incorrectly implemented for short keys. The C version has
5720 been fixed, but is untested. The assembler versions are also fixed, but
5721 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5722 to regenerate it if needed.
5723 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5724 Hagino <itojun@kame.net>]
5726 *) File was opened incorrectly in randfile.c.
5727 [Ulf Möller <ulf@fitug.de>]
5729 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5730 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5731 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5732 al: it's just almost always a UTCTime. Note this patch adds new error
5733 codes so do a "make errors" if there are problems.
5736 *) Correct Linux 1 recognition in config.
5737 [Ulf Möller <ulf@fitug.de>]
5739 *) Remove pointless MD5 hash when using DSA keys in ca.
5740 [Anonymous <nobody@replay.com>]
5742 *) Generate an error if given an empty string as a cert directory. Also
5743 generate an error if handed NULL (previously returned 0 to indicate an
5744 error, but didn't set one).
5745 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5747 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5750 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5751 parameters. This was causing a warning which killed off the Win32 compile.
5754 *) Remove C++ style comments from crypto/bn/bn_local.h.
5755 [Neil Costigan <neil.costigan@celocom.com>]
5757 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5758 based on a text string, looking up short and long names and finally
5759 "dot" format. The "dot" format stuff didn't work. Added new function
5760 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5761 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5762 OID is not part of the table.
5765 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5766 X509_LOOKUP_by_alias().
5769 *) Sort openssl functions by name.
5772 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5773 encryption from sample DSA keys (in case anyone is interested the password
5777 *) Make _all_ *_free functions accept a NULL pointer.
5778 [Frans Heymans <fheymans@isaserver.be>]
5780 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5782 [Anonymous <nobody@replay.com>]
5784 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5785 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5787 *) Don't blow it for numeric -newkey arguments to apps/req.
5788 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5790 *) Temp key "for export" tests were wrong in s3_srvr.c.
5791 [Anonymous <nobody@replay.com>]
5793 *) Add prototype for temp key callback functions
5794 SSL_CTX_set_tmp_{rsa,dh}_callback().
5797 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5798 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5801 *) X509_name_add_entry() freed the wrong thing after an error.
5802 [Arne Ansper <arne@ats.cyber.ee>]
5804 *) rsa_eay.c would attempt to free a NULL context.
5805 [Arne Ansper <arne@ats.cyber.ee>]
5807 *) BIO_s_socket() had a broken should_retry() on Windoze.
5808 [Arne Ansper <arne@ats.cyber.ee>]
5810 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5811 [Arne Ansper <arne@ats.cyber.ee>]
5813 *) Make sure the already existing X509_STORE->depth variable is initialized
5814 in X509_STORE_new(), but document the fact that this variable is still
5815 unused in the certificate verification process.
5816 [Ralf S. Engelschall]
5818 *) Fix the various library and apps files to free up pkeys obtained from
5819 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5822 *) Fix reference counting in X509_PUBKEY_get(). This makes
5823 demos/maurice/example2.c work, amongst others, probably.
5824 [Steve Henson and Ben Laurie]
5826 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5827 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5828 are no longer created. This way we have a single and consistent command
5829 line interface `openssl <command>', similar to `cvs <command>'.
5830 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5832 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5833 BIT STRING wrapper always have zero unused bits.
5836 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5839 *) Make the top-level INSTALL documentation easier to understand.
5842 *) Makefiles updated to exit if an error occurs in a sub-directory
5843 make (including if user presses ^C) [Paul Sutton]
5845 *) Make Montgomery context stuff explicit in RSA data structure.
5848 *) Fix build order of pem and err to allow for generated pem.h.
5851 *) Fix renumbering bug in X509_NAME_delete_entry().
5854 *) Enhanced the err-ins.pl script so it makes the error library number
5855 global and can add a library name. This is needed for external ASN1 and
5856 other error libraries.
5859 *) Fixed sk_insert which never worked properly.
5862 *) Fix ASN1 macros so they can handle indefinite length construted
5863 EXPLICIT tags. Some non standard certificates use these: they can now
5867 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5868 into a single doc/ssleay.txt bundle. This way the information is still
5869 preserved but no longer messes up this directory. Now it's new room for
5870 the new set of documenation files.
5871 [Ralf S. Engelschall]
5873 *) SETs were incorrectly DER encoded. This was a major pain, because they
5874 shared code with SEQUENCEs, which aren't coded the same. This means that
5875 almost everything to do with SETs or SEQUENCEs has either changed name or
5876 number of arguments.
5877 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5879 *) Fix test data to work with the above.
5882 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5883 was already fixed by Eric for 0.9.1 it seems.
5884 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5886 *) Autodetect FreeBSD3.
5889 *) Fix various bugs in Configure. This affects the following platforms:
5893 unixware-2.0-pentium
5897 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5898 before they are needed.
5901 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5905 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5907 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5908 changed SSLeay to OpenSSL in version strings.
5909 [Ralf S. Engelschall]
5911 *) Some fixups to the top-level documents.
5914 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5915 because the symlink to include/ was missing.
5916 [Ralf S. Engelschall]
5918 *) Incorporated the popular no-RSA/DSA-only patches
5919 which allow to compile a RSA-free SSLeay.
5920 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5922 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5923 when "ssleay" is still not found.
5924 [Ralf S. Engelschall]
5926 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5927 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5929 *) Updated the README file.
5930 [Ralf S. Engelschall]
5932 *) Added various .cvsignore files in the CVS repository subdirs
5933 to make a "cvs update" really silent.
5934 [Ralf S. Engelschall]
5936 *) Recompiled the error-definition header files and added
5937 missing symbols to the Win32 linker tables.
5938 [Ralf S. Engelschall]
5940 *) Cleaned up the top-level documents;
5941 o new files: CHANGES and LICENSE
5942 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5943 o merged COPYRIGHT into LICENSE
5944 o removed obsolete TODO file
5945 o renamed MICROSOFT to INSTALL.W32
5946 [Ralf S. Engelschall]
5948 *) Removed dummy files from the 0.9.1b source tree:
5949 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5950 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5951 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5952 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5953 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5954 [Ralf S. Engelschall]
5956 *) Added various platform portability fixes.
5959 *) The Genesis of the OpenSSL rpject:
5960 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5961 Young and Tim J. Hudson created while they were working for C2Net until
5963 [The OpenSSL Project]
5966 Changes between 0.9.0b and 0.9.1b [not released]
5968 *) Updated a few CA certificates under certs/
5971 *) Changed some BIGNUM api stuff.
5974 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5975 DGUX x86, Linux Alpha, etc.
5978 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5979 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5983 *) Add -strparse option to asn1pars program which parses nested
5985 [Dr Stephen Henson <shenson@bigfoot.com>]
5987 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5990 *) DSA fix for "ca" program.
5993 *) Added "-genkey" option to "dsaparam" program.
5996 *) Added RIPE MD160 (rmd160) message digest.
5999 *) Added -a (all) option to "ssleay version" command.
6002 *) Added PLATFORM define which is the id given to Configure.
6005 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6008 *) Extended the ASN.1 parser routines.
6011 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6014 *) Added a BN_CTX to the BN library.
6017 *) Fixed the weak key values in DES library
6020 *) Changed API in EVP library for cipher aliases.
6023 *) Added support for RC2/64bit cipher.
6026 *) Converted the lhash library to the crypto/mem.c functions.
6029 *) Added more recognized ASN.1 object ids.
6032 *) Added more RSA padding checks for SSL/TLS.
6035 *) Added BIO proxy/filter functionality.
6038 *) Added extra_certs to SSL_CTX which can be used
6039 send extra CA certificates to the client in the CA cert chain sending
6040 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6043 *) Now Fortezza is denied in the authentication phase because
6044 this is key exchange mechanism is not supported by SSLeay at all.
6047 *) Additional PKCS1 checks.
6050 *) Support the string "TLSv1" for all TLS v1 ciphers.
6053 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6054 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6057 *) Fixed a few memory leaks.
6060 *) Fixed various code and comment typos.
6063 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6064 bytes sent in the client random.
6065 [Edward Bishop <ebishop@spyglass.com>]