5 Changes between 0.9.6h and 0.9.7 [XX xxx 2002]
7 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
8 seems that in spite of existing for more than a year, many application
9 author have done nothing to provide the necessary callbacks, which
10 means that this particular engine will not work properly anywhere.
11 This is a very unfortunate situation which forces us, in the name
12 of usability, to give the hw_ncipher.c a static lock, which is part
14 NOTE: This is for the 0.9.7 series ONLY. This hack will never
15 appear in 0.9.8 or later. We EXPECT application authors to have
16 dealt properly with this when 0.9.8 is released (unless we actually
17 make such changes in the libcrypto locking code that changes will
18 have to be made anyway).
21 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
22 octets have been read, EOF or an error occurs. Without this change
23 some truncated ASN1 structures will not produce an error.
26 *) Disable Heimdal support, since it hasn't been fully implemented.
27 Still give the possibility to force the use of Heimdal, but with
28 warnings and a request that patches get sent to openssl-dev.
31 *) Add the VC-CE target, introduce the WINCE sysname, and add
32 INSTALL.WCE and appropriate conditionals to make it build.
33 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
35 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
36 cygssl-x.y.z.dll, where x, y and z are the major, minor and
37 edit numbers of the version.
38 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
40 *) Introduce safe string copy and catenation functions
41 (BUF_strlcpy() and BUF_strlcat()).
42 [Ben Laurie (CHATS) and Richard Levitte]
44 *) Avoid using fixed-size buffers for one-line DNs.
47 *) Add BUF_MEM_grow_clean() to avoid information leakage when
48 resizing buffers containing secrets, and use where appropriate.
51 *) Avoid using fixed size buffers for configuration file location.
54 *) Avoid filename truncation for various CA files.
57 *) Use sizeof in preference to magic numbers.
60 *) Avoid filename truncation in cert requests.
63 *) Add assertions to check for (supposedly impossible) buffer
67 *) Don't cache truncated DNS entries in the local cache (this could
68 potentially lead to a spoofing attack).
71 *) Fix various buffers to be large enough for hex/decimal
72 representations in a platform independent manner.
75 *) Add CRYPTO_realloc_clean() to avoid information leakage when
76 resizing buffers containing secrets, and use where appropriate.
79 *) Add BIO_indent() to avoid much slightly worrying code to do
83 *) Convert sprintf()/BIO_puts() to BIO_printf().
86 *) buffer_gets() could terminate with the buffer only half
90 *) Add assertions to prevent user-supplied crypto functions from
91 overflowing internal buffers by having large block sizes, etc.
94 *) New OPENSSL_assert() macro (similar to assert(), but enabled
98 *) Eliminate unused copy of key in RC4.
101 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
104 *) Fix off-by-one error in EGD path.
107 *) If RANDFILE path is too long, ignore instead of truncating.
110 *) Eliminate unused and incorrectly sized X.509 structure
114 *) Eliminate unused and dangerous function knumber().
117 *) Eliminate unused and dangerous structure, KSSL_ERR.
120 *) Protect against overlong session ID context length in an encoded
121 session object. Since these are local, this does not appear to be
125 *) Change from security patch (see 0.9.6e below) that did not affect
126 the 0.9.6 release series:
128 Remote buffer overflow in SSL3 protocol - an attacker could
129 supply an oversized master key in Kerberos-enabled versions.
133 *) Change the SSL kerb5 codes to match RFC 2712.
136 *) Make -nameopt work fully for req and add -reqopt switch.
137 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
139 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
140 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
142 *) Make sure tests can be performed even if the corresponding algorithms
143 have been removed entirely. This was also the last step to make
144 OpenSSL compilable with DJGPP under all reasonable conditions.
145 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
147 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
148 to allow version independent disabling of normally unselected ciphers,
149 which may be activated as a side-effect of selecting a single cipher.
151 (E.g., cipher list string "RSA" enables ciphersuites that are left
152 out of "ALL" because they do not provide symmetric encryption.
153 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
154 [Lutz Jaenicke, Bodo Moeller]
156 *) Add appropriate support for separate platform-dependent build
157 directories. The recommended way to make a platform-dependent
158 build directory is the following (tested on Linux), maybe with
161 # Place yourself outside of the OpenSSL source tree. In
162 # this example, the environment variable OPENSSL_SOURCE
163 # is assumed to contain the absolute OpenSSL source directory.
164 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
165 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
166 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
167 mkdir -p `dirname $F`
168 ln -s $OPENSSL_SOURCE/$F $F
171 To be absolutely sure not to disturb the source tree, a "make clean"
172 is a good thing. If it isn't successfull, don't worry about it,
173 it probably means the source directory is very clean.
176 *) Make sure any ENGINE control commands make local copies of string
177 pointers passed to them whenever necessary. Otherwise it is possible
178 the caller may have overwritten (or deallocated) the original string
179 data when a later ENGINE operation tries to use the stored values.
180 [Götz Babin-Ebell <babinebell@trustcenter.de>]
182 *) Improve diagnostics in file reading and command-line digests.
183 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
185 *) Add AES modes CFB and OFB to the object database. Correct an
186 error in AES-CFB decryption.
189 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
190 allows existing EVP_CIPHER_CTX structures to be reused after
191 calling EVP_*Final(). This behaviour is used by encryption
192 BIOs and some applications. This has the side effect that
193 applications must explicitly clean up cipher contexts with
194 EVP_CIPHER_CTX_cleanup() or they will leak memory.
197 *) Check the values of dna and dnb in bn_mul_recursive before calling
198 bn_mul_comba (a non zero value means the a or b arrays do not contain
199 n2 elements) and fallback to bn_mul_normal if either is not zero.
202 *) Fix escaping of non-ASCII characters when using the -subj option
203 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
206 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
207 form for "surname", serialNumber has no short form.
208 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
209 therefore remove "mail" short name for "internet 7".
210 The OID for unique identifiers in X509 certificates is
211 x500UniqueIdentifier, not uniqueIdentifier.
212 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
215 *) Add an "init" command to the ENGINE config module and auto initialize
216 ENGINEs. Without any "init" command the ENGINE will be initialized
217 after all ctrl commands have been executed on it. If init=1 the
218 ENGINE is initailized at that point (ctrls before that point are run
219 on the uninitialized ENGINE and after on the initialized one). If
220 init=0 then the ENGINE will not be iniatialized at all.
223 *) Fix the 'app_verify_callback' interface so that the user-defined
224 argument is actually passed to the callback: In the
225 SSL_CTX_set_cert_verify_callback() prototype, the callback
226 declaration has been changed from
229 int (*cb)(X509_STORE_CTX *,void *);
230 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
231 i=s->ctx->app_verify_callback(&ctx)
232 has been changed into
233 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
235 To update applications using SSL_CTX_set_cert_verify_callback(),
236 a dummy argument can be added to their callback functions.
237 [D. K. Smetters <smetters@parc.xerox.com>]
239 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
240 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
242 *) Add and OPENSSL_LOAD_CONF define which will cause
243 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
244 This allows older applications to transparently support certain
245 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
246 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
247 load the config file and OPENSSL_add_all_algorithms_conf() which will
248 always load it have also been added.
251 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
252 Adjust NIDs and EVP layer.
253 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
255 *) Config modules support in openssl utility.
257 Most commands now load modules from the config file,
258 though in a few (such as version) this isn't done
259 because it couldn't be used for anything.
261 In the case of ca and req the config file used is
262 the same as the utility itself: that is the -config
263 command line option can be used to specify an
267 *) Move default behaviour from OPENSSL_config(). If appname is NULL
268 use "openssl_conf" if filename is NULL use default openssl config file.
271 *) Add an argument to OPENSSL_config() to allow the use of an alternative
272 config section name. Add a new flag to tolerate a missing config file
273 and move code to CONF_modules_load_file().
276 *) Support for crypto accelerator cards from Accelerated Encryption
277 Processing, www.aep.ie. (Use engine 'aep')
278 The support was copied from 0.9.6c [engine] and adapted/corrected
279 to work with the new engine framework.
280 [AEP Inc. and Richard Levitte]
282 *) Support for SureWare crypto accelerator cards from Baltimore
283 Technologies. (Use engine 'sureware')
284 The support was copied from 0.9.6c [engine] and adapted
285 to work with the new engine framework.
288 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
289 make the newer ENGINE framework commands for the CHIL engine work.
290 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
292 *) Make it possible to produce shared libraries on ReliantUNIX.
293 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
295 *) Add the configuration target debug-linux-ppro.
296 Make 'openssl rsa' use the general key loading routines
297 implemented in apps.c, and make those routines able to
298 handle the key format FORMAT_NETSCAPE and the variant
300 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
302 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
303 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
305 *) Add -keyform to rsautl, and document -engine.
306 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
308 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
309 BIO_R_NO_SUCH_FILE error code rather than the generic
310 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
315 ERR_peek_last_error_line
316 ERR_peek_last_error_line_data.
320 ERR_peek_error_line_data,
321 but report on the latest error recorded rather than the first one
322 still in the error queue.
323 [Ben Laurie, Bodo Moeller]
325 *) default_algorithms option in ENGINE config module. This allows things
327 default_algorithms = ALL
328 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
331 *) Prelminary ENGINE config module.
334 *) New experimental application configuration code.
337 *) Change the AES code to follow the same name structure as all other
338 symmetric ciphers, and behave the same way. Move everything to
339 the directory crypto/aes, thereby obsoleting crypto/rijndael.
340 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
342 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
343 [Ben Laurie and Theo de Raadt]
345 *) Add option to output public keys in req command.
346 [Massimiliano Pala madwolf@openca.org]
348 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
349 (up to about 10% better than before for P-192 and P-224).
352 *) New functions/macros
354 SSL_CTX_set_msg_callback(ctx, cb)
355 SSL_CTX_set_msg_callback_arg(ctx, arg)
356 SSL_set_msg_callback(ssl, cb)
357 SSL_set_msg_callback_arg(ssl, arg)
359 to request calling a callback function
361 void cb(int write_p, int version, int content_type,
362 const void *buf, size_t len, SSL *ssl, void *arg)
364 whenever a protocol message has been completely received
365 (write_p == 0) or sent (write_p == 1). Here 'version' is the
366 protocol version according to which the SSL library interprets
367 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
368 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
369 the content type as defined in the SSL 3.0/TLS 1.0 protocol
370 specification (change_cipher_spec(20), alert(21), handshake(22)).
371 'buf' and 'len' point to the actual message, 'ssl' to the
372 SSL object, and 'arg' is the application-defined value set by
373 SSL[_CTX]_set_msg_callback_arg().
375 'openssl s_client' and 'openssl s_server' have new '-msg' options
376 to enable a callback that displays all protocol messages.
379 *) Change the shared library support so shared libraries are built as
380 soon as the corresponding static library is finished, and thereby get
381 openssl and the test programs linked against the shared library.
382 This still only happens when the keyword "shard" has been given to
383 the configuration scripts.
385 NOTE: shared library support is still an experimental thing, and
386 backward binary compatibility is still not guaranteed.
387 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
389 *) Add support for Subject Information Access extension.
390 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
392 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
393 additional bytes when new memory had to be allocated, not just
394 when reusing an existing buffer.
397 *) New command line and configuration option 'utf8' for the req command.
398 This allows field values to be specified as UTF8 strings.
401 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
402 runs for the former and machine-readable output for the latter.
405 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
406 of the e-mail address in the DN (i.e., it will go into a certificate
407 extension only). The new configuration file option 'email_in_dn = no'
409 [Massimiliano Pala madwolf@openca.org]
411 *) Change all functions with names starting with des_ to be starting
412 with DES_ instead. Add wrappers that are compatible with libdes,
413 but are named _ossl_old_des_*. Finally, add macros that map the
414 des_* symbols to the corresponding _ossl_old_des_* if libdes
415 compatibility is desired. If OpenSSL 0.9.6c compatibility is
416 desired, the des_* symbols will be mapped to DES_*, with one
419 Since we provide two compatibility mappings, the user needs to
420 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
421 compatibility is desired. The default (i.e., when that macro
422 isn't defined) is OpenSSL 0.9.6c compatibility.
424 There are also macros that enable and disable the support of old
425 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
426 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
427 are defined, the default will apply: to support the old des routines.
429 In either case, one must include openssl/des.h to get the correct
430 definitions. Do not try to just include openssl/des_old.h, that
433 NOTE: This is a major break of an old API into a new one. Software
434 authors are encouraged to switch to the DES_ style functions. Some
435 time in the future, des_old.h and the libdes compatibility functions
436 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
437 default), and then completely removed.
440 *) Test for certificates which contain unsupported critical extensions.
441 If such a certificate is found during a verify operation it is
442 rejected by default: this behaviour can be overridden by either
443 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
444 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
445 X509_supported_extension() has also been added which returns 1 if a
446 particular extension is supported.
449 *) Modify the behaviour of EVP cipher functions in similar way to digests
450 to retain compatibility with existing code.
453 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
454 compatibility with existing code. In particular the 'ctx' parameter does
455 not have to be to be initialized before the call to EVP_DigestInit() and
456 it is tidied up after a call to EVP_DigestFinal(). New function
457 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
458 EVP_MD_CTX_copy() changed to not require the destination to be
459 initialized valid and new function EVP_MD_CTX_copy_ex() added which
460 requires the destination to be valid.
462 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
463 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
466 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
467 so that complete 'Handshake' protocol structures are kept in memory
468 instead of overwriting 'msg_type' and 'length' with 'body' data.
471 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
472 [Massimo Santin via Richard Levitte]
474 *) Major restructuring to the underlying ENGINE code. This includes
475 reduction of linker bloat, separation of pure "ENGINE" manipulation
476 (initialisation, etc) from functionality dealing with implementations
477 of specific crypto iterfaces. This change also introduces integrated
478 support for symmetric ciphers and digest implementations - so ENGINEs
479 can now accelerate these by providing EVP_CIPHER and EVP_MD
480 implementations of their own. This is detailed in crypto/engine/README
481 as it couldn't be adequately described here. However, there are a few
482 API changes worth noting - some RSA, DSA, DH, and RAND functions that
483 were changed in the original introduction of ENGINE code have now
484 reverted back - the hooking from this code to ENGINE is now a good
485 deal more passive and at run-time, operations deal directly with
486 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
487 dereferencing through an ENGINE pointer any more. Also, the ENGINE
488 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
489 they were not being used by the framework as there is no concept of a
490 BIGNUM_METHOD and they could not be generalised to the new
491 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
492 ENGINE_cpy() has been removed as it cannot be consistently defined in
496 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
499 *) Change mkdef.pl to sort symbols that get the same entry number,
500 and make sure the automatically generated functions ERR_load_*
501 become part of libeay.num as well.
504 *) New function SSL_renegotiate_pending(). This returns true once
505 renegotiation has been requested (either SSL_renegotiate() call
506 or HelloRequest/ClientHello receveived from the peer) and becomes
507 false once a handshake has been completed.
508 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
509 sends a HelloRequest, but does not ensure that a handshake takes
510 place. SSL_renegotiate_pending() is useful for checking if the
511 client has followed the request.)
514 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
515 By default, clients may request session resumption even during
516 renegotiation (if session ID contexts permit); with this option,
517 session resumption is possible only in the first handshake.
519 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
520 more bits available for options that should not be part of
521 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
524 *) Add some demos for certificate and certificate request creation.
527 *) Make maximum certificate chain size accepted from the peer application
528 settable (SSL*_get/set_max_cert_list()), as proposed by
529 "Douglas E. Engert" <deengert@anl.gov>.
532 *) Add support for shared libraries for Unixware-7
533 (Boyd Lynn Gerber <gerberb@zenez.com>).
536 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
537 be done prior to destruction. Use this to unload error strings from
538 ENGINEs that load their own error strings. NB: This adds two new API
539 functions to "get" and "set" this destroy handler in an ENGINE.
542 *) Alter all existing ENGINE implementations (except "openssl" and
543 "openbsd") to dynamically instantiate their own error strings. This
544 makes them more flexible to be built both as statically-linked ENGINEs
545 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
546 Also, add stub code to each that makes building them as self-contained
547 shared-libraries easier (see README.ENGINE).
550 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
551 implementations into applications that are completely implemented in
552 self-contained shared-libraries. The "dynamic" ENGINE exposes control
553 commands that can be used to configure what shared-library to load and
554 to control aspects of the way it is handled. Also, made an update to
555 the README.ENGINE file that brings its information up-to-date and
556 provides some information and instructions on the "dynamic" ENGINE
557 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
560 *) Make it possible to unload ranges of ERR strings with a new
561 "ERR_unload_strings" function.
564 *) Add a copy() function to EVP_MD.
567 *) Make EVP_MD routines take a context pointer instead of just the
568 md_data void pointer.
571 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
572 that the digest can only process a single chunk of data
573 (typically because it is provided by a piece of
574 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
575 is only going to provide a single chunk of data, and hence the
576 framework needn't accumulate the data for oneshot drivers.
579 *) As with "ERR", make it possible to replace the underlying "ex_data"
580 functions. This change also alters the storage and management of global
581 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
582 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
583 index counters. The API functions that use this state have been changed
584 to take a "class_index" rather than pointers to the class's local STACK
585 and counter, and there is now an API function to dynamically create new
586 classes. This centralisation allows us to (a) plug a lot of the
587 thread-safety problems that existed, and (b) makes it possible to clean
588 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
589 such data would previously have always leaked in application code and
590 workarounds were in place to make the memory debugging turn a blind eye
591 to it. Application code that doesn't use this new function will still
592 leak as before, but their memory debugging output will announce it now
593 rather than letting it slide.
595 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
596 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
597 has a return value to indicate success or failure.
600 *) Make it possible to replace the underlying "ERR" functions such that the
601 global state (2 LHASH tables and 2 locks) is only used by the "default"
602 implementation. This change also adds two functions to "get" and "set"
603 the implementation prior to it being automatically set the first time
604 any other ERR function takes place. Ie. an application can call "get",
605 pass the return value to a module it has just loaded, and that module
606 can call its own "set" function using that value. This means the
607 module's "ERR" operations will use (and modify) the error state in the
608 application and not in its own statically linked copy of OpenSSL code.
611 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
612 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
613 the operation, and provides a more encapsulated way for external code
614 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
615 to use these functions rather than manually incrementing the counts.
617 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
620 *) Add EVP test program.
623 *) Add symmetric cipher support to ENGINE. Expect the API to change!
626 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
627 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
628 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
629 These allow a CRL to be built without having to access X509_CRL fields
630 directly. Modify 'ca' application to use new functions.
633 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
634 bug workarounds. Rollback attack detection is a security feature.
635 The problem will only arise on OpenSSL servers when TLSv1 is not
636 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
637 Software authors not wanting to support TLSv1 will have special reasons
638 for their choice and can explicitly enable this option.
639 [Bodo Moeller, Lutz Jaenicke]
641 *) Rationalise EVP so it can be extended: don't include a union of
642 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
643 (similar to those existing for EVP_CIPHER_CTX).
648 EVP_MD_CTX_init(&md); /* new function call */
649 EVP_DigestInit(&md, EVP_sha1());
650 EVP_DigestUpdate(&md, in, len);
651 EVP_DigestFinal(&md, out, NULL);
652 EVP_MD_CTX_cleanup(&md); /* new function call */
656 *) Make DES key schedule conform to the usual scheme, as well as
657 correcting its structure. This means that calls to DES functions
658 now have to pass a pointer to a des_key_schedule instead of a
659 plain des_key_schedule (which was actually always a pointer
664 des_set_key_checked(..., &ks);
665 des_ncbc_encrypt(..., &ks, ...);
667 (Note that a later change renames 'des_...' into 'DES_...'.)
670 *) Initial reduction of linker bloat: the use of some functions, such as
671 PEM causes large amounts of unused functions to be linked in due to
672 poor organisation. For example pem_all.c contains every PEM function
673 which has a knock on effect of linking in large amounts of (unused)
674 ASN1 code. Grouping together similar functions and splitting unrelated
675 functions prevents this.
678 *) Cleanup of EVP macros.
681 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
685 *) Add initial OCSP responder support to ocsp application. The
686 revocation information is handled using the text based index
687 use by the ca application. The responder can either handle
688 requests generated internally, supplied in files (for example
689 via a CGI script) or using an internal minimal server.
692 *) Add configuration choices to get zlib compression for TLS.
695 *) Changes to Kerberos SSL for RFC 2712 compliance:
696 1. Implemented real KerberosWrapper, instead of just using
697 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
698 2. Implemented optional authenticator field of KerberosWrapper.
700 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
701 and authenticator structs; see crypto/krb5/.
703 Generalized Kerberos calls to support multiple Kerberos libraries.
704 [Vern Staats <staatsvr@asc.hpc.mil>,
705 Jeffrey Altman <jaltman@columbia.edu>
708 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
709 already does with RSA. testdsa.h now has 'priv_key/pub_key'
710 values for each of the key sizes rather than having just
711 parameters (and 'speed' generating keys each time).
714 *) Speed up EVP routines.
717 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
718 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
719 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
720 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
722 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
723 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
724 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
727 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
729 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
732 *) Added the OS2-EMX target.
733 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
735 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
736 to support NCONF routines in extension code. New function CONF_set_nconf()
737 to allow functions which take an NCONF to also handle the old LHASH
738 structure: this means that the old CONF compatible routines can be
739 retained (in particular wrt extensions) without having to duplicate the
740 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
743 *) Enhance the general user interface with mechanisms for inner control
744 and with possibilities to have yes/no kind of prompts.
747 *) Change all calls to low level digest routines in the library and
748 applications to use EVP. Add missing calls to HMAC_cleanup() and
749 don't assume HMAC_CTX can be copied using memcpy().
750 [Verdon Walker <VWalker@novell.com>, Steve Henson]
752 *) Add the possibility to control engines through control names but with
753 arbitrary arguments instead of just a string.
754 Change the key loaders to take a UI_METHOD instead of a callback
755 function pointer. NOTE: this breaks binary compatibility with earlier
756 versions of OpenSSL [engine].
757 Adapt the nCipher code for these new conditions and add a card insertion
761 *) Enhance the general user interface with mechanisms to better support
762 dialog box interfaces, application-defined prompts, the possibility
763 to use defaults (for example default passwords from somewhere else)
764 and interrupts/cancellations.
767 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
768 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
771 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
772 tidy up some unnecessarily weird code in 'sk_new()').
773 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
775 *) Change the key loading routines for ENGINEs to use the same kind
776 callback (pem_password_cb) as all other routines that need this
780 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
781 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
782 than this minimum value is recommended.
785 *) New random seeder for OpenVMS, using the system process statistics
786 that are easily reachable.
789 *) Windows apparently can't transparently handle global
790 variables defined in DLLs. Initialisations such as:
792 const ASN1_ITEM *it = &ASN1_INTEGER_it;
794 wont compile. This is used by the any applications that need to
795 declare their own ASN1 modules. This was fixed by adding the option
796 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
797 needed for static libraries under Win32.
800 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
801 setting of purpose and trust fields. New X509_STORE trust and
802 purpose functions and tidy up setting in other SSL functions.
805 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
806 structure. These are inherited by X509_STORE_CTX when it is
807 initialised. This allows various defaults to be set in the
808 X509_STORE structure (such as flags for CRL checking and custom
809 purpose or trust settings) for functions which only use X509_STORE_CTX
810 internally such as S/MIME.
812 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
813 trust settings if they are not set in X509_STORE. This allows X509_STORE
814 purposes and trust (in S/MIME for example) to override any set by default.
816 Add command line options for CRL checking to smime, s_client and s_server
820 *) Initial CRL based revocation checking. If the CRL checking flag(s)
821 are set then the CRL is looked up in the X509_STORE structure and
822 its validity and signature checked, then if the certificate is found
823 in the CRL the verify fails with a revoked error.
825 Various new CRL related callbacks added to X509_STORE_CTX structure.
827 Command line options added to 'verify' application to support this.
829 This needs some additional work, such as being able to handle multiple
830 CRLs with different times, extension based lookup (rather than just
831 by subject name) and ultimately more complete V2 CRL extension
835 *) Add a general user interface API (crypto/ui/). This is designed
836 to replace things like des_read_password and friends (backward
837 compatibility functions using this new API are provided).
838 The purpose is to remove prompting functions from the DES code
839 section as well as provide for prompting through dialog boxes in
840 a window system and the like.
843 *) Add "ex_data" support to ENGINE so implementations can add state at a
844 per-structure level rather than having to store it globally.
847 *) Make it possible for ENGINE structures to be copied when retrieved by
848 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
849 This causes the "original" ENGINE structure to act like a template,
850 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
851 operational state can be localised to each ENGINE structure, despite the
852 fact they all share the same "methods". New ENGINE structures returned in
853 this case have no functional references and the return value is the single
854 structural reference. This matches the single structural reference returned
855 by ENGINE_by_id() normally, when it is incremented on the pre-existing
859 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
860 needs to match any other type at all we need to manually clear the
864 *) Changes to the "openssl engine" utility to include;
865 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
866 about an ENGINE's available control commands.
867 - executing control commands from command line arguments using the
868 '-pre' and '-post' switches. '-post' is only used if '-t' is
869 specified and the ENGINE is successfully initialised. The syntax for
870 the individual commands are colon-separated, for example;
871 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
874 *) New dynamic control command support for ENGINEs. ENGINEs can now
875 declare their own commands (numbers), names (strings), descriptions,
876 and input types for run-time discovery by calling applications. A
877 subset of these commands are implicitly classed as "executable"
878 depending on their input type, and only these can be invoked through
879 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
880 can be based on user input, config files, etc). The distinction is
881 that "executable" commands cannot return anything other than a boolean
882 result and can only support numeric or string input, whereas some
883 discoverable commands may only be for direct use through
884 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
885 pointers, or other custom uses. The "executable" commands are to
886 support parameterisations of ENGINE behaviour that can be
887 unambiguously defined by ENGINEs and used consistently across any
888 OpenSSL-based application. Commands have been added to all the
889 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
890 control over shared-library paths without source code alterations.
893 *) Changed all ENGINE implementations to dynamically allocate their
894 ENGINEs rather than declaring them statically. Apart from this being
895 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
896 this also allows the implementations to compile without using the
897 internal engine_int.h header.
900 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
901 'const' value. Any code that should be able to modify a RAND_METHOD
902 should already have non-const pointers to it (ie. they should only
903 modify their own ones).
906 *) Made a variety of little tweaks to the ENGINE code.
907 - "atalla" and "ubsec" string definitions were moved from header files
908 to C code. "nuron" string definitions were placed in variables
909 rather than hard-coded - allowing parameterisation of these values
910 later on via ctrl() commands.
911 - Removed unused "#if 0"'d code.
912 - Fixed engine list iteration code so it uses ENGINE_free() to release
913 structural references.
914 - Constified the RAND_METHOD element of ENGINE structures.
915 - Constified various get/set functions as appropriate and added
916 missing functions (including a catch-all ENGINE_cpy that duplicates
917 all ENGINE values onto a new ENGINE except reference counts/state).
918 - Removed NULL parameter checks in get/set functions. Setting a method
919 or function to NULL is a way of cancelling out a previously set
920 value. Passing a NULL ENGINE parameter is just plain stupid anyway
921 and doesn't justify the extra error symbols and code.
922 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
923 flags from engine_int.h to engine.h.
924 - Changed prototypes for ENGINE handler functions (init(), finish(),
925 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
928 *) Implement binary inversion algorithm for BN_mod_inverse in addition
929 to the algorithm using long division. The binary algorithm can be
930 used only if the modulus is odd. On 32-bit systems, it is faster
931 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
932 roughly 5-15% for 256-bit moduli), so we use it only for moduli
933 up to 450 bits. In 64-bit environments, the binary algorithm
934 appears to be advantageous for much longer moduli; here we use it
935 for moduli up to 2048 bits.
938 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
939 could not support the combine flag in choice fields.
942 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
943 extensions from a certificate request to the certificate.
946 *) Allow multiple 'certopt' and 'nameopt' options to be separated
947 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
948 file: this allows the display of the certificate about to be
949 signed to be customised, to allow certain fields to be included
950 or excluded and extension details. The old system didn't display
951 multicharacter strings properly, omitted fields not in the policy
952 and couldn't display additional details such as extensions.
955 *) Function EC_POINTs_mul for multiple scalar multiplication
956 of an arbitrary number of elliptic curve points
957 \sum scalars[i]*points[i],
958 optionally including the generator defined for the EC_GROUP:
959 scalar*generator + \sum scalars[i]*points[i].
961 EC_POINT_mul is a simple wrapper function for the typical case
962 that the point list has just one item (besides the optional
966 *) First EC_METHODs for curves over GF(p):
968 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
969 operations and provides various method functions that can also
970 operate with faster implementations of modular arithmetic.
972 EC_GFp_mont_method() reuses most functions that are part of
973 EC_GFp_simple_method, but uses Montgomery arithmetic.
975 [Bodo Moeller; point addition and point doubling
976 implementation directly derived from source code provided by
977 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
979 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
982 Curves are EC_GROUP objects (with an optional group generator)
983 based on EC_METHODs that are built into the library.
985 Points are EC_POINT objects based on EC_GROUP objects.
987 Most of the framework would be able to handle curves over arbitrary
988 finite fields, but as there are no obvious types for fields other
989 than GF(p), some functions are limited to that for now.
992 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
993 that the file contains a complete HTTP response.
996 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
997 change the def and num file printf format specifier from "%-40sXXX"
998 to "%-39s XXX". The latter will always guarantee a space after the
999 field while the former will cause them to run together if the field
1000 is 40 of more characters long.
1003 *) Constify the cipher and digest 'method' functions and structures
1004 and modify related functions to take constant EVP_MD and EVP_CIPHER
1008 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1009 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1012 *) Modify EVP_Digest*() routines so they now return values. Although the
1013 internal software routines can never fail additional hardware versions
1017 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1019 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1020 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1023 ERR_R_NESTED_ASN1_ERROR
1025 ERR_R_MISSING_ASN1_EOS
1026 were 4 .. 9, conflicting with
1027 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1029 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1030 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1032 Add new error code 'ERR_R_INTERNAL_ERROR'.
1035 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1039 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1040 sets the subject name for a new request or supersedes the
1041 subject name in a given request. Formats that can be parsed are
1042 'CN=Some Name, OU=myOU, C=IT'
1044 'CN=Some Name/OU=myOU/C=IT'.
1046 Add options '-batch' and '-verbose' to 'openssl req'.
1047 [Massimiliano Pala <madwolf@hackmasters.net>]
1049 *) Introduce the possibility to access global variables through
1050 functions on platform were that's the best way to handle exporting
1051 global variables in shared libraries. To enable this functionality,
1052 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1053 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1054 is normally done by Configure or something similar).
1056 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1057 in the source file (foo.c) like this:
1059 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1060 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1062 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1063 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1065 OPENSSL_DECLARE_GLOBAL(int,foo);
1066 #define foo OPENSSL_GLOBAL_REF(foo)
1067 OPENSSL_DECLARE_GLOBAL(double,bar);
1068 #define bar OPENSSL_GLOBAL_REF(bar)
1070 The #defines are very important, and therefore so is including the
1071 header file everywhere where the defined globals are used.
1073 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1074 of ASN.1 items, but that structure is a bit different.
1076 The largest change is in util/mkdef.pl which has been enhanced with
1077 better and easier to understand logic to choose which symbols should
1078 go into the Windows .def files as well as a number of fixes and code
1079 cleanup (among others, algorithm keywords are now sorted
1080 lexicographically to avoid constant rewrites).
1083 *) In BN_div() keep a copy of the sign of 'num' before writing the
1084 result to 'rm' because if rm==num the value will be overwritten
1085 and produce the wrong result if 'num' is negative: this caused
1086 problems with BN_mod() and BN_nnmod().
1089 *) Function OCSP_request_verify(). This checks the signature on an
1090 OCSP request and verifies the signer certificate. The signer
1091 certificate is just checked for a generic purpose and OCSP request
1095 *) Add OCSP_check_validity() function to check the validity of OCSP
1096 responses. OCSP responses are prepared in real time and may only
1097 be a few seconds old. Simply checking that the current time lies
1098 between thisUpdate and nextUpdate max reject otherwise valid responses
1099 caused by either OCSP responder or client clock inaccuracy. Instead
1100 we allow thisUpdate and nextUpdate to fall within a certain period of
1101 the current time. The age of the response can also optionally be
1102 checked. Two new options -validity_period and -status_age added to
1106 *) If signature or public key algorithm is unrecognized print out its
1107 OID rather that just UNKNOWN.
1110 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1111 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1112 ID to be generated from the issuer certificate alone which can then be
1113 passed to OCSP_id_issuer_cmp().
1116 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1117 ASN1 modules to export functions returning ASN1_ITEM pointers
1118 instead of the ASN1_ITEM structures themselves. This adds several
1119 new macros which allow the underlying ASN1 function/structure to
1120 be accessed transparently. As a result code should not use ASN1_ITEM
1121 references directly (such as &X509_it) but instead use the relevant
1122 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1123 use of the new ASN1 code on platforms where exporting structures
1124 is problematical (for example in shared libraries) but exporting
1125 functions returning pointers to structures is not.
1128 *) Add support for overriding the generation of SSL/TLS session IDs.
1129 These callbacks can be registered either in an SSL_CTX or per SSL.
1130 The purpose of this is to allow applications to control, if they wish,
1131 the arbitrary values chosen for use as session IDs, particularly as it
1132 can be useful for session caching in multiple-server environments. A
1133 command-line switch for testing this (and any client code that wishes
1134 to use such a feature) has been added to "s_server".
1135 [Geoff Thorpe, Lutz Jaenicke]
1137 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1138 of the form '#if defined(...) || defined(...) || ...' and
1139 '#if !defined(...) && !defined(...) && ...'. This also avoids
1140 the growing number of special cases it was previously handling.
1143 *) Make all configuration macros available for application by making
1144 sure they are available in opensslconf.h, by giving them names starting
1145 with "OPENSSL_" to avoid conflicts with other packages and by making
1146 sure e_os2.h will cover all platform-specific cases together with
1148 Additionally, it is now possible to define configuration/platform-
1149 specific names (called "system identities"). In the C code, these
1150 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1151 macro with the name beginning with "OPENSSL_SYS_", which is determined
1152 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1156 *) New option -set_serial to 'req' and 'x509' this allows the serial
1157 number to use to be specified on the command line. Previously self
1158 signed certificates were hard coded with serial number 0 and the
1159 CA options of 'x509' had to use a serial number in a file which was
1163 *) New options to 'ca' utility to support V2 CRL entry extensions.
1164 Currently CRL reason, invalidity date and hold instruction are
1165 supported. Add new CRL extensions to V3 code and some new objects.
1168 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1169 disable standard block padding (aka PKCS#5 padding) in the EVP
1170 API, which was previously mandatory. This means that the data is
1171 not padded in any way and so the total length much be a multiple
1172 of the block size, otherwise an error occurs.
1175 *) Initial (incomplete) OCSP SSL support.
1178 *) New function OCSP_parse_url(). This splits up a URL into its host,
1179 port and path components: primarily to parse OCSP URLs. New -url
1180 option to ocsp utility.
1183 *) New nonce behavior. The return value of OCSP_check_nonce() now
1184 reflects the various checks performed. Applications can decide
1185 whether to tolerate certain situations such as an absent nonce
1186 in a response when one was present in a request: the ocsp application
1187 just prints out a warning. New function OCSP_add1_basic_nonce()
1188 this is to allow responders to include a nonce in a response even if
1189 the request is nonce-less.
1192 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1193 skipped when using openssl x509 multiple times on a single input file,
1194 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1197 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1198 set string type: to handle setting ASN1_TIME structures. Fix ca
1199 utility to correctly initialize revocation date of CRLs.
1202 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1203 the clients preferred ciphersuites and rather use its own preferences.
1204 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1205 Internet Explorer by ensuring unchanged hash method during stepup.
1206 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1209 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1210 to aes and add a new 'exist' option to print out symbols that don't
1214 *) Additional options to ocsp utility to allow flags to be set and
1215 additional certificates supplied.
1218 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1219 OCSP client a number of certificate to only verify the response
1223 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1224 handle the new API. Currently only ECB, CBC modes supported. Add new
1227 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1228 Encryption Standard (AES) Ciphersuites for Transport Layer
1229 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1230 not enabled by default and were not part of the "ALL" ciphersuite
1231 alias because they were not yet official; they could be
1232 explicitly requested by specifying the "AESdraft" ciphersuite
1233 group alias. In the final release of OpenSSL 0.9.7, the group
1234 alias is called "AES" and is part of "ALL".)
1235 [Ben Laurie, Steve Henson, Bodo Moeller]
1237 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1238 request to response.
1241 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1242 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1243 extract information from a certificate request. OCSP_response_create()
1244 creates a response and optionally adds a basic response structure.
1245 OCSP_basic_add1_status() adds a complete single response to a basic
1246 response and returns the OCSP_SINGLERESP structure just added (to allow
1247 extensions to be included for example). OCSP_basic_add1_cert() adds a
1248 certificate to a basic response and OCSP_basic_sign() signs a basic
1249 response with various flags. New helper functions ASN1_TIME_check()
1250 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1251 (converts ASN1_TIME to GeneralizedTime).
1254 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1255 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1256 structure from a certificate. X509_pubkey_digest() digests the public_key
1257 contents: this is used in various key identifiers.
1260 *) Make sk_sort() tolerate a NULL argument.
1261 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1263 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1264 passed by the function are trusted implicitly. If any of them signed the
1265 response then it is assumed to be valid and is not verified.
1268 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1269 to data. This was previously part of the PKCS7 ASN1 code. This
1270 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1271 [Steve Henson, reported by Kenneth R. Robinette
1272 <support@securenetterm.com>]
1274 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1275 routines: without these tracing memory leaks is very painful.
1276 Fix leaks in PKCS12 and PKCS7 routines.
1279 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1280 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1281 effectively meant GeneralizedTime would never be used. Now it
1282 is initialised to -1 but X509_time_adj() now has to check the value
1283 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1284 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1285 [Steve Henson, reported by Kenneth R. Robinette
1286 <support@securenetterm.com>]
1288 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1289 result in a zero length in the ASN1_INTEGER structure which was
1290 not consistent with the structure when d2i_ASN1_INTEGER() was used
1291 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1292 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1293 where it did not print out a minus for negative ASN1_INTEGER.
1296 *) Add summary printout to ocsp utility. The various functions which
1297 convert status values to strings have been renamed to:
1298 OCSP_response_status_str(), OCSP_cert_status_str() and
1299 OCSP_crl_reason_str() and are no longer static. New options
1300 to verify nonce values and to disable verification. OCSP response
1301 printout format cleaned up.
1304 *) Add additional OCSP certificate checks. These are those specified
1305 in RFC2560. This consists of two separate checks: the CA of the
1306 certificate being checked must either be the OCSP signer certificate
1307 or the issuer of the OCSP signer certificate. In the latter case the
1308 OCSP signer certificate must contain the OCSP signing extended key
1309 usage. This check is performed by attempting to match the OCSP
1310 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1311 in the OCSP_CERTID structures of the response.
1314 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1315 and related routines. This uses the standard OpenSSL certificate
1316 verify routines to perform initial checks (just CA validity) and
1317 to obtain the certificate chain. Then additional checks will be
1318 performed on the chain. Currently the root CA is checked to see
1319 if it is explicitly trusted for OCSP signing. This is used to set
1320 a root CA as a global signing root: that is any certificate that
1321 chains to that CA is an acceptable OCSP signing certificate.
1324 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1325 extensions from a separate configuration file.
1326 As when reading extensions from the main configuration file,
1327 the '-extensions ...' option may be used for specifying the
1329 [Massimiliano Pala <madwolf@comune.modena.it>]
1331 *) New OCSP utility. Allows OCSP requests to be generated or
1332 read. The request can be sent to a responder and the output
1333 parsed, outputed or printed in text form. Not complete yet:
1334 still needs to check the OCSP response validity.
1337 *) New subcommands for 'openssl ca':
1338 'openssl ca -status <serial>' prints the status of the cert with
1339 the given serial number (according to the index file).
1340 'openssl ca -updatedb' updates the expiry status of certificates
1342 [Massimiliano Pala <madwolf@comune.modena.it>]
1344 *) New '-newreq-nodes' command option to CA.pl. This is like
1345 '-newreq', but calls 'openssl req' with the '-nodes' option
1346 so that the resulting key is not encrypted.
1347 [Damien Miller <djm@mindrot.org>]
1349 *) New configuration for the GNU Hurd.
1350 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1352 *) Initial code to implement OCSP basic response verify. This
1353 is currently incomplete. Currently just finds the signer's
1354 certificate and verifies the signature on the response.
1357 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1358 value of OPENSSLDIR. This is available via the new '-d' option
1359 to 'openssl version', and is also included in 'openssl version -a'.
1362 *) Allowing defining memory allocation callbacks that will be given
1363 file name and line number information in additional arguments
1364 (a const char* and an int). The basic functionality remains, as
1365 well as the original possibility to just replace malloc(),
1366 realloc() and free() by functions that do not know about these
1367 additional arguments. To register and find out the current
1368 settings for extended allocation functions, the following
1369 functions are provided:
1371 CRYPTO_set_mem_ex_functions
1372 CRYPTO_set_locked_mem_ex_functions
1373 CRYPTO_get_mem_ex_functions
1374 CRYPTO_get_locked_mem_ex_functions
1376 These work the same way as CRYPTO_set_mem_functions and friends.
1377 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1378 extended allocation function is enabled.
1379 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1380 a conventional allocation function is enabled.
1381 [Richard Levitte, Bodo Moeller]
1383 *) Finish off removing the remaining LHASH function pointer casts.
1384 There should no longer be any prototype-casting required when using
1385 the LHASH abstraction, and any casts that remain are "bugs". See
1386 the callback types and macros at the head of lhash.h for details
1387 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1390 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1391 If /dev/[u]random devices are not available or do not return enough
1392 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1394 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1395 /etc/entropy will be queried once each in this sequence, quering stops
1396 when enough entropy was collected without querying more sockets.
1399 *) Change the Unix RAND_poll() variant to be able to poll several
1400 random devices, as specified by DEVRANDOM, until a sufficient amount
1401 of data has been collected. We spend at most 10 ms on each file
1402 (select timeout) and read in non-blocking mode. DEVRANDOM now
1403 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1404 (previously it was just the string "/dev/urandom"), so on typical
1405 platforms the 10 ms delay will never occur.
1406 Also separate out the Unix variant to its own file, rand_unix.c.
1407 For VMS, there's a currently-empty rand_vms.c.
1410 *) Move OCSP client related routines to ocsp_cl.c. These
1411 provide utility functions which an application needing
1412 to issue a request to an OCSP responder and analyse the
1413 response will typically need: as opposed to those which an
1414 OCSP responder itself would need which will be added later.
1416 OCSP_request_sign() signs an OCSP request with an API similar
1417 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1418 response. OCSP_response_get1_basic() extracts basic response
1419 from response. OCSP_resp_find_status(): finds and extracts status
1420 information from an OCSP_CERTID structure (which will be created
1421 when the request structure is built). These are built from lower
1422 level functions which work on OCSP_SINGLERESP structures but
1423 wont normally be used unless the application wishes to examine
1424 extensions in the OCSP response for example.
1426 Replace nonce routines with a pair of functions.
1427 OCSP_request_add1_nonce() adds a nonce value and optionally
1428 generates a random value. OCSP_check_nonce() checks the
1429 validity of the nonce in an OCSP response.
1432 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1433 This doesn't copy the supplied OCSP_CERTID and avoids the
1434 need to free up the newly created id. Change return type
1435 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1436 This can then be used to add extensions to the request.
1437 Deleted OCSP_request_new(), since most of its functionality
1438 is now in OCSP_REQUEST_new() (and the case insensitive name
1439 clash) apart from the ability to set the request name which
1440 will be added elsewhere.
1443 *) Update OCSP API. Remove obsolete extensions argument from
1444 various functions. Extensions are now handled using the new
1445 OCSP extension code. New simple OCSP HTTP function which
1446 can be used to send requests and parse the response.
1449 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1450 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1451 uses the special reorder version of SET OF to sort the attributes
1452 and reorder them to match the encoded order. This resolves a long
1453 standing problem: a verify on a PKCS7 structure just after signing
1454 it used to fail because the attribute order did not match the
1455 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1456 it uses the received order. This is necessary to tolerate some broken
1457 software that does not order SET OF. This is handled by encoding
1458 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1459 to produce the required SET OF.
1462 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1463 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1464 files to get correct declarations of the ASN.1 item variables.
1467 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1468 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1469 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1470 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1471 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1472 ASN1_ITEM and no wrapper functions.
1475 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1476 replace the old function pointer based I/O routines. Change most of
1477 the *_d2i_bio() and *_d2i_fp() functions to use these.
1480 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1481 lines, recognice more "algorithms" that can be deselected, and make
1482 it complain about algorithm deselection that isn't recognised.
1485 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1486 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1487 to use new functions. Add NO_ASN1_OLD which can be set to remove
1488 some old style ASN1 functions: this can be used to determine if old
1489 code will still work when these eventually go away.
1492 *) New extension functions for OCSP structures, these follow the
1493 same conventions as certificates and CRLs.
1496 *) New function X509V3_add1_i2d(). This automatically encodes and
1497 adds an extension. Its behaviour can be customised with various
1498 flags to append, replace or delete. Various wrappers added for
1499 certifcates and CRLs.
1502 *) Fix to avoid calling the underlying ASN1 print routine when
1503 an extension cannot be parsed. Correct a typo in the
1504 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1507 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1508 entries for variables.
1511 *) Add functionality to apps/openssl.c for detecting locking
1512 problems: As the program is single-threaded, all we have
1513 to do is register a locking callback using an array for
1514 storing which locks are currently held by the program.
1517 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1518 SSL_get_ex_data_X509_STORE_idx(), which is used in
1519 ssl_verify_cert_chain() and thus can be called at any time
1520 during TLS/SSL handshakes so that thread-safety is essential.
1521 Unfortunately, the ex_data design is not at all suited
1522 for multi-threaded use, so it probably should be abolished.
1525 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1526 [Broadcom, tweaked and integrated by Geoff Thorpe]
1528 *) Move common extension printing code to new function
1529 X509V3_print_extensions(). Reorganise OCSP print routines and
1530 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1533 *) New function X509_signature_print() to remove duplication in some
1537 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1538 set (this was treated exactly the same as SET OF previously). This
1539 is used to reorder the STACK representing the structure to match the
1540 encoding. This will be used to get round a problem where a PKCS7
1541 structure which was signed could not be verified because the STACK
1542 order did not reflect the encoded order.
1545 *) Reimplement the OCSP ASN1 module using the new code.
1548 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1549 for its ASN1 operations. The old style function pointers still exist
1550 for now but they will eventually go away.
1553 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1554 completely replaces the old ASN1 functionality with a table driven
1555 encoder and decoder which interprets an ASN1_ITEM structure describing
1556 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1557 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1558 has also been converted to the new form.
1561 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1562 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1563 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1564 for negative moduli.
1567 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1568 of not touching the result's sign bit.
1571 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1575 *) Changed the LHASH code to use prototypes for callbacks, and created
1576 macros to declare and implement thin (optionally static) functions
1577 that provide type-safety and avoid function pointer casting for the
1578 type-specific callbacks.
1581 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1583 [Veers Staats <staatsvr@asc.hpc.mil>,
1584 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1586 *) Reformat the FAQ so the different questions and answers can be divided
1587 in sections depending on the subject.
1590 *) Have the zlib compression code load ZLIB.DLL dynamically under
1594 *) New function BN_mod_sqrt for computing square roots modulo a prime
1595 (using the probabilistic Tonelli-Shanks algorithm unless
1596 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1597 be handled deterministically).
1598 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1600 *) Make BN_mod_inverse faster by explicitly handling small quotients
1601 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1602 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1605 *) New function BN_kronecker.
1608 *) Fix BN_gcd so that it works on negative inputs; the result is
1609 positive unless both parameters are zero.
1610 Previously something reasonably close to an infinite loop was
1611 possible because numbers could be growing instead of shrinking
1612 in the implementation of Euclid's algorithm.
1615 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1616 sign of the number in question.
1618 Fix BN_is_word(a,w) to work correctly for w == 0.
1620 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1621 because its test if the absolute value of 'a' equals 'w'.
1622 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1623 it exists mostly for use in the implementations of BN_is_zero(),
1624 BN_is_one(), and BN_is_word().
1627 *) New function BN_swap.
1630 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1631 the exponentiation functions are more likely to produce reasonable
1632 results on negative inputs.
1635 *) Change BN_mod_mul so that the result is always non-negative.
1636 Previously, it could be negative if one of the factors was negative;
1637 I don't think anyone really wanted that behaviour.
1640 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1641 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1642 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1643 and add new functions:
1652 BN_mod_lshift1_quick
1656 These functions always generate non-negative results.
1658 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1659 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1661 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1662 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1663 be reduced modulo m.
1664 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1666 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1667 was actually never needed) and in BN_mul(). The removal in BN_mul()
1668 required a small change in bn_mul_part_recursive() and the addition
1669 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1670 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1671 bn_sub_words() and bn_add_words() except they take arrays with
1675 *) In 'openssl passwd', verify passwords read from the terminal
1676 unless the '-salt' option is used (which usually means that
1677 verification would just waste user's time since the resulting
1678 hash is going to be compared with some given password hash)
1679 or the new '-noverify' option is used.
1681 This is an incompatible change, but it does not affect
1682 non-interactive use of 'openssl passwd' (passwords on the command
1683 line, '-stdin' option, '-in ...' option) and thus should not
1687 *) Remove all references to RSAref, since there's no more need for it.
1690 *) Make DSO load along a path given through an environment variable
1691 (SHLIB_PATH) with shl_load().
1694 *) Constify the ENGINE code as a result of BIGNUM constification.
1695 Also constify the RSA code and most things related to it. In a
1696 few places, most notable in the depth of the ASN.1 code, ugly
1697 casts back to non-const were required (to be solved at a later
1701 *) Make it so the openssl application has all engines loaded by default.
1704 *) Constify the BIGNUM routines a little more.
1707 *) Add the following functions:
1709 ENGINE_load_cswift()
1711 ENGINE_load_atalla()
1713 ENGINE_load_builtin_engines()
1715 That way, an application can itself choose if external engines that
1716 are built-in in OpenSSL shall ever be used or not. The benefit is
1717 that applications won't have to be linked with libdl or other dso
1718 libraries unless it's really needed.
1720 Changed 'openssl engine' to load all engines on demand.
1721 Changed the engine header files to avoid the duplication of some
1722 declarations (they differed!).
1725 *) 'openssl engine' can now list capabilities.
1728 *) Better error reporting in 'openssl engine'.
1731 *) Never call load_dh_param(NULL) in s_server.
1734 *) Add engine application. It can currently list engines by name and
1735 identity, and test if they are actually available.
1738 *) Improve RPM specification file by forcing symbolic linking and making
1739 sure the installed documentation is also owned by root.root.
1740 [Damien Miller <djm@mindrot.org>]
1742 *) Give the OpenSSL applications more possibilities to make use of
1743 keys (public as well as private) handled by engines.
1746 *) Add OCSP code that comes from CertCo.
1749 *) Add VMS support for the Rijndael code.
1752 *) Added untested support for Nuron crypto accelerator.
1755 *) Add support for external cryptographic devices. This code was
1756 previously distributed separately as the "engine" branch.
1757 [Geoff Thorpe, Richard Levitte]
1759 *) Rework the filename-translation in the DSO code. It is now possible to
1760 have far greater control over how a "name" is turned into a filename
1761 depending on the operating environment and any oddities about the
1762 different shared library filenames on each system.
1765 *) Support threads on FreeBSD-elf in Configure.
1768 *) Fix for SHA1 assembly problem with MASM: it produces
1769 warnings about corrupt line number information when assembling
1770 with debugging information. This is caused by the overlapping
1772 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1775 NCONF_get_number() has no error checking at all. As a replacement,
1776 NCONF_get_number_e() is defined (_e for "error checking") and is
1777 promoted strongly. The old NCONF_get_number is kept around for
1778 binary backward compatibility.
1779 Make it possible for methods to load from something other than a BIO,
1780 by providing a function pointer that is given a name instead of a BIO.
1781 For example, this could be used to load configuration data from an
1785 *) Fix for non blocking accept BIOs. Added new I/O special reason
1786 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1787 with non blocking I/O was not possible because no retry code was
1788 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1792 *) Added the beginnings of Rijndael support.
1795 *) Fix for bug in DirectoryString mask setting. Add support for
1796 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1797 to allow certificate printing to more controllable, additional
1798 'certopt' option to 'x509' to allow new printing options to be
1802 *) Clean old EAY MD5 hack from e_os.h.
1805 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
1807 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
1808 memory from it's contents. This is done with a counter that will
1809 place alternating values in each byte. This can be used to solve
1810 two issues: 1) the removal of calls to memset() by highly optimizing
1811 compilers, and 2) cleansing with other values than 0, since those can
1812 be read through on certain media, for example a swap space on disk.
1815 *) Bugfix: client side session caching did not work with external caching,
1816 because the session->cipher setting was not restored when reloading
1817 from the external cache. This problem was masked, when
1818 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
1819 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
1822 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
1823 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
1824 [Zeev Lieber <zeev-l@yahoo.com>]
1826 *) Undo an undocumented change introduced in 0.9.6e which caused
1827 repeated calls to OpenSSL_add_all_ciphers() and
1828 OpenSSL_add_all_digests() to be ignored, even after calling
1832 *) Change the default configuration reader to deal with last line not
1833 being properly terminated.
1836 *) Change X509_NAME_cmp() so it applies the special rules on handling
1837 DN values that are of type PrintableString, as well as RDNs of type
1838 emailAddress where the value has the type ia5String.
1839 [stefank@valicert.com via Richard Levitte]
1841 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
1842 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
1843 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
1844 the bitwise-OR of the two for use by the majority of applications
1845 wanting this behaviour, and update the docs. The documented
1846 behaviour and actual behaviour were inconsistent and had been
1847 changing anyway, so this is more a bug-fix than a behavioural
1849 [Geoff Thorpe, diagnosed by Nadav Har'El]
1851 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
1852 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
1855 *) Fix initialization code race conditions in
1856 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
1857 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
1858 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
1859 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
1860 ssl2_get_cipher_by_char(),
1861 ssl3_get_cipher_by_char().
1862 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
1864 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
1865 the cached sessions are flushed, as the remove_cb() might use ex_data
1866 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
1867 (see [openssl.org #212]).
1868 [Geoff Thorpe, Lutz Jaenicke]
1870 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
1871 length, instead of the encoding length to d2i_ASN1_OBJECT.
1874 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
1876 *) [In 0.9.6g-engine release:]
1877 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
1878 [Lynn Gazis <lgazis@rainbow.com>]
1880 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
1882 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
1883 and get fix the header length calculation.
1884 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
1885 Alon Kantor <alonk@checkpoint.com> (and others),
1888 *) Use proper error handling instead of 'assertions' in buffer
1889 overflow checks added in 0.9.6e. This prevents DoS (the
1890 assertions could call abort()).
1891 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
1893 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
1895 *) Add various sanity checks to asn1_get_length() to reject
1896 the ASN1 length bytes if they exceed sizeof(long), will appear
1897 negative or the content length exceeds the length of the
1899 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
1901 *) Fix cipher selection routines: ciphers without encryption had no flags
1902 for the cipher strength set and where therefore not handled correctly
1903 by the selection routines (PR #130).
1906 *) Fix EVP_dsa_sha macro.
1910 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
1911 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
1912 that was added in OpenSSL 0.9.6d.
1914 As the countermeasure turned out to be incompatible with some
1915 broken SSL implementations, the new option is part of SSL_OP_ALL.
1916 SSL_OP_ALL is usually employed when compatibility with weird SSL
1917 implementations is desired (e.g. '-bugs' option to 's_client' and
1918 's_server'), so the new option is automatically set in many
1922 *) Changes in security patch:
1924 Changes marked "(CHATS)" were sponsored by the Defense Advanced
1925 Research Projects Agency (DARPA) and Air Force Research Laboratory,
1926 Air Force Materiel Command, USAF, under agreement number
1929 *) Add various sanity checks to asn1_get_length() to reject
1930 the ASN1 length bytes if they exceed sizeof(long), will appear
1931 negative or the content length exceeds the length of the
1932 supplied buffer. (CAN-2002-0659)
1933 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
1935 *) Assertions for various potential buffer overflows, not known to
1937 [Ben Laurie (CHATS)]
1939 *) Various temporary buffers to hold ASCII versions of integers were
1940 too small for 64 bit platforms. (CAN-2002-0655)
1941 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
1943 *) Remote buffer overflow in SSL3 protocol - an attacker could
1944 supply an oversized session ID to a client. (CAN-2002-0656)
1945 [Ben Laurie (CHATS)]
1947 *) Remote buffer overflow in SSL2 protocol - an attacker could
1948 supply an oversized client master key. (CAN-2002-0656)
1949 [Ben Laurie (CHATS)]
1951 Changes between 0.9.6c and 0.9.6d [9 May 2002]
1953 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
1954 encoded as NULL) with id-dsa-with-sha1.
1955 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
1957 *) Check various X509_...() return values in apps/req.c.
1958 [Nils Larsch <nla@trustcenter.de>]
1960 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
1961 an end-of-file condition would erronously be flagged, when the CRLF
1962 was just at the end of a processed block. The bug was discovered when
1963 processing data through a buffering memory BIO handing the data to a
1964 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
1965 <ptsekov@syntrex.com> and Nedelcho Stanev.
1968 *) Implement a countermeasure against a vulnerability recently found
1969 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
1970 before application data chunks to avoid the use of known IVs
1971 with data potentially chosen by the attacker.
1974 *) Fix length checks in ssl3_get_client_hello().
1977 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
1978 to prevent ssl3_read_internal() from incorrectly assuming that
1979 ssl3_read_bytes() found application data while handshake
1980 processing was enabled when in fact s->s3->in_read_app_data was
1981 merely automatically cleared during the initial handshake.
1982 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
1984 *) Fix object definitions for Private and Enterprise: they were not
1985 recognized in their shortname (=lowercase) representation. Extend
1986 obj_dat.pl to issue an error when using undefined keywords instead
1987 of silently ignoring the problem (Svenning Sorensen
1988 <sss@sss.dnsalias.net>).
1991 *) Fix DH_generate_parameters() so that it works for 'non-standard'
1992 generators, i.e. generators other than 2 and 5. (Previously, the
1993 code did not properly initialise the 'add' and 'rem' values to
1994 BN_generate_prime().)
1996 In the new general case, we do not insist that 'generator' is
1997 actually a primitive root: This requirement is rather pointless;
1998 a generator of the order-q subgroup is just as good, if not
2002 *) Map new X509 verification errors to alerts. Discovered and submitted by
2003 Tom Wu <tom@arcot.com>.
2006 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2007 returning non-zero before the data has been completely received
2008 when using non-blocking I/O.
2009 [Bodo Moeller; problem pointed out by John Hughes]
2011 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2012 [Ben Laurie, Lutz Jaenicke]
2014 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2015 Yoram Zahavi <YoramZ@gilian.com>).
2018 *) Add information about CygWin 1.3 and on, and preserve proper
2019 configuration for the versions before that.
2020 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2022 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2023 check whether we deal with a copy of a session and do not delete from
2024 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2025 <izhar@checkpoint.com>.
2028 *) Do not store session data into the internal session cache, if it
2029 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2030 flag is set). Proposed by Aslam <aslam@funk.com>.
2033 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2037 *) [In 0.9.6d-engine release:]
2038 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2039 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2041 *) Add the configuration target linux-s390x.
2042 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2044 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2045 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2046 variable as an indication that a ClientHello message has been
2047 received. As the flag value will be lost between multiple
2048 invocations of ssl3_accept when using non-blocking I/O, the
2049 function may not be aware that a handshake has actually taken
2050 place, thus preventing a new session from being added to the
2053 To avoid this problem, we now set s->new_session to 2 instead of
2054 using a local variable.
2055 [Lutz Jaenicke, Bodo Moeller]
2057 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2058 if the SSL_R_LENGTH_MISMATCH error is detected.
2059 [Geoff Thorpe, Bodo Moeller]
2061 *) New 'shared_ldflag' column in Configure platform table.
2064 *) Fix EVP_CIPHER_mode macro.
2065 ["Dan S. Camper" <dan@bti.net>]
2067 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2068 type, we must throw them away by setting rr->length to 0.
2069 [D P Chang <dpc@qualys.com>]
2071 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2073 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2074 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2075 worked incorrectly for those cases where range = 10..._2 and
2076 3*range is two bits longer than range.)
2079 *) Only add signing time to PKCS7 structures if it is not already
2083 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2084 OBJ_ld_ce should be OBJ_id_ce.
2085 Also some ip-pda OIDs in crypto/objects/objects.txt were
2086 incorrect (cf. RFC 3039).
2087 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2089 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2090 returns early because it has nothing to do.
2091 [Andy Schneider <andy.schneider@bjss.co.uk>]
2093 *) [In 0.9.6c-engine release:]
2094 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2095 [Andy Schneider <andy.schneider@bjss.co.uk>]
2097 *) [In 0.9.6c-engine release:]
2098 Add support for Cryptographic Appliance's keyserver technology.
2099 (Use engine 'keyclient')
2100 [Cryptographic Appliances and Geoff Thorpe]
2102 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2103 is called via tools/c89.sh because arguments have to be
2104 rearranged (all '-L' options must appear before the first object
2106 [Richard Shapiro <rshapiro@abinitio.com>]
2108 *) [In 0.9.6c-engine release:]
2109 Add support for Broadcom crypto accelerator cards, backported
2111 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2113 *) [In 0.9.6c-engine release:]
2114 Add support for SureWare crypto accelerator cards from
2115 Baltimore Technologies. (Use engine 'sureware')
2116 [Baltimore Technologies and Mark Cox]
2118 *) [In 0.9.6c-engine release:]
2119 Add support for crypto accelerator cards from Accelerated
2120 Encryption Processing, www.aep.ie. (Use engine 'aep')
2121 [AEP Inc. and Mark Cox]
2123 *) Add a configuration entry for gcc on UnixWare.
2124 [Gary Benson <gbenson@redhat.com>]
2126 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2127 messages are stored in a single piece (fixed-length part and
2128 variable-length part combined) and fix various bugs found on the way.
2131 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2132 instead. BIO_gethostbyname() does not know what timeouts are
2133 appropriate, so entries would stay in cache even when they have
2135 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2137 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2138 faced with a pathologically small ClientHello fragment that does
2139 not contain client_version: Instead of aborting with an error,
2140 simply choose the highest available protocol version (i.e.,
2141 TLS 1.0 unless it is disabled). In practice, ClientHello
2142 messages are never sent like this, but this change gives us
2143 strictly correct behaviour at least for TLS.
2146 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2147 never resets s->method to s->ctx->method when called from within
2148 one of the SSL handshake functions.
2149 [Bodo Moeller; problem pointed out by Niko Baric]
2151 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2152 (sent using the client's version number) if client_version is
2153 smaller than the protocol version in use. Also change
2154 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2155 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2156 the client will at least see that alert.
2159 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2163 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2164 client receives HelloRequest while in a handshake.
2165 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2167 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2168 should end in 'break', not 'goto end' which circuments various
2169 cleanups done in state SSL_ST_OK. But session related stuff
2170 must be disabled for SSL_ST_OK in the case that we just sent a
2173 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2174 before just sending a HelloRequest.
2175 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2177 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2178 reveal whether illegal block cipher padding was found or a MAC
2179 verification error occured. (Neither SSLerr() codes nor alerts
2180 are directly visible to potential attackers, but the information
2181 may leak via logfiles.)
2183 Similar changes are not required for the SSL 2.0 implementation
2184 because the number of padding bytes is sent in clear for SSL 2.0,
2185 and the extra bytes are just ignored. However ssl/s2_pkt.c
2186 failed to verify that the purported number of padding bytes is in
2190 *) Add OpenUNIX-8 support including shared libraries
2191 (Boyd Lynn Gerber <gerberb@zenez.com>).
2194 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2195 'wristwatch attack' using huge encoding parameters (cf.
2196 James H. Manger's CRYPTO 2001 paper). Note that the
2197 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2198 encoding parameters and hence was not vulnerable.
2201 *) BN_sqr() bug fix.
2202 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2204 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2205 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2206 followed by modular reduction.
2207 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2209 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2210 equivalent based on BN_pseudo_rand() instead of BN_rand().
2213 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2214 This function was broken, as the check for a new client hello message
2215 to handle SGC did not allow these large messages.
2216 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2219 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2222 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2223 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2226 *) Rework the configuration and shared library support for Tru64 Unix.
2227 The configuration part makes use of modern compiler features and
2228 still retains old compiler behavior for those that run older versions
2229 of the OS. The shared library support part includes a variant that
2230 uses the RPATH feature, and is available through the special
2231 configuration target "alpha-cc-rpath", which will never be selected
2233 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2235 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2236 with the same message size as in ssl3_get_certificate_request().
2237 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2238 messages might inadvertently be reject as too long.
2239 [Petr Lampa <lampa@fee.vutbr.cz>]
2241 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2244 *) Modified SSL library such that the verify_callback that has been set
2245 specificly for an SSL object with SSL_set_verify() is actually being
2246 used. Before the change, a verify_callback set with this function was
2247 ignored and the verify_callback() set in the SSL_CTX at the time of
2248 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2249 to allow the necessary settings.
2252 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2253 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2254 done automatically (in contradiction to the requirements of the C
2255 standard). This made problems when used from OpenSSH.
2258 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2259 dh->length and always used
2261 BN_rand_range(priv_key, dh->p).
2263 BN_rand_range() is not necessary for Diffie-Hellman, and this
2264 specific range makes Diffie-Hellman unnecessarily inefficient if
2265 dh->length (recommended exponent length) is much smaller than the
2266 length of dh->p. We could use BN_rand_range() if the order of
2267 the subgroup was stored in the DH structure, but we only have
2272 BN_rand(priv_key, l, ...)
2274 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2280 RSA_eay_public_encrypt
2281 RSA_eay_private_decrypt
2282 RSA_eay_private_encrypt (signing)
2283 RSA_eay_public_decrypt (signature verification)
2285 (default implementations for RSA_public_encrypt,
2286 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2287 always reject numbers >= n.
2290 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2291 to synchronize access to 'locking_thread'. This is necessary on
2292 systems where access to 'locking_thread' (an 'unsigned long'
2293 variable) is not atomic.
2296 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2297 *before* setting the 'crypto_lock_rand' flag. The previous code had
2298 a race condition if 0 is a valid thread ID.
2299 [Travis Vitek <vitek@roguewave.com>]
2301 *) Add support for shared libraries under Irix.
2302 [Albert Chin-A-Young <china@thewrittenword.com>]
2304 *) Add configuration option to build on Linux on both big-endian and
2306 [Ralf Baechle <ralf@uni-koblenz.de>]
2308 *) Add the possibility to create shared libraries on HP-UX.
2311 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2313 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2314 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2315 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2316 PRNG state recovery was possible based on the output of
2317 one PRNG request appropriately sized to gain knowledge on
2318 'md' followed by enough consecutive 1-byte PRNG requests
2319 to traverse all of 'state'.
2321 1. When updating 'md_local' (the current thread's copy of 'md')
2322 during PRNG output generation, hash all of the previous
2323 'md_local' value, not just the half used for PRNG output.
2325 2. Make the number of bytes from 'state' included into the hash
2326 independent from the number of PRNG bytes requested.
2328 The first measure alone would be sufficient to avoid
2329 Markku-Juhani's attack. (Actually it had never occurred
2330 to me that the half of 'md_local' used for chaining was the
2331 half from which PRNG output bytes were taken -- I had always
2332 assumed that the secret half would be used.) The second
2333 measure makes sure that additional data from 'state' is never
2334 mixed into 'md_local' in small portions; this heuristically
2335 further strengthens the PRNG.
2338 *) Fix crypto/bn/asm/mips3.s.
2341 *) When only the key is given to "enc", the IV is undefined. Print out
2342 an error message in this case.
2345 *) Handle special case when X509_NAME is empty in X509 printing routines.
2348 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2349 positive and less than q.
2352 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2353 used: it isn't thread safe and the add_lock_callback should handle
2355 [Paul Rose <Paul.Rose@bridge.com>]
2357 *) Verify that incoming data obeys the block size in
2358 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2362 [Ulf Möller, Bodo Möller]
2364 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2365 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2366 when fixing the server behaviour for backwards-compatible 'client
2367 hello' messages. (Note that the attack is impractical against
2368 SSL 3.0 and TLS 1.0 anyway because length and version checking
2369 means that the probability of guessing a valid ciphertext is
2370 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2373 Before 0.9.5, the countermeasure (hide the error by generating a
2374 random 'decryption result') did not work properly because
2375 ERR_clear_error() was missing, meaning that SSL_get_error() would
2376 detect the supposedly ignored error.
2378 Both problems are now fixed.
2381 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2382 (previously it was 1024).
2385 *) Fix for compatibility mode trust settings: ignore trust settings
2386 unless some valid trust or reject settings are present.
2389 *) Fix for blowfish EVP: its a variable length cipher.
2392 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2393 parameters in DSA public key structures and return an error in the
2394 DSA routines if parameters are absent.
2397 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2398 in the current directory if neither $RANDFILE nor $HOME was set.
2399 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2400 caused some confusion to Windows users who haven't defined $HOME.
2401 Thus RAND_file_name() is changed again: e_os.h can define a
2402 DEFAULT_HOME, which will be used if $HOME is not set.
2403 For Windows, we use "C:"; on other platforms, we still require
2404 environment variables.
2406 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2407 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2408 having multiple threads call RAND_poll() concurrently.
2411 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2412 combination of a flag and a thread ID variable.
2413 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2414 flag), *other* threads can enter ssleay_add_bytes without obeying
2415 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2416 that they do not hold after the first thread unsets add_do_not_lock).
2419 *) Change bctest again: '-x' expressions are not available in all
2423 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2425 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2426 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2428 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2429 the default extension for executables, if any. Also, make the perl
2430 scripts that use symlink() to test if it really exists and use "cp"
2431 if it doesn't. All this made OpenSSL compilable and installable in
2435 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2436 If SEQUENCE is length is indefinite just set c->slen to the total
2437 amount of data available.
2438 [Steve Henson, reported by shige@FreeBSD.org]
2439 [This change does not apply to 0.9.7.]
2441 *) Change bctest to avoid here-documents inside command substitution
2442 (workaround for FreeBSD /bin/sh bug).
2443 For compatibility with Ultrix, avoid shell functions (introduced
2444 in the bctest version that searches along $PATH).
2447 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2448 with des_encrypt() defined on some operating systems, like Solaris
2452 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2453 On the Importance of Eliminating Errors in Cryptographic
2454 Computations, J. Cryptology 14 (2001) 2, 101-119,
2455 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2458 *) MIPS assembler BIGNUM division bug fix.
2461 *) Disabled incorrect Alpha assembler code.
2464 *) Fix PKCS#7 decode routines so they correctly update the length
2465 after reading an EOC for the EXPLICIT tag.
2467 [This change does not apply to 0.9.7.]
2469 *) Fix bug in PKCS#12 key generation routines. This was triggered
2470 if a 3DES key was generated with a 0 initial byte. Include
2471 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2472 (but broken) behaviour.
2475 *) Enhance bctest to search for a working bc along $PATH and print
2477 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2479 *) Fix memory leaks in err.c: free err_data string if necessary;
2480 don't write to the wrong index in ERR_set_error_data.
2483 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2487 *) Replace rdtsc with _emit statements for VC++ version 5.
2488 [Jeremy Cooper <jeremy@baymoo.org>]
2490 *) Make it possible to reuse SSLv2 sessions.
2493 *) In copy_email() check for >= 0 as a return value for
2494 X509_NAME_get_index_by_NID() since 0 is a valid index.
2495 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2497 *) Avoid coredump with unsupported or invalid public keys by checking if
2498 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2499 PKCS7_verify() fails with non detached data.
2502 *) Don't use getenv in library functions when run as setuid/setgid.
2503 New function OPENSSL_issetugid().
2506 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2507 due to incorrect handling of multi-threading:
2509 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2511 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2513 3. Count how many times MemCheck_off() has been called so that
2514 nested use can be treated correctly. This also avoids
2515 inband-signalling in the previous code (which relied on the
2516 assumption that thread ID 0 is impossible).
2519 *) Add "-rand" option also to s_client and s_server.
2522 *) Fix CPU detection on Irix 6.x.
2523 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2524 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2526 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2529 [This change does not apply to 0.9.7.]
2531 *) Use the cached encoding of an X509_NAME structure rather than
2532 copying it. This is apparently the reason for the libsafe "errors"
2533 but the code is actually correct.
2536 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2537 Bleichenbacher's DSA attack.
2538 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2539 to be set and top=0 forces the highest bit to be set; top=-1 is new
2540 and leaves the highest bit random.
2541 [Ulf Moeller, Bodo Moeller]
2543 *) In the NCONF_...-based implementations for CONF_... queries
2544 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2545 a temporary CONF structure with the data component set to NULL
2546 (which gives segmentation faults in lh_retrieve).
2547 Instead, use NULL for the CONF pointer in CONF_get_string and
2548 CONF_get_number (which may use environment variables) and directly
2549 return NULL from CONF_get_section.
2552 *) Fix potential buffer overrun for EBCDIC.
2555 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2556 keyUsage if basicConstraints absent for a CA.
2559 *) Make SMIME_write_PKCS7() write mail header values with a format that
2560 is more generally accepted (no spaces before the semicolon), since
2561 some programs can't parse those values properly otherwise. Also make
2562 sure BIO's that break lines after each write do not create invalid
2566 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2567 macros previously used would not encode an empty SEQUENCE OF
2568 and break the signature.
2570 [This change does not apply to 0.9.7.]
2572 *) Zero the premaster secret after deriving the master secret in
2576 *) Add some EVP_add_digest_alias registrations (as found in
2577 OpenSSL_add_all_digests()) to SSL_library_init()
2578 aka OpenSSL_add_ssl_algorithms(). This provides improved
2579 compatibility with peers using X.509 certificates
2580 with unconventional AlgorithmIdentifier OIDs.
2583 *) Fix for Irix with NO_ASM.
2584 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2586 *) ./config script fixes.
2587 [Ulf Moeller, Richard Levitte]
2589 *) Fix 'openssl passwd -1'.
2592 *) Change PKCS12_key_gen_asc() so it can cope with non null
2593 terminated strings whose length is passed in the passlen
2594 parameter, for example from PEM callbacks. This was done
2595 by adding an extra length parameter to asc2uni().
2596 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2598 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2599 call failed, free the DSA structure.
2602 *) Fix to uni2asc() to cope with zero length Unicode strings.
2603 These are present in some PKCS#12 files.
2606 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2607 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2608 when writing a 32767 byte record.
2609 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2611 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2612 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2614 (RSA objects have a reference count access to which is protected
2615 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2616 so they are meant to be shared between threads.)
2617 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2618 "Reddie, Steven" <Steven.Reddie@ca.com>]
2620 *) Fix a deadlock in CRYPTO_mem_leaks().
2623 *) Use better test patterns in bntest.
2626 *) rand_win.c fix for Borland C.
2629 *) BN_rshift bugfix for n == 0.
2632 *) Add a 'bctest' script that checks for some known 'bc' bugs
2633 so that 'make test' does not abort just because 'bc' is broken.
2636 *) Store verify_result within SSL_SESSION also for client side to
2637 avoid potential security hole. (Re-used sessions on the client side
2638 always resulted in verify_result==X509_V_OK, not using the original
2639 result of the server certificate verification.)
2642 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2643 SSL3_RT_APPLICATION_DATA, return 0.
2644 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2648 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2649 releases, have been re-implemented by renaming the previous
2650 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2651 and ssl3_read_internal, respectively, and adding 'peek' parameters
2652 to them. The new ssl[23]_{read,peek} functions are calls to
2653 ssl[23]_read_internal with the 'peek' flag set appropriately.
2654 A 'peek' parameter has also been added to ssl3_read_bytes, which
2655 does the actual work for ssl3_read_internal.
2658 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2659 the method-specific "init()" handler. Also clean up ex_data after
2660 calling the method-specific "finish()" handler. Previously, this was
2661 happening the other way round.
2664 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2665 The previous value, 12, was not always sufficient for BN_mod_exp().
2668 *) Make sure that shared libraries get the internal name engine with
2669 the full version number and not just 0. This should mark the
2670 shared libraries as not backward compatible. Of course, this should
2671 be changed again when we can guarantee backward binary compatibility.
2674 *) Fix typo in get_cert_by_subject() in by_dir.c
2675 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2677 *) Rework the system to generate shared libraries:
2679 - Make note of the expected extension for the shared libraries and
2680 if there is a need for symbolic links from for example libcrypto.so.0
2681 to libcrypto.so.0.9.7. There is extended info in Configure for
2684 - Make as few rebuilds of the shared libraries as possible.
2686 - Still avoid linking the OpenSSL programs with the shared libraries.
2688 - When installing, install the shared libraries separately from the
2692 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2694 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2695 and not in SSL_clear because the latter is also used by the
2696 accept/connect functions; previously, the settings made by
2697 SSL_set_read_ahead would be lost during the handshake.
2698 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2700 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2701 Previously, it would create entries for disableed algorithms no
2705 *) Added several new manual pages for SSL_* function.
2708 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2710 *) In ssl23_get_client_hello, generate an error message when faced
2711 with an initial SSL 3.0/TLS record that is too small to contain the
2712 first two bytes of the ClientHello message, i.e. client_version.
2713 (Note that this is a pathologic case that probably has never happened
2714 in real life.) The previous approach was to use the version number
2715 from the record header as a substitute; but our protocol choice
2716 should not depend on that one because it is not authenticated
2717 by the Finished messages.
2720 *) More robust randomness gathering functions for Windows.
2721 [Jeffrey Altman <jaltman@columbia.edu>]
2723 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2724 not set then we don't setup the error code for issuer check errors
2725 to avoid possibly overwriting other errors which the callback does
2726 handle. If an application does set the flag then we assume it knows
2727 what it is doing and can handle the new informational codes
2731 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2732 a general "ANY" type, as such it should be able to decode anything
2733 including tagged types. However it didn't check the class so it would
2734 wrongly interpret tagged types in the same way as their universal
2735 counterpart and unknown types were just rejected. Changed so that the
2736 tagged and unknown types are handled in the same way as a SEQUENCE:
2737 that is the encoding is stored intact. There is also a new type
2738 "V_ASN1_OTHER" which is used when the class is not universal, in this
2739 case we have no idea what the actual type is so we just lump them all
2743 *) On VMS, stdout may very well lead to a file that is written to
2744 in a record-oriented fashion. That means that every write() will
2745 write a separate record, which will be read separately by the
2746 programs trying to read from it. This can be very confusing.
2748 The solution is to put a BIO filter in the way that will buffer
2749 text until a linefeed is reached, and then write everything a
2750 line at a time, so every record written will be an actual line,
2751 not chunks of lines and not (usually doesn't happen, but I've
2752 seen it once) several lines in one record. BIO_f_linebuffer() is
2755 Currently, it's a VMS-only method, because that's where it has
2756 been tested well enough.
2759 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2760 it can return incorrect results.
2761 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2762 but it was in 0.9.6-beta[12].)
2765 *) Disable the check for content being present when verifying detached
2766 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2767 include zero length content when signing messages.
2770 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2771 BIO_ctrl (for BIO pairs).
2774 *) Add DSO method for VMS.
2777 *) Bug fix: Montgomery multiplication could produce results with the
2781 *) Add RPM specification openssl.spec and modify it to build three
2782 packages. The default package contains applications, application
2783 documentation and run-time libraries. The devel package contains
2784 include files, static libraries and function documentation. The
2785 doc package contains the contents of the doc directory. The original
2786 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2789 *) Add a large number of documentation files for many SSL routines.
2790 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2792 *) Add a configuration entry for Sony News 4.
2793 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2795 *) Don't set the two most significant bits to one when generating a
2796 random number < q in the DSA library.
2799 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2800 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2801 the underlying transport is blocking) if a handshake took place.
2802 (The default behaviour is needed by applications such as s_client
2803 and s_server that use select() to determine when to use SSL_read;
2804 but for applications that know in advance when to expect data, it
2805 just makes things more complicated.)
2808 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2812 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2813 work better on such systems.
2814 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2816 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2817 Update PKCS12_parse() so it copies the friendlyName and the
2818 keyid to the certificates aux info.
2821 *) Fix bug in PKCS7_verify() which caused an infinite loop
2822 if there was more than one signature.
2823 [Sven Uszpelkat <su@celocom.de>]
2825 *) Major change in util/mkdef.pl to include extra information
2826 about each symbol, as well as presentig variables as well
2827 as functions. This change means that there's n more need
2828 to rebuild the .num files when some algorithms are excluded.
2831 *) Allow the verify time to be set by an application,
2832 rather than always using the current time.
2835 *) Phase 2 verify code reorganisation. The certificate
2836 verify code now looks up an issuer certificate by a
2837 number of criteria: subject name, authority key id
2838 and key usage. It also verifies self signed certificates
2839 by the same criteria. The main comparison function is
2840 X509_check_issued() which performs these checks.
2842 Lot of changes were necessary in order to support this
2843 without completely rewriting the lookup code.
2845 Authority and subject key identifier are now cached.
2847 The LHASH 'certs' is X509_STORE has now been replaced
2848 by a STACK_OF(X509_OBJECT). This is mainly because an
2849 LHASH can't store or retrieve multiple objects with
2850 the same hash value.
2852 As a result various functions (which were all internal
2853 use only) have changed to handle the new X509_STORE
2854 structure. This will break anything that messed round
2855 with X509_STORE internally.
2857 The functions X509_STORE_add_cert() now checks for an
2858 exact match, rather than just subject name.
2860 The X509_STORE API doesn't directly support the retrieval
2861 of multiple certificates matching a given criteria, however
2862 this can be worked round by performing a lookup first
2863 (which will fill the cache with candidate certificates)
2864 and then examining the cache for matches. This is probably
2865 the best we can do without throwing out X509_LOOKUP
2866 entirely (maybe later...).
2868 The X509_VERIFY_CTX structure has been enhanced considerably.
2870 All certificate lookup operations now go via a get_issuer()
2871 callback. Although this currently uses an X509_STORE it
2872 can be replaced by custom lookups. This is a simple way
2873 to bypass the X509_STORE hackery necessary to make this
2874 work and makes it possible to use more efficient techniques
2875 in future. A very simple version which uses a simple
2876 STACK for its trusted certificate store is also provided
2877 using X509_STORE_CTX_trusted_stack().
2879 The verify_cb() and verify() callbacks now have equivalents
2880 in the X509_STORE_CTX structure.
2882 X509_STORE_CTX also has a 'flags' field which can be used
2883 to customise the verify behaviour.
2886 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2887 excludes S/MIME capabilities.
2890 *) When a certificate request is read in keep a copy of the
2891 original encoding of the signed data and use it when outputing
2892 again. Signatures then use the original encoding rather than
2893 a decoded, encoded version which may cause problems if the
2894 request is improperly encoded.
2897 *) For consistency with other BIO_puts implementations, call
2898 buffer_write(b, ...) directly in buffer_puts instead of calling
2901 In BIO_puts, increment b->num_write as in BIO_write.
2902 [Peter.Sylvester@EdelWeb.fr]
2904 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2905 BN_zero, we may not return a BIGNUM with an array consisting of
2909 *) Avoid calling abort() from within the library when problems are
2910 detected, except if preprocessor symbols have been defined
2911 (such as REF_CHECK, BN_DEBUG etc.).
2914 *) New openssl application 'rsautl'. This utility can be
2915 used for low level RSA operations. DER public key
2916 BIO/fp routines also added.
2919 *) New Configure entry and patches for compiling on QNX 4.
2920 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2922 *) A demo state-machine implementation was sponsored by
2923 Nuron (http://www.nuron.com/) and is now available in
2924 demos/state_machine.
2927 *) New options added to the 'dgst' utility for signature
2928 generation and verification.
2931 *) Unrecognized PKCS#7 content types are now handled via a
2932 catch all ASN1_TYPE structure. This allows unsupported
2933 types to be stored as a "blob" and an application can
2934 encode and decode it manually.
2937 *) Fix various signed/unsigned issues to make a_strex.c
2939 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2941 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2942 length if passed a buffer. ASN1_INTEGER_to_BN failed
2943 if passed a NULL BN and its argument was negative.
2944 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2946 *) Modification to PKCS#7 encoding routines to output definite
2947 length encoding. Since currently the whole structures are in
2948 memory there's not real point in using indefinite length
2949 constructed encoding. However if OpenSSL is compiled with
2950 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2953 *) Added BIO_vprintf() and BIO_vsnprintf().
2956 *) Added more prefixes to parse for in the the strings written
2957 through a logging bio, to cover all the levels that are available
2958 through syslog. The prefixes are now:
2960 PANIC, EMERG, EMR => LOG_EMERG
2961 ALERT, ALR => LOG_ALERT
2962 CRIT, CRI => LOG_CRIT
2963 ERROR, ERR => LOG_ERR
2964 WARNING, WARN, WAR => LOG_WARNING
2965 NOTICE, NOTE, NOT => LOG_NOTICE
2966 INFO, INF => LOG_INFO
2967 DEBUG, DBG => LOG_DEBUG
2969 and as before, if none of those prefixes are present at the
2970 beginning of the string, LOG_ERR is chosen.
2972 On Win32, the LOG_* levels are mapped according to this:
2974 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2975 LOG_WARNING => EVENTLOG_WARNING_TYPE
2976 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2980 *) Made it possible to reconfigure with just the configuration
2981 argument "reconf" or "reconfigure". The command line arguments
2982 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2983 and are retrieved from there when reconfiguring.
2987 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2989 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2992 *) The obj_dat.pl script was messing up the sorting of object
2993 names. The reason was that it compared the quoted version
2994 of strings as a result "OCSP" > "OCSP Signing" because
2995 " > SPACE. Changed script to store unquoted versions of
2996 names and add quotes on output. It was also omitting some
2997 names from the lookup table if they were given a default
2998 value (that is if SN is missing it is given the same
2999 value as LN and vice versa), these are now added on the
3000 grounds that if an object has a name we should be able to
3001 look it up. Finally added warning output when duplicate
3002 short or long names are found.
3005 *) Changes needed for Tandem NSK.
3006 [Scott Uroff <scott@xypro.com>]
3008 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3009 RSA_padding_check_SSLv23(), special padding was never detected
3010 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3011 version rollback attacks was not effective.
3013 In s23_clnt.c, don't use special rollback-attack detection padding
3014 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3015 client; similarly, in s23_srvr.c, don't do the rollback check if
3016 SSL 2.0 is the only protocol enabled in the server.
3019 *) Make it possible to get hexdumps of unprintable data with 'openssl
3020 asn1parse'. By implication, the functions ASN1_parse_dump() and
3021 BIO_dump_indent() are added.
3024 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3025 these print out strings and name structures based on various
3026 flags including RFC2253 support and proper handling of
3027 multibyte characters. Added options to the 'x509' utility
3028 to allow the various flags to be set.
3031 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3032 Also change the functions X509_cmp_current_time() and
3033 X509_gmtime_adj() work with an ASN1_TIME structure,
3034 this will enable certificates using GeneralizedTime in validity
3035 dates to be checked.
3038 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3039 negative public key encodings) on by default,
3040 NO_NEG_PUBKEY_BUG can be set to disable it.
3043 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3044 content octets. An i2c_ASN1_OBJECT is unnecessary because
3045 the encoding can be trivially obtained from the structure.
3048 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3049 not read locks (CRYPTO_r_[un]lock).
3052 *) A first attempt at creating official support for shared
3053 libraries through configuration. I've kept it so the
3054 default is static libraries only, and the OpenSSL programs
3055 are always statically linked for now, but there are
3056 preparations for dynamic linking in place.
3057 This has been tested on Linux and Tru64.
3060 *) Randomness polling function for Win9x, as described in:
3061 Peter Gutmann, Software Generation of Practically Strong
3065 *) Fix so PRNG is seeded in req if using an already existing
3069 *) New options to smime application. -inform and -outform
3070 allow alternative formats for the S/MIME message including
3071 PEM and DER. The -content option allows the content to be
3072 specified separately. This should allow things like Netscape
3073 form signing output easier to verify.
3076 *) Fix the ASN1 encoding of tags using the 'long form'.
3079 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3080 STRING types. These convert content octets to and from the
3081 underlying type. The actual tag and length octets are
3082 already assumed to have been read in and checked. These
3083 are needed because all other string types have virtually
3084 identical handling apart from the tag. By having versions
3085 of the ASN1 functions that just operate on content octets
3086 IMPLICIT tagging can be handled properly. It also allows
3087 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3088 and ASN1_INTEGER are identical apart from the tag.
3091 *) Change the handling of OID objects as follows:
3093 - New object identifiers are inserted in objects.txt, following
3094 the syntax given in objects.README.
3095 - objects.pl is used to process obj_mac.num and create a new
3097 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3100 This is currently kind of a hack, and the perl code in objects.pl
3101 isn't very elegant, but it works as I intended. The simplest way
3102 to check that it worked correctly is to look in obj_dat.h and
3103 check the array nid_objs and make sure the objects haven't moved
3104 around (this is important!). Additions are OK, as well as
3105 consistent name changes.
3108 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3111 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3112 The given file adds to whatever has already been seeded into the
3113 random pool through the RANDFILE configuration file option or
3114 environment variable, or the default random state file.
3117 *) mkstack.pl now sorts each macro group into lexical order.
3118 Previously the output order depended on the order the files
3119 appeared in the directory, resulting in needless rewriting
3123 *) Patches to make OpenSSL compile under Win32 again. Mostly
3124 work arounds for the VC++ problem that it treats func() as
3125 func(void). Also stripped out the parts of mkdef.pl that
3126 added extra typesafe functions: these no longer exist.
3129 *) Reorganisation of the stack code. The macros are now all
3130 collected in safestack.h . Each macro is defined in terms of
3131 a "stack macro" of the form SKM_<name>(type, a, b). The
3132 DEBUG_SAFESTACK is now handled in terms of function casts,
3133 this has the advantage of retaining type safety without the
3134 use of additional functions. If DEBUG_SAFESTACK is not defined
3135 then the non typesafe macros are used instead. Also modified the
3136 mkstack.pl script to handle the new form. Needs testing to see
3137 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3138 the default if no major problems. Similar behaviour for ASN1_SET_OF
3139 and PKCS12_STACK_OF.
3142 *) When some versions of IIS use the 'NET' form of private key the
3143 key derivation algorithm is different. Normally MD5(password) is
3144 used as a 128 bit RC4 key. In the modified case
3145 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3146 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3147 as the old Netscape_RSA functions except they have an additional
3148 'sgckey' parameter which uses the modified algorithm. Also added
3149 an -sgckey command line option to the rsa utility. Thanks to
3150 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3151 algorithm to openssl-dev.
3154 *) The evp_local.h macros were using 'c.##kname' which resulted in
3155 invalid expansion on some systems (SCO 5.0.5 for example).
3156 Corrected to 'c.kname'.
3157 [Phillip Porch <root@theporch.com>]
3159 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3160 a STACK of email addresses from a certificate or request, these look
3161 in the subject name and the subject alternative name extensions and
3162 omit any duplicate addresses.
3165 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3166 This makes DSA verification about 2 % faster.
3169 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3170 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3171 plus overhead for 1024 bit moduli).
3172 This makes exponentiations about 0.5 % faster for 1024 bit
3173 exponents (as measured by "openssl speed rsa2048").
3176 *) Rename memory handling macros to avoid conflicts with other
3178 Malloc => OPENSSL_malloc
3179 Malloc_locked => OPENSSL_malloc_locked
3180 Realloc => OPENSSL_realloc
3181 Free => OPENSSL_free
3184 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3185 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3188 *) CygWin32 support.
3189 [John Jarvie <jjarvie@newsguy.com>]
3191 *) The type-safe stack code has been rejigged. It is now only compiled
3192 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3193 by default all type-specific stack functions are "#define"d back to
3194 standard stack functions. This results in more streamlined output
3195 but retains the type-safety checking possibilities of the original
3199 *) The STACK code has been cleaned up, and certain type declarations
3200 that didn't make a lot of sense have been brought in line. This has
3201 also involved a cleanup of sorts in safestack.h to more correctly
3202 map type-safe stack functions onto their plain stack counterparts.
3203 This work has also resulted in a variety of "const"ifications of
3204 lots of the code, especially "_cmp" operations which should normally
3205 be prototyped with "const" parameters anyway.
3208 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3209 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3210 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3211 where all of 'md' is used each time the PRNG is used, but 'state'
3212 is used only indexed by a cyclic counter. As entropy may not be
3213 well distributed from the beginning, 'md' is important as a
3214 chaining variable. However, the output function chains only half
3215 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3216 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3217 in all of 'state' being rewritten, with the new values depending
3218 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3221 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3222 the handshake is continued after ssl_verify_cert_chain();
3223 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3224 can lead to 'unexplainable' connection aborts later.
3225 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3227 *) Major EVP API cipher revision.
3228 Add hooks for extra EVP features. This allows various cipher
3229 parameters to be set in the EVP interface. Support added for variable
3230 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3231 setting of RC2 and RC5 parameters.
3233 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3236 Remove lots of duplicated code from the EVP library. For example *every*
3237 cipher init() function handles the 'iv' in the same way according to the
3238 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3239 for CFB and OFB modes they zero ctx->num.
3241 New functionality allows removal of S/MIME code RC2 hack.
3243 Most of the routines have the same form and so can be declared in terms
3246 By shifting this to the top level EVP_CipherInit() it can be removed from
3247 all individual ciphers. If the cipher wants to handle IVs or keys
3248 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3251 Change lots of functions like EVP_EncryptUpdate() to now return a
3252 value: although software versions of the algorithms cannot fail
3253 any installed hardware versions can.
3256 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3257 this option is set, tolerate broken clients that send the negotiated
3258 protocol version number instead of the requested protocol version
3262 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3263 i.e. non-zero for export ciphersuites, zero otherwise.
3264 Previous versions had this flag inverted, inconsistent with
3265 rsa_tmp_cb (..._TMP_RSA_CB).
3266 [Bodo Moeller; problem reported by Amit Chopra]
3268 *) Add missing DSA library text string. Work around for some IIS
3269 key files with invalid SEQUENCE encoding.
3272 *) Add a document (doc/standards.txt) that list all kinds of standards
3273 and so on that are implemented in OpenSSL.
3276 *) Enhance c_rehash script. Old version would mishandle certificates
3277 with the same subject name hash and wouldn't handle CRLs at all.
3278 Added -fingerprint option to crl utility, to support new c_rehash
3282 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3285 *) Fix for SSL server purpose checking. Server checking was
3286 rejecting certificates which had extended key usage present
3287 but no ssl client purpose.
3288 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3290 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3291 is a little unclear about how a blank password is handled.
3292 Since the password in encoded as a BMPString with terminating
3293 double NULL a zero length password would end up as just the
3294 double NULL. However no password at all is different and is
3295 handled differently in the PKCS#12 key generation code. NS
3296 treats a blank password as zero length. MSIE treats it as no
3297 password on export: but it will try both on import. We now do
3298 the same: PKCS12_parse() tries zero length and no password if
3299 the password is set to "" or NULL (NULL is now a valid password:
3300 it wasn't before) as does the pkcs12 application.
3303 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3304 perror when PEM_read_bio_X509_REQ fails, the error message must
3305 be obtained from the error queue.
3308 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3309 it in ERR_remove_state if appropriate, and change ERR_get_state
3310 accordingly to avoid race conditions (this is necessary because
3311 thread_hash is no longer constant once set).
3314 *) Bugfix for linux-elf makefile.one.
3317 *) RSA_get_default_method() will now cause a default
3318 RSA_METHOD to be chosen if one doesn't exist already.
3319 Previously this was only set during a call to RSA_new()
3320 or RSA_new_method(NULL) meaning it was possible for
3321 RSA_get_default_method() to return NULL.
3324 *) Added native name translation to the existing DSO code
3325 that will convert (if the flag to do so is set) filenames
3326 that are sufficiently small and have no path information
3327 into a canonical native form. Eg. "blah" converted to
3328 "libblah.so" or "blah.dll" etc.
3331 *) New function ERR_error_string_n(e, buf, len) which is like
3332 ERR_error_string(e, buf), but writes at most 'len' bytes
3333 including the 0 terminator. For ERR_error_string_n, 'buf'
3335 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3337 *) CONF library reworked to become more general. A new CONF
3338 configuration file reader "class" is implemented as well as a
3339 new functions (NCONF_*, for "New CONF") to handle it. The now
3340 old CONF_* functions are still there, but are reimplemented to
3341 work in terms of the new functions. Also, a set of functions
3342 to handle the internal storage of the configuration data is
3343 provided to make it easier to write new configuration file
3344 reader "classes" (I can definitely see something reading a
3345 configuration file in XML format, for example), called _CONF_*,
3346 or "the configuration storage API"...
3348 The new configuration file reading functions are:
3350 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3351 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3353 NCONF_default, NCONF_WIN32
3355 NCONF_dump_fp, NCONF_dump_bio
3357 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3358 NCONF_new creates a new CONF object. This works in the same way
3359 as other interfaces in OpenSSL, like the BIO interface.
3360 NCONF_dump_* dump the internal storage of the configuration file,
3361 which is useful for debugging. All other functions take the same
3362 arguments as the old CONF_* functions wth the exception of the
3363 first that must be a `CONF *' instead of a `LHASH *'.
3365 To make it easer to use the new classes with the old CONF_* functions,
3366 the function CONF_set_default_method is provided.
3369 *) Add '-tls1' option to 'openssl ciphers', which was already
3370 mentioned in the documentation but had not been implemented.
3371 (This option is not yet really useful because even the additional
3372 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3375 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3376 OpenSSL-based applications) load shared libraries and bind to
3377 them in a portable way.
3378 [Geoff Thorpe, with contributions from Richard Levitte]
3380 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3382 *) Make sure _lrotl and _lrotr are only used with MSVC.
3384 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3385 (the default implementation of RAND_status).
3387 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3388 to '-clrext' (= clear extensions), as intended and documented.
3389 [Bodo Moeller; inconsistency pointed out by Michael Attili
3392 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3393 was larger than the MD block size.
3394 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3396 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3397 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3398 using the passed key: if the passed key was a private key the result
3399 of X509_print(), for example, would be to print out all the private key
3403 *) des_quad_cksum() byte order bug fix.
3404 [Ulf Möller, using the problem description in krb4-0.9.7, where
3405 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3407 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3409 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3411 *) For easily testing in shell scripts whether some command
3412 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3413 returns with exit code 0 iff no command of the given name is available.
3414 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3415 the output goes to stdout and nothing is printed to stderr.
3416 Additional arguments are always ignored.
3418 Since for each cipher there is a command of the same name,
3419 the 'no-cipher' compilation switches can be tested this way.
3421 ('openssl no-XXX' is not able to detect pseudo-commands such
3422 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3425 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3428 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3429 is set; it will be thrown away anyway because each handshake creates
3431 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3432 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3433 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3434 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3437 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3438 'Q' and 'R' lose their special meanings (quit/renegotiate).
3439 This is part of what -quiet does; unlike -quiet, -ign_eof
3440 does not suppress any output.
3443 *) Add compatibility options to the purpose and trust code. The
3444 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3445 accepts a certificate or CA, this was the previous behaviour,
3446 with all the associated security issues.
3448 X509_TRUST_COMPAT is the old trust behaviour: only and
3449 automatically trust self signed roots in certificate store. A
3450 new trust setting X509_TRUST_DEFAULT is used to specify that
3451 a purpose has no associated trust setting and it should instead
3452 use the value in the default purpose.
3455 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3456 and fix a memory leak.
3459 *) In util/mkerr.pl (which implements 'make errors'), preserve
3460 reason strings from the previous version of the .c file, as
3461 the default to have only downcase letters (and digits) in
3462 automatically generated reasons codes is not always appropriate.
3465 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3466 using strerror. Previously, ERR_reason_error_string() returned
3467 library names as reason strings for SYSerr; but SYSerr is a special
3468 case where small numbers are errno values, not library numbers.
3471 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3472 converts DSA parameters into DH parameters. (When creating parameters,
3473 DSA_generate_parameters is used.)
3476 *) Include 'length' (recommended exponent length) in C code generated
3477 by 'openssl dhparam -C'.
3480 *) The second argument to set_label in perlasm was already being used
3481 so couldn't be used as a "file scope" flag. Moved to third argument
3485 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3486 instead of RAND_bytes for encryption IVs and salts.
3489 *) Include RAND_status() into RAND_METHOD instead of implementing
3490 it only for md_rand.c Otherwise replacing the PRNG by calling
3491 RAND_set_rand_method would be impossible.
3494 *) Don't let DSA_generate_key() enter an infinite loop if the random
3495 number generation fails.
3498 *) New 'rand' application for creating pseudo-random output.
3501 *) Added configuration support for Linux/IA64
3502 [Rolf Haberrecker <rolf@suse.de>]
3504 *) Assembler module support for Mingw32.
3507 *) Shared library support for HPUX (in shlib/).
3508 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3510 *) Shared library support for Solaris gcc.
3511 [Lutz Behnke <behnke@trustcenter.de>]
3513 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3515 *) PKCS7_encrypt() was adding text MIME headers twice because they
3516 were added manually and by SMIME_crlf_copy().
3519 *) In bntest.c don't call BN_rand with zero bits argument.
3520 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3522 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3523 case was implemented. This caused BN_div_recp() to fail occasionally.
3526 *) Add an optional second argument to the set_label() in the perl
3527 assembly language builder. If this argument exists and is set
3528 to 1 it signals that the assembler should use a symbol whose
3529 scope is the entire file, not just the current function. This
3530 is needed with MASM which uses the format label:: for this scope.
3531 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3533 *) Change the ASN1 types so they are typedefs by default. Before
3534 almost all types were #define'd to ASN1_STRING which was causing
3535 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3539 *) Change names of new functions to the new get1/get0 naming
3540 convention: After 'get1', the caller owns a reference count
3541 and has to call ..._free; 'get0' returns a pointer to some
3542 data structure without incrementing reference counters.
3543 (Some of the existing 'get' functions increment a reference
3544 counter, some don't.)
3545 Similarly, 'set1' and 'add1' functions increase reference
3546 counters or duplicate objects.
3549 *) Allow for the possibility of temp RSA key generation failure:
3550 the code used to assume it always worked and crashed on failure.
3553 *) Fix potential buffer overrun problem in BIO_printf().
3554 [Ulf Möller, using public domain code by Patrick Powell; problem
3555 pointed out by David Sacerdote <das33@cornell.edu>]
3557 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3558 RAND_egd() and RAND_status(). In the command line application,
3559 the EGD socket can be specified like a seed file using RANDFILE
3563 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3564 Some CAs (e.g. Verisign) distribute certificates in this form.
3567 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3568 list to exclude them. This means that no special compilation option
3569 is needed to use anonymous DH: it just needs to be included in the
3573 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3574 EVP_MD_type. The old functionality is available in a new macro called
3575 EVP_MD_md(). Change code that uses it and update docs.
3578 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3579 where the 'void *' argument is replaced by a function pointer argument.
3580 Previously 'void *' was abused to point to functions, which works on
3581 many platforms, but is not correct. As these functions are usually
3582 called by macros defined in OpenSSL header files, most source code
3583 should work without changes.
3586 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3587 sections with information on -D... compiler switches used for
3588 compiling the library so that applications can see them. To enable
3589 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3590 must be defined. E.g.,
3591 #define OPENSSL_ALGORITHM_DEFINES
3592 #include <openssl/opensslconf.h>
3593 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3594 [Richard Levitte, Ulf and Bodo Möller]
3596 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3600 *) Change the 'other' type in certificate aux info to a STACK_OF
3601 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3602 the required ASN1 format: arbitrary types determined by an OID.
3605 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3606 argument to 'req'. This is not because the function is newer or
3607 better than others it just uses the work 'NEW' in the certificate
3608 request header lines. Some software needs this.
3611 *) Reorganise password command line arguments: now passwords can be
3612 obtained from various sources. Delete the PEM_cb function and make
3613 it the default behaviour: i.e. if the callback is NULL and the
3614 usrdata argument is not NULL interpret it as a null terminated pass
3615 phrase. If usrdata and the callback are NULL then the pass phrase
3616 is prompted for as usual.
3619 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3620 the support is automatically enabled. The resulting binaries will
3621 autodetect the card and use it if present.
3622 [Ben Laurie and Compaq Inc.]
3624 *) Work around for Netscape hang bug. This sends certificate request
3625 and server done in one record. Since this is perfectly legal in the
3626 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3627 the bugs/SSLv3 entry for more info.
3630 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3633 *) Add -rand argument to smime and pkcs12 applications and read/write
3637 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3640 *) Add command line password options to the remaining applications.
3643 *) Bug fix for BN_div_recp() for numerators with an even number of
3647 *) More tests in bntest.c, and changed test_bn output.
3650 *) ./config recognizes MacOS X now.
3653 *) Bug fix for BN_div() when the first words of num and divsor are
3654 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3657 *) Add support for various broken PKCS#8 formats, and command line
3658 options to produce them.
3661 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3662 get temporary BIGNUMs from a BN_CTX.
3665 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3669 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3670 include a #define from the old name to the new. The original intent
3671 was that statically linked binaries could for example just call
3672 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3673 link with digests. This never worked becayse SSLeay_add_all_digests()
3674 and SSLeay_add_all_ciphers() were in the same source file so calling
3675 one would link with the other. They are now in separate source files.
3678 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3681 *) Use a less unusual form of the Miller-Rabin primality test (it used
3682 a binary algorithm for exponentiation integrated into the Miller-Rabin
3683 loop, our standard modexp algorithms are faster).
3686 *) Support for the EBCDIC character set completed.
3687 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3689 *) Source code cleanups: use const where appropriate, eliminate casts,
3690 use void * instead of char * in lhash.
3693 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3694 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3695 this the server could overwrite ephemeral keys that the client
3699 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3700 using 50 iterations of the Rabin-Miller test.
3702 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3703 iterations of the Rabin-Miller test as required by the appendix
3704 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3705 As BN_is_prime_fasttest includes trial division, DSA parameter
3706 generation becomes much faster.
3708 This implies a change for the callback functions in DSA_is_prime
3709 and DSA_generate_parameters: The callback function is called once
3710 for each positive witness in the Rabin-Miller test, not just
3711 occasionally in the inner loop; and the parameters to the
3712 callback function now provide an iteration count for the outer
3713 loop rather than for the current invocation of the inner loop.
3714 DSA_generate_parameters additionally can call the callback
3715 function with an 'iteration count' of -1, meaning that a
3716 candidate has passed the trial division test (when q is generated
3717 from an application-provided seed, trial division is skipped).
3720 *) New function BN_is_prime_fasttest that optionally does trial
3721 division before starting the Rabin-Miller test and has
3722 an additional BN_CTX * argument (whereas BN_is_prime always
3723 has to allocate at least one BN_CTX).
3724 'callback(1, -1, cb_arg)' is called when a number has passed the
3725 trial division stage.
3728 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3732 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3735 *) New function BN_pseudo_rand().
3738 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3739 bignum version of BN_from_montgomery() with the working code from
3740 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3744 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3745 made it impossible to use the same SSL_SESSION data structure in
3746 SSL2 clients in multiple threads.
3749 *) The return value of RAND_load_file() no longer counts bytes obtained
3750 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3751 to seed the PRNG (previously an explicit byte count was required).
3752 [Ulf Möller, Bodo Möller]
3754 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3755 used (char *) instead of (void *) and had casts all over the place.
3758 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3761 *) Retain source code compatibility for BN_prime_checks macro:
3762 BN_is_prime(..., BN_prime_checks, ...) now uses
3763 BN_prime_checks_for_size to determine the appropriate number of
3764 Rabin-Miller iterations.
3767 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3768 DH_CHECK_P_NOT_SAFE_PRIME.
3769 (Check if this is true? OpenPGP calls them "strong".)
3772 *) Merge the functionality of "dh" and "gendh" programs into a new program
3773 "dhparam". The old programs are retained for now but will handle DH keys
3774 (instead of parameters) in future.
3777 *) Make the ciphers, s_server and s_client programs check the return values
3778 when a new cipher list is set.
3781 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3782 ciphers. Before when the 56bit ciphers were enabled the sorting was
3785 The syntax for the cipher sorting has been extended to support sorting by
3786 cipher-strength (using the strength_bits hard coded in the tables).
3787 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3789 Fix a bug in the cipher-command parser: when supplying a cipher command
3790 string with an "undefined" symbol (neither command nor alphanumeric
3791 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3792 an error is flagged.
3794 Due to the strength-sorting extension, the code of the
3795 ssl_create_cipher_list() function was completely rearranged. I hope that
3796 the readability was also increased :-)
3797 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3799 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3800 for the first serial number and places 2 in the serial number file. This
3801 avoids problems when the root CA is created with serial number zero and
3802 the first user certificate has the same issuer name and serial number
3806 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3807 the new code. Add documentation for this stuff.
3810 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3811 X509_*() to X509at_*() on the grounds that they don't handle X509
3812 structures and behave in an analagous way to the X509v3 functions:
3813 they shouldn't be called directly but wrapper functions should be used
3816 So we also now have some wrapper functions that call the X509at functions
3817 when passed certificate requests. (TO DO: similar things can be done with
3818 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3819 things. Some of these need some d2i or i2d and print functionality
3820 because they handle more complex structures.)
3823 *) Add missing #ifndefs that caused missing symbols when building libssl
3824 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3825 NO_RSA in ssl/s2*.c.
3826 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3828 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3829 has a return value which indicates the quality of the random data
3830 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3831 error queue. New function RAND_pseudo_bytes() generates output that is
3832 guaranteed to be unique but not unpredictable. RAND_add is like
3833 RAND_seed, but takes an extra argument for an entropy estimate
3834 (RAND_seed always assumes full entropy).
3837 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3838 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3839 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3840 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3841 false-positive rate of at most 2^-80 for random input.
3844 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3847 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3848 in the 0.9.5 release), this returns the chain
3849 from an X509_CTX structure with a dup of the stack and all
3850 the X509 reference counts upped: so the stack will exist
3851 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3854 Also make SSL_SESSION_print() print out the verify return
3858 *) Add manpage for the pkcs12 command. Also change the default
3859 behaviour so MAC iteration counts are used unless the new
3860 -nomaciter option is used. This improves file security and
3861 only older versions of MSIE (4.0 for example) need it.
3864 *) Honor the no-xxx Configure options when creating .DEF files.
3867 *) Add PKCS#10 attributes to field table: challengePassword,
3868 unstructuredName and unstructuredAddress. These are taken from
3869 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3870 international characters are used.
3872 More changes to X509_ATTRIBUTE code: allow the setting of types
3873 based on strings. Remove the 'loc' parameter when adding
3874 attributes because these will be a SET OF encoding which is sorted
3878 *) Initial changes to the 'req' utility to allow request generation
3879 automation. This will allow an application to just generate a template
3880 file containing all the field values and have req construct the
3883 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3884 used all over the place including certificate requests and PKCS#7
3885 structures. They are currently handled manually where necessary with
3886 some primitive wrappers for PKCS#7. The new functions behave in a
3887 manner analogous to the X509 extension functions: they allow
3888 attributes to be looked up by NID and added.
3890 Later something similar to the X509V3 code would be desirable to
3891 automatically handle the encoding, decoding and printing of the
3892 more complex types. The string types like challengePassword can
3893 be handled by the string table functions.
3895 Also modified the multi byte string table handling. Now there is
3896 a 'global mask' which masks out certain types. The table itself
3897 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3898 is useful when for example there is only one permissible type
3899 (as in countryName) and using the mask might result in no valid
3903 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3904 SSL_get_peer_finished to allow applications to obtain the latest
3905 Finished messages sent to the peer or expected from the peer,
3906 respectively. (SSL_get_peer_finished is usually the Finished message
3907 actually received from the peer, otherwise the protocol will be aborted.)
3909 As the Finished message are message digests of the complete handshake
3910 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3911 be used for external authentication procedures when the authentication
3912 provided by SSL/TLS is not desired or is not enough.
3915 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3916 the host supports BWX extension and if Compaq C is present on the
3917 $PATH. Just exploiting of the BWX extension results in 20-30%
3918 performance kick for some algorithms, e.g. DES and RC4 to mention
3919 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3923 *) Add support for MS "fast SGC". This is arguably a violation of the
3924 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3925 weak crypto and after checking the certificate is SGC a second one
3926 with strong crypto. MS SGC stops the first handshake after receiving
3927 the server certificate message and sends a second client hello. Since
3928 a server will typically do all the time consuming operations before
3929 expecting any further messages from the client (server key exchange
3930 is the most expensive) there is little difference between the two.
3932 To get OpenSSL to support MS SGC we have to permit a second client
3933 hello message after we have sent server done. In addition we have to
3934 reset the MAC if we do get this second client hello.
3937 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3938 if a DER encoded private key is RSA or DSA traditional format. Changed
3939 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3940 format DER encoded private key. Newer code should use PKCS#8 format which
3941 has the key type encoded in the ASN1 structure. Added DER private key
3942 support to pkcs8 application.
3945 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3946 ciphersuites has been selected (as required by the SSL 3/TLS 1
3947 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3948 is set, we interpret this as a request to violate the specification
3949 (the worst that can happen is a handshake failure, and 'correct'
3950 behaviour would result in a handshake failure anyway).
3953 *) In SSL_CTX_add_session, take into account that there might be multiple
3954 SSL_SESSION structures with the same session ID (e.g. when two threads
3955 concurrently obtain them from an external cache).
3956 The internal cache can handle only one SSL_SESSION with a given ID,
3957 so if there's a conflict, we now throw out the old one to achieve
3961 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3962 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3963 some routines that use cipher OIDs: some ciphers do not have OIDs
3964 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3968 *) Simplify the trust setting structure and code. Now we just have
3969 two sequences of OIDs for trusted and rejected settings. These will
3970 typically have values the same as the extended key usage extension
3971 and any application specific purposes.
3973 The trust checking code now has a default behaviour: it will just
3974 check for an object with the same NID as the passed id. Functions can
3975 be provided to override either the default behaviour or the behaviour
3976 for a given id. SSL client, server and email already have functions
3977 in place for compatibility: they check the NID and also return "trusted"
3978 if the certificate is self signed.
3981 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3982 traditional format into an EVP_PKEY structure.
3985 *) Add a password callback function PEM_cb() which either prompts for
3986 a password if usr_data is NULL or otherwise assumes it is a null
3987 terminated password. Allow passwords to be passed on command line
3988 environment or config files in a few more utilities.
3991 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3992 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3993 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3994 Update documentation.
3997 *) Support for ASN1 "NULL" type. This could be handled before by using
3998 ASN1_TYPE but there wasn't any function that would try to read a NULL
3999 and produce an error if it couldn't. For compatibility we also have
4000 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4001 don't allocate anything because they don't need to.
4004 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4006 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4008 *) Rebuild of the memory allocation routines used by OpenSSL code and
4009 possibly others as well. The purpose is to make an interface that
4010 provide hooks so anyone can build a separate set of allocation and
4011 deallocation routines to be used by OpenSSL, for example memory
4012 pool implementations, or something else, which was previously hard
4013 since Malloc(), Realloc() and Free() were defined as macros having
4014 the values malloc, realloc and free, respectively (except for Win32
4015 compilations). The same is provided for memory debugging code.
4016 OpenSSL already comes with functionality to find memory leaks, but
4017 this gives people a chance to debug other memory problems.
4019 With these changes, a new set of functions and macros have appeared:
4021 CRYPTO_set_mem_debug_functions() [F]
4022 CRYPTO_get_mem_debug_functions() [F]
4023 CRYPTO_dbg_set_options() [F]
4024 CRYPTO_dbg_get_options() [F]
4025 CRYPTO_malloc_debug_init() [M]
4027 The memory debug functions are NULL by default, unless the library
4028 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4029 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4030 gives the standard debugging functions that come with OpenSSL) or
4031 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4032 provided by the library user) must be used. When the standard
4033 debugging functions are used, CRYPTO_dbg_set_options can be used to
4034 request additional information:
4035 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4036 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4038 Also, things like CRYPTO_set_mem_functions will always give the
4039 expected result (the new set of functions is used for allocation
4040 and deallocation) at all times, regardless of platform and compiler
4043 To finish it up, some functions that were never use in any other
4044 way than through macros have a new API and new semantic:
4047 CRYPTO_dbg_realloc()
4050 All macros of value have retained their old syntax.
4051 [Richard Levitte and Bodo Moeller]
4053 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4054 ordering of SMIMECapabilities wasn't in "strength order" and there
4055 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4059 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4060 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4061 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4063 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4064 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4065 functionality to handle multipart/signed properly) and a utility
4066 called 'smime' to call all this stuff. This is based on code I
4067 originally wrote for Celo who have kindly allowed it to be
4068 included in OpenSSL.
4071 *) Add variants des_set_key_checked and des_set_key_unchecked of
4072 des_set_key (aka des_key_sched). Global variable des_check_key
4073 decides which of these is called by des_set_key; this way
4074 des_check_key behaves as it always did, but applications and
4075 the library itself, which was buggy for des_check_key == 1,
4076 have a cleaner way to pick the version they need.
4079 *) New function PKCS12_newpass() which changes the password of a
4083 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4084 dynamic mix. In both cases the ids can be used as an index into the
4085 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4086 functions so they accept a list of the field values and the
4087 application doesn't need to directly manipulate the X509_TRUST
4091 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4095 *) Modify the way the V3 extension code looks up extensions. This now
4096 works in a similar way to the object code: we have some "standard"
4097 extensions in a static table which is searched with OBJ_bsearch()
4098 and the application can add dynamic ones if needed. The file
4099 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4100 updated whenever a new extension is added to the core code and kept
4101 in ext_nid order. There is a simple program 'tabtest.c' which checks
4102 this. New extensions are not added too often so this file can readily
4103 be maintained manually.
4105 There are two big advantages in doing things this way. The extensions
4106 can be looked up immediately and no longer need to be "added" using
4107 X509V3_add_standard_extensions(): this function now does nothing.
4108 [Side note: I get *lots* of email saying the extension code doesn't
4109 work because people forget to call this function]
4110 Also no dynamic allocation is done unless new extensions are added:
4111 so if we don't add custom extensions there is no need to call
4112 X509V3_EXT_cleanup().
4115 *) Modify enc utility's salting as follows: make salting the default. Add a
4116 magic header, so unsalted files fail gracefully instead of just decrypting
4117 to garbage. This is because not salting is a big security hole, so people
4118 should be discouraged from doing it.
4121 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4122 digest to be passed on the command line but it only used this
4123 parameter when signing a certificate. Modified so all relevant
4124 operations are affected by the digest parameter including the
4125 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4126 DSA key was used because it didn't fix the digest.
4129 *) Initial certificate chain verify code. Currently tests the untrusted
4130 certificates for consistency with the verify purpose (which is set
4131 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4133 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4134 this is because it will reject chains with invalid extensions whereas
4135 every previous version of OpenSSL and SSLeay made no checks at all.
4137 Trust code: checks the root CA for the relevant trust settings. Trust
4138 settings have an initial value consistent with the verify purpose: e.g.
4139 if the verify purpose is for SSL client use it expects the CA to be
4140 trusted for SSL client use. However the default value can be changed to
4141 permit custom trust settings: one example of this would be to only trust
4142 certificates from a specific "secure" set of CAs.
4144 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4145 which should be used for version portability: especially since the
4146 verify structure is likely to change more often now.
4148 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4149 to set them. If not set then assume SSL clients will verify SSL servers
4152 Two new options to the verify program: -untrusted allows a set of
4153 untrusted certificates to be passed in and -purpose which sets the
4154 intended purpose of the certificate. If a purpose is set then the
4155 new chain verify code is used to check extension consistency.
4158 *) Support for the authority information access extension.
4161 *) Modify RSA and DSA PEM read routines to transparently handle
4162 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4163 public keys in a format compatible with certificate
4164 SubjectPublicKeyInfo structures. Unfortunately there were already
4165 functions called *_PublicKey_* which used various odd formats so
4166 these are retained for compatibility: however the DSA variants were
4167 never in a public release so they have been deleted. Changed dsa/rsa
4168 utilities to handle the new format: note no releases ever handled public
4169 keys so we should be OK.
4171 The primary motivation for this change is to avoid the same fiasco
4172 that dogs private keys: there are several incompatible private key
4173 formats some of which are standard and some OpenSSL specific and
4174 require various evil hacks to allow partial transparent handling and
4175 even then it doesn't work with DER formats. Given the option anything
4176 other than PKCS#8 should be dumped: but the other formats have to
4177 stay in the name of compatibility.
4179 With public keys and the benefit of hindsight one standard format
4180 is used which works with EVP_PKEY, RSA or DSA structures: though
4181 it clearly returns an error if you try to read the wrong kind of key.
4183 Added a -pubkey option to the 'x509' utility to output the public key.
4184 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4185 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4186 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4187 that do the same as the EVP_PKEY_assign_*() except they up the
4188 reference count of the added key (they don't "swallow" the
4192 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4193 CRLs would fail if the file contained no certificates or no CRLs:
4194 added a new function to read in both types and return the number
4195 read: this means that if none are read it will be an error. The
4196 DER versions of the certificate and CRL reader would always fail
4197 because it isn't possible to mix certificates and CRLs in DER format
4198 without choking one or the other routine. Changed this to just read
4199 a certificate: this is the best we can do. Also modified the code
4200 in apps/verify.c to take notice of return codes: it was previously
4201 attempting to read in certificates from NULL pointers and ignoring
4202 any errors: this is one reason why the cert and CRL reader seemed
4203 to work. It doesn't check return codes from the default certificate
4204 routines: these may well fail if the certificates aren't installed.
4207 *) Code to support otherName option in GeneralName.
4210 *) First update to verify code. Change the verify utility
4211 so it warns if it is passed a self signed certificate:
4212 for consistency with the normal behaviour. X509_verify
4213 has been modified to it will now verify a self signed
4214 certificate if *exactly* the same certificate appears
4215 in the store: it was previously impossible to trust a
4216 single self signed certificate. This means that:
4217 openssl verify ss.pem
4218 now gives a warning about a self signed certificate but
4219 openssl verify -CAfile ss.pem ss.pem
4223 *) For servers, store verify_result in SSL_SESSION data structure
4224 (and add it to external session representation).
4225 This is needed when client certificate verifications fails,
4226 but an application-provided verification callback (set by
4227 SSL_CTX_set_cert_verify_callback) allows accepting the session
4228 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4229 but returns 1): When the session is reused, we have to set
4230 ssl->verify_result to the appropriate error code to avoid
4232 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4234 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4235 case in PKCS7_dataInit() where the signed PKCS7 structure
4236 didn't contain any existing data because it was being created.
4237 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4239 *) Add a salt to the key derivation routines in enc.c. This
4240 forms the first 8 bytes of the encrypted file. Also add a
4241 -S option to allow a salt to be input on the command line.
4244 *) New function X509_cmp(). Oddly enough there wasn't a function
4245 to compare two certificates. We do this by working out the SHA1
4246 hash and comparing that. X509_cmp() will be needed by the trust
4250 *) SSL_get1_session() is like SSL_get_session(), but increments
4251 the reference count in the SSL_SESSION returned.
4252 [Geoff Thorpe <geoff@eu.c2.net>]
4254 *) Fix for 'req': it was adding a null to request attributes.
4255 Also change the X509_LOOKUP and X509_INFO code to handle
4256 certificate auxiliary information.
4259 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4263 *) Add the possibility to add extra information to the memory leak
4264 detecting output, to form tracebacks, showing from where each
4265 allocation was originated: CRYPTO_push_info("constant string") adds
4266 the string plus current file name and line number to a per-thread
4267 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4268 is like calling CYRPTO_pop_info() until the stack is empty.
4269 Also updated memory leak detection code to be multi-thread-safe.
4272 *) Add options -text and -noout to pkcs7 utility and delete the
4273 encryption options which never did anything. Update docs.
4276 *) Add options to some of the utilities to allow the pass phrase
4277 to be included on either the command line (not recommended on
4278 OSes like Unix) or read from the environment. Update the
4279 manpages and fix a few bugs.
4282 *) Add a few manpages for some of the openssl commands.
4285 *) Fix the -revoke option in ca. It was freeing up memory twice,
4286 leaking and not finding already revoked certificates.
4289 *) Extensive changes to support certificate auxiliary information.
4290 This involves the use of X509_CERT_AUX structure and X509_AUX
4291 functions. An X509_AUX function such as PEM_read_X509_AUX()
4292 can still read in a certificate file in the usual way but it
4293 will also read in any additional "auxiliary information". By
4294 doing things this way a fair degree of compatibility can be
4295 retained: existing certificates can have this information added
4296 using the new 'x509' options.
4298 Current auxiliary information includes an "alias" and some trust
4299 settings. The trust settings will ultimately be used in enhanced
4300 certificate chain verification routines: currently a certificate
4301 can only be trusted if it is self signed and then it is trusted
4305 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4306 The problem was that one of the replacement routines had not been working
4307 since SSLeay releases. For now the offending routine has been replaced
4308 with non-optimised assembler. Even so, this now gives around 95%
4309 performance improvement for 1024 bit RSA signs.
4312 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4313 handling. Most clients have the effective key size in bits equal to
4314 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4315 A few however don't do this and instead use the size of the decrypted key
4316 to determine the RC2 key length and the AlgorithmIdentifier to determine
4317 the effective key length. In this case the effective key length can still
4318 be 40 bits but the key length can be 168 bits for example. This is fixed
4319 by manually forcing an RC2 key into the EVP_PKEY structure because the
4320 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4321 the key length and effective key length are equal.
4324 *) Add a bunch of functions that should simplify the creation of
4325 X509_NAME structures. Now you should be able to do:
4326 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4327 and have it automatically work out the correct field type and fill in
4328 the structures. The more adventurous can try:
4329 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4330 and it will (hopefully) work out the correct multibyte encoding.
4333 *) Change the 'req' utility to use the new field handling and multibyte
4334 copy routines. Before the DN field creation was handled in an ad hoc
4335 way in req, ca, and x509 which was rather broken and didn't support
4336 BMPStrings or UTF8Strings. Since some software doesn't implement
4337 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4338 using the dirstring_type option. See the new comment in the default
4339 openssl.cnf for more info.
4342 *) Make crypto/rand/md_rand.c more robust:
4343 - Assure unique random numbers after fork().
4344 - Make sure that concurrent threads access the global counter and
4345 md serializably so that we never lose entropy in them
4346 or use exactly the same state in multiple threads.
4347 Access to the large state is not always serializable because
4348 the additional locking could be a performance killer, and
4349 md should be large enough anyway.
4352 *) New file apps/app_rand.c with commonly needed functionality
4353 for handling the random seed file.
4355 Use the random seed file in some applications that previously did not:
4357 dsaparam -genkey (which also ignored its '-rand' option),
4360 x509 (when signing).
4361 Except on systems with /dev/urandom, it is crucial to have a random
4362 seed file at least for key creation, DSA signing, and for DH exchanges;
4363 for RSA signatures we could do without one.
4365 gendh and gendsa (unlike genrsa) used to read only the first byte
4366 of each file listed in the '-rand' option. The function as previously
4367 found in genrsa is now in app_rand.c and is used by all programs
4368 that support '-rand'.
4371 *) In RAND_write_file, use mode 0600 for creating files;
4372 don't just chmod when it may be too late.
4375 *) Report an error from X509_STORE_load_locations
4376 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4379 *) New function ASN1_mbstring_copy() this copies a string in either
4380 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4381 into an ASN1_STRING type. A mask of permissible types is passed
4382 and it chooses the "minimal" type to use or an error if not type
4386 *) Add function equivalents to the various macros in asn1.h. The old
4387 macros are retained with an M_ prefix. Code inside the library can
4388 use the M_ macros. External code (including the openssl utility)
4389 should *NOT* in order to be "shared library friendly".
4392 *) Add various functions that can check a certificate's extensions
4393 to see if it usable for various purposes such as SSL client,
4394 server or S/MIME and CAs of these types. This is currently
4395 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4396 verification. Also added a -purpose flag to x509 utility to
4397 print out all the purposes.
4400 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4404 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4405 for, obtain and decode and extension and obtain its critical flag.
4406 This allows all the necessary extension code to be handled in a
4407 single function call.
4410 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4411 platforms. See crypto/rc4/rc4_enc.c for further details.
4414 *) New -noout option to asn1parse. This causes no output to be produced
4415 its main use is when combined with -strparse and -out to extract data
4416 from a file (which may not be in ASN.1 format).
4419 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4420 when producing the local key id.
4421 [Richard Levitte <levitte@stacken.kth.se>]
4423 *) New option -dhparam in s_server. This allows a DH parameter file to be
4424 stated explicitly. If it is not stated then it tries the first server
4425 certificate file. The previous behaviour hard coded the filename
4429 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4430 a public key to be input or output. For example:
4431 openssl rsa -in key.pem -pubout -out pubkey.pem
4432 Also added necessary DSA public key functions to handle this.
4435 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4436 in the message. This was handled by allowing
4437 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4438 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4440 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4441 to the end of the strings whereas this didn't. This would cause problems
4442 if strings read with d2i_ASN1_bytes() were later modified.
4443 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4445 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4446 data and it contains EOF it will end up returning an error. This is
4447 caused by input 46 bytes long. The cause is due to the way base64
4448 BIOs find the start of base64 encoded data. They do this by trying a
4449 trial decode on each line until they find one that works. When they
4450 do a flag is set and it starts again knowing it can pass all the
4451 data directly through the decoder. Unfortunately it doesn't reset
4452 the context it uses. This means that if EOF is reached an attempt
4453 is made to pass two EOFs through the context and this causes the
4454 resulting error. This can also cause other problems as well. As is
4455 usual with these problems it takes *ages* to find and the fix is
4456 trivial: move one line.
4457 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4459 *) Ugly workaround to get s_client and s_server working under Windows. The
4460 old code wouldn't work because it needed to select() on sockets and the
4461 tty (for keypresses and to see if data could be written). Win32 only
4462 supports select() on sockets so we select() with a 1s timeout on the
4463 sockets and then see if any characters are waiting to be read, if none
4464 are present then we retry, we also assume we can always write data to
4465 the tty. This isn't nice because the code then blocks until we've
4466 received a complete line of data and it is effectively polling the
4467 keyboard at 1s intervals: however it's quite a bit better than not
4468 working at all :-) A dedicated Windows application might handle this
4469 with an event loop for example.
4472 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4473 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4474 will be called when RSA_sign() and RSA_verify() are used. This is useful
4475 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4476 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4477 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4478 This necessitated the support of an extra signature type NID_md5_sha1
4479 for SSL signatures and modifications to the SSL library to use it instead
4480 of calling RSA_public_decrypt() and RSA_private_encrypt().
4483 *) Add new -verify -CAfile and -CApath options to the crl program, these
4484 will lookup a CRL issuers certificate and verify the signature in a
4485 similar way to the verify program. Tidy up the crl program so it
4486 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4487 less strict. It will now permit CRL extensions even if it is not
4488 a V2 CRL: this will allow it to tolerate some broken CRLs.
4491 *) Initialize all non-automatic variables each time one of the openssl
4492 sub-programs is started (this is necessary as they may be started
4493 multiple times from the "OpenSSL>" prompt).
4494 [Lennart Bang, Bodo Moeller]
4496 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4497 removing all other RSA functionality (this is what NO_RSA does). This
4498 is so (for example) those in the US can disable those operations covered
4499 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4503 *) Non-copying interface to BIO pairs.
4504 (still largely untested)
4507 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4508 ASCII string. This was handled independently in various places before.
4511 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4512 UTF8 strings a character at a time.
4515 *) Use client_version from client hello to select the protocol
4516 (s23_srvr.c) and for RSA client key exchange verification
4517 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4520 *) Add various utility functions to handle SPKACs, these were previously
4521 handled by poking round in the structure internals. Added new function
4522 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4523 print, verify and generate SPKACs. Based on an original idea from
4524 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4527 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4530 *) Allow the config file extension section to be overwritten on the
4531 command line. Based on an original idea from Massimiliano Pala
4532 <madwolf@comune.modena.it>. The new option is called -extensions
4533 and can be applied to ca, req and x509. Also -reqexts to override
4534 the request extensions in req and -crlexts to override the crl extensions
4538 *) Add new feature to the SPKAC handling in ca. Now you can include
4539 the same field multiple times by preceding it by "XXXX." for example:
4542 this is the same syntax as used in the req config file.
4545 *) Allow certificate extensions to be added to certificate requests. These
4546 are specified in a 'req_extensions' option of the req section of the
4547 config file. They can be printed out with the -text option to req but
4548 are otherwise ignored at present.
4551 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4552 data read consists of only the final block it would not decrypted because
4553 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4554 A misplaced 'break' also meant the decrypted final block might not be
4555 copied until the next read.
4558 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4559 a few extra parameters to the DH structure: these will be useful if
4560 for example we want the value of 'q' or implement X9.42 DH.
4563 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4564 provides hooks that allow the default DSA functions or functions on a
4565 "per key" basis to be replaced. This allows hardware acceleration and
4566 hardware key storage to be handled without major modification to the
4567 library. Also added low level modexp hooks and CRYPTO_EX structure and
4568 associated functions.
4571 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4572 as "read only": it can't be written to and the buffer it points to will
4573 not be freed. Reading from a read only BIO is much more efficient than
4574 a normal memory BIO. This was added because there are several times when
4575 an area of memory needs to be read from a BIO. The previous method was
4576 to create a memory BIO and write the data to it, this results in two
4577 copies of the data and an O(n^2) reading algorithm. There is a new
4578 function BIO_new_mem_buf() which creates a read only memory BIO from
4579 an area of memory. Also modified the PKCS#7 routines to use read only
4583 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4584 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4585 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4586 but a retry condition occured while trying to read the rest.
4589 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4590 NID_pkcs7_encrypted by default: this was wrong since this should almost
4591 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4592 the encrypted data type: this is a more sensible place to put it and it
4593 allows the PKCS#12 code to be tidied up that duplicated this
4597 *) Changed obj_dat.pl script so it takes its input and output files on
4598 the command line. This should avoid shell escape redirection problems
4602 *) Initial support for certificate extension requests, these are included
4603 in things like Xenroll certificate requests. Included functions to allow
4604 extensions to be obtained and added.
4607 *) -crlf option to s_client and s_server for sending newlines as
4608 CRLF (as required by many protocols).
4611 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4613 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4614 [Ralf S. Engelschall]
4616 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4617 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4619 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4623 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4624 DH parameters/keys (q is lost during that conversion, but the resulting
4625 DH parameters contain its length).
4627 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4628 much faster than DH_generate_parameters (which creates parameters
4629 where p = 2*q + 1), and also the smaller q makes DH computations
4630 much more efficient (160-bit exponentiation instead of 1024-bit
4631 exponentiation); so this provides a convenient way to support DHE
4632 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4633 utter importance to use
4634 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4636 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4637 when such DH parameters are used, because otherwise small subgroup
4638 attacks may become possible!
4641 *) Avoid memory leak in i2d_DHparams.
4644 *) Allow the -k option to be used more than once in the enc program:
4645 this allows the same encrypted message to be read by multiple recipients.
4648 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4649 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4650 it will always use the numerical form of the OID, even if it has a short
4654 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4655 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4656 otherwise bn_mod_exp was called. In the case of hardware keys for example
4657 no private key components need be present and it might store extra data
4658 in the RSA structure, which cannot be accessed from bn_mod_exp.
4659 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4660 private key operations.
4663 *) Added support for SPARC Linux.
4666 *) pem_password_cb function type incompatibly changed from
4667 typedef int pem_password_cb(char *buf, int size, int rwflag);
4669 ....(char *buf, int size, int rwflag, void *userdata);
4670 so that applications can pass data to their callbacks:
4671 The PEM[_ASN1]_{read,write}... functions and macros now take an
4672 additional void * argument, which is just handed through whenever
4673 the password callback is called.
4674 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4676 New function SSL_CTX_set_default_passwd_cb_userdata.
4678 Compatibility note: As many C implementations push function arguments
4679 onto the stack in reverse order, the new library version is likely to
4680 interoperate with programs that have been compiled with the old
4681 pem_password_cb definition (PEM_whatever takes some data that
4682 happens to be on the stack as its last argument, and the callback
4683 just ignores this garbage); but there is no guarantee whatsoever that
4686 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4687 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4688 problems not only on Windows, but also on some Unix platforms.
4689 To avoid problematic command lines, these definitions are now in an
4690 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4691 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4694 *) MIPS III/IV assembler module is reimplemented.
4697 *) More DES library cleanups: remove references to srand/rand and
4698 delete an unused file.
4701 *) Add support for the the free Netwide assembler (NASM) under Win32,
4702 since not many people have MASM (ml) and it can be hard to obtain.
4703 This is currently experimental but it seems to work OK and pass all
4704 the tests. Check out INSTALL.W32 for info.
4707 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4708 without temporary keys kept an extra copy of the server key,
4709 and connections with temporary keys did not free everything in case
4713 *) New function RSA_check_key and new openssl rsa option -check
4714 for verifying the consistency of RSA keys.
4715 [Ulf Moeller, Bodo Moeller]
4717 *) Various changes to make Win32 compile work:
4718 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4719 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4720 comparison" warnings.
4721 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4724 *) Add a debugging option to PKCS#5 v2 key generation function: when
4725 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4726 derived keys are printed to stderr.
4729 *) Copy the flags in ASN1_STRING_dup().
4730 [Roman E. Pavlov <pre@mo.msk.ru>]
4732 *) The x509 application mishandled signing requests containing DSA
4733 keys when the signing key was also DSA and the parameters didn't match.
4735 It was supposed to omit the parameters when they matched the signing key:
4736 the verifying software was then supposed to automatically use the CA's
4737 parameters if they were absent from the end user certificate.
4739 Omitting parameters is no longer recommended. The test was also
4740 the wrong way round! This was probably due to unusual behaviour in
4741 EVP_cmp_parameters() which returns 1 if the parameters match.
4742 This meant that parameters were omitted when they *didn't* match and
4743 the certificate was useless. Certificates signed with 'ca' didn't have
4745 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4747 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4748 The interface is as follows:
4749 Applications can use
4750 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4751 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4752 "off" is now the default.
4753 The library internally uses
4754 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4755 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4756 to disable memory-checking temporarily.
4758 Some inconsistent states that previously were possible (and were
4759 even the default) are now avoided.
4761 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4762 with each memory chunk allocated; this is occasionally more helpful
4763 than just having a counter.
4765 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4767 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4771 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4772 which largely parallels "options", but is for changing API behaviour,
4773 whereas "options" are about protocol behaviour.
4774 Initial "mode" flags are:
4776 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4777 a single record has been written.
4778 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4779 retries use the same buffer location.
4780 (But all of the contents must be
4784 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4787 *) Fix problems with no-hmac etc.
4788 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4790 *) New functions RSA_get_default_method(), RSA_set_method() and
4791 RSA_get_method(). These allows replacement of RSA_METHODs without having
4792 to mess around with the internals of an RSA structure.
4795 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4796 Also really enable memory leak checks in openssl.c and in some
4798 [Chad C. Mulligan, Bodo Moeller]
4800 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4801 up the length of negative integers. This has now been simplified to just
4802 store the length when it is first determined and use it later, rather
4803 than trying to keep track of where data is copied and updating it to
4805 [Steve Henson, reported by Brien Wheeler
4806 <bwheeler@authentica-security.com>]
4808 *) Add a new function PKCS7_signatureVerify. This allows the verification
4809 of a PKCS#7 signature but with the signing certificate passed to the
4810 function itself. This contrasts with PKCS7_dataVerify which assumes the
4811 certificate is present in the PKCS#7 structure. This isn't always the
4812 case: certificates can be omitted from a PKCS#7 structure and be
4813 distributed by "out of band" means (such as a certificate database).
4816 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4817 function prototypes in pem.h, also change util/mkdef.pl to add the
4818 necessary function names.
4821 *) mk1mf.pl (used by Windows builds) did not properly read the
4822 options set by Configure in the top level Makefile, and Configure
4823 was not even able to write more than one option correctly.
4824 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4827 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4828 file to be loaded from a BIO or FILE pointer. The BIO version will
4829 for example allow memory BIOs to contain config info.
4832 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4833 Whoever hopes to achieve shared-library compatibility across versions
4834 must use this, not the compile-time macro.
4835 (Exercise 0.9.4: Which is the minimum library version required by
4837 Note: All this applies only to multi-threaded programs, others don't
4841 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4842 through a BIO pair triggered the default case, i.e.
4843 SSLerr(...,SSL_R_UNKNOWN_STATE).
4846 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4847 can use the SSL library even if none of the specific BIOs is
4851 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4852 for the encoded length.
4853 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4855 *) Add initial documentation of the X509V3 functions.
4858 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4859 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4860 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4861 secure PKCS#8 private key format with a high iteration count.
4864 *) Fix determination of Perl interpreter: A perl or perl5
4865 _directory_ in $PATH was also accepted as the interpreter.
4866 [Ralf S. Engelschall]
4868 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4869 wrong with it but it was very old and did things like calling
4870 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4874 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4875 to use the new extension code.
4878 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4879 with macros. This should make it easier to change their form, add extra
4880 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4884 *) Add to configuration table a new entry that can specify an alternative
4885 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4886 according to Mark Crispin <MRC@Panda.COM>.
4890 *) DES CBC did not update the IV. Weird.
4893 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4894 Changing the behaviour of the former might break existing programs --
4895 where IV updating is needed, des_ncbc_encrypt can be used.
4898 *) When bntest is run from "make test" it drives bc to check its
4899 calculations, as well as internally checking them. If an internal check
4900 fails, it needs to cause bc to give a non-zero result or make test carries
4901 on without noticing the failure. Fixed.
4904 *) DES library cleanups.
4907 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4908 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4909 ciphers. NOTE: although the key derivation function has been verified
4910 against some published test vectors it has not been extensively tested
4911 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4915 *) Instead of "mkdir -p", which is not fully portable, use new
4916 Perl script "util/mkdir-p.pl".
4919 *) Rewrite the way password based encryption (PBE) is handled. It used to
4920 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4921 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4922 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4923 the 'parameter' field of the AlgorithmIdentifier is passed to the
4924 underlying key generation function so it must do its own ASN1 parsing.
4925 This has also changed the EVP_PBE_CipherInit() function which now has a
4926 'parameter' argument instead of literal salt and iteration count values
4927 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4930 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4931 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4932 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4933 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4934 value was just used as a "magic string" and not used directly its
4935 value doesn't matter.
4938 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4942 *) "linux-sparc64" configuration (ultrapenguin).
4943 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4944 "linux-sparc" configuration.
4945 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4947 *) config now generates no-xxx options for missing ciphers.
4950 *) Support the EBCDIC character set (work in progress).
4951 File ebcdic.c not yet included because it has a different license.
4952 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4954 *) Support BS2000/OSD-POSIX.
4955 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4957 *) Make callbacks for key generation use void * instead of char *.
4960 *) Make S/MIME samples compile (not yet tested).
4963 *) Additional typesafe stacks.
4966 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4970 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4972 *) New configuration variant "sco5-gcc".
4974 *) Updated some demos.
4975 [Sean O Riordain, Wade Scholine]
4977 *) Add missing BIO_free at exit of pkcs12 application.
4980 *) Fix memory leak in conf.c.
4983 *) Updates for Win32 to assembler version of MD5.
4986 *) Set #! path to perl in apps/der_chop to where we found it
4987 instead of using a fixed path.
4990 *) SHA library changes for irix64-mips4-cc.
4993 *) Improvements for VMS support.
4997 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4999 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5000 This also avoids the problems with SC4.2 and unpatched SC5.
5001 [Andy Polyakov <appro@fy.chalmers.se>]
5003 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5004 These are required because of the typesafe stack would otherwise break
5005 existing code. If old code used a structure member which used to be STACK
5006 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5007 sk_num or sk_value it would produce an error because the num, data members
5008 are not present in STACK_OF. Now it just produces a warning. sk_set
5009 replaces the old method of assigning a value to sk_value
5010 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5011 that does this will no longer work (and should use sk_set instead) but
5012 this could be regarded as a "questionable" behaviour anyway.
5015 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5016 correctly handle encrypted S/MIME data.
5019 *) Change type of various DES function arguments from des_cblock
5020 (which means, in function argument declarations, pointer to char)
5021 to des_cblock * (meaning pointer to array with 8 char elements),
5022 which allows the compiler to do more typechecking; it was like
5023 that back in SSLeay, but with lots of ugly casts.
5025 Introduce new type const_des_cblock.
5028 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5029 problems: find RecipientInfo structure that matches recipient certificate
5030 and initialise the ASN1 structures properly based on passed cipher.
5033 *) Belatedly make the BN tests actually check the results.
5036 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5037 to and from BNs: it was completely broken. New compilation option
5038 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5039 key elements as negative integers.
5042 *) Reorganize and speed up MD5.
5043 [Andy Polyakov <appro@fy.chalmers.se>]
5046 [Richard Levitte <richard@levitte.org>]
5048 *) New option -out to asn1parse to allow the parsed structure to be
5049 output to a file. This is most useful when combined with the -strparse
5050 option to examine the output of things like OCTET STRINGS.
5053 *) Make SSL library a little more fool-proof by not requiring any longer
5054 that SSL_set_{accept,connect}_state be called before
5055 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5056 in many applications because usually everything *appeared* to work as
5057 intended anyway -- now it really works as intended).
5060 *) Move openssl.cnf out of lib/.
5063 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5064 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5065 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5066 [Ralf S. Engelschall]
5068 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5069 handle PKCS#7 enveloped data properly.
5070 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5072 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5073 copying pointers. The cert_st handling is changed by this in
5074 various ways (and thus what used to be known as ctx->default_cert
5075 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5076 any longer when s->cert does not give us what we need).
5077 ssl_cert_instantiate becomes obsolete by this change.
5078 As soon as we've got the new code right (possibly it already is?),
5079 we have solved a couple of bugs of the earlier code where s->cert
5080 was used as if it could not have been shared with other SSL structures.
5082 Note that using the SSL API in certain dirty ways now will result
5083 in different behaviour than observed with earlier library versions:
5084 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5085 does not influence s as it used to.
5087 In order to clean up things more thoroughly, inside SSL_SESSION
5088 we don't use CERT any longer, but a new structure SESS_CERT
5089 that holds per-session data (if available); currently, this is
5090 the peer's certificate chain and, for clients, the server's certificate
5091 and temporary key. CERT holds only those values that can have
5092 meaningful defaults in an SSL_CTX.
5095 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5096 from the internal representation. Various PKCS#7 fixes: remove some
5097 evil casts and set the enc_dig_alg field properly based on the signing
5101 *) Allow PKCS#12 password to be set from the command line or the
5102 environment. Let 'ca' get its config file name from the environment
5103 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5107 *) Allow certificate policies extension to use an IA5STRING for the
5108 organization field. This is contrary to the PKIX definition but
5109 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5113 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5114 without disallowing inline assembler and the like for non-pedantic builds.
5117 *) Support Borland C++ builder.
5118 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5123 *) SHA-1 cleanups and performance enhancements.
5124 [Andy Polyakov <appro@fy.chalmers.se>]
5126 *) Sparc v8plus assembler for the bignum library.
5127 [Andy Polyakov <appro@fy.chalmers.se>]
5129 *) Accept any -xxx and +xxx compiler options in Configure.
5132 *) Update HPUX configuration.
5135 *) Add missing sk_<type>_unshift() function to safestack.h
5136 [Ralf S. Engelschall]
5138 *) New function SSL_CTX_use_certificate_chain_file that sets the
5139 "extra_cert"s in addition to the certificate. (This makes sense
5140 only for "PEM" format files, as chains as a whole are not
5144 *) Support verify_depth from the SSL API.
5145 x509_vfy.c had what can be considered an off-by-one-error:
5146 Its depth (which was not part of the external interface)
5147 was actually counting the number of certificates in a chain;
5148 now it really counts the depth.
5151 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5152 instead of X509err, which often resulted in confusing error
5153 messages since the error codes are not globally unique
5154 (e.g. an alleged error in ssl3_accept when a certificate
5155 didn't match the private key).
5157 *) New function SSL_CTX_set_session_id_context that allows to set a default
5158 value (so that you don't need SSL_set_session_id_context for each
5159 connection using the SSL_CTX).
5162 *) OAEP decoding bug fix.
5165 *) Support INSTALL_PREFIX for package builders, as proposed by
5169 *) New Configure options "threads" and "no-threads". For systems
5170 where the proper compiler options are known (currently Solaris
5171 and Linux), "threads" is the default.
5174 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5177 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5178 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5179 such as /usr/local/bin.
5182 *) "make linux-shared" to build shared libraries.
5183 [Niels Poppe <niels@netbox.org>]
5185 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5188 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5189 extension adding in x509 utility.
5192 *) Remove NOPROTO sections and error code comments.
5195 *) Partial rewrite of the DEF file generator to now parse the ANSI
5199 *) New Configure options --prefix=DIR and --openssldir=DIR.
5202 *) Complete rewrite of the error code script(s). It is all now handled
5203 by one script at the top level which handles error code gathering,
5204 header rewriting and C source file generation. It should be much better
5205 than the old method: it now uses a modified version of Ulf's parser to
5206 read the ANSI prototypes in all header files (thus the old K&R definitions
5207 aren't needed for error creation any more) and do a better job of
5208 translating function codes into names. The old 'ASN1 error code imbedded
5209 in a comment' is no longer necessary and it doesn't use .err files which
5210 have now been deleted. Also the error code call doesn't have to appear all
5211 on one line (which resulted in some large lines...).
5214 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5217 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5218 0 (which usually indicates a closed connection), but continue reading.
5221 *) Fix some race conditions.
5224 *) Add support for CRL distribution points extension. Add Certificate
5225 Policies and CRL distribution points documentation.
5228 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5231 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5232 8 of keying material. Merlin has also confirmed interop with this fix
5233 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5234 [Merlin Hughes <merlin@baltimore.ie>]
5236 *) Fix lots of warnings.
5237 [Richard Levitte <levitte@stacken.kth.se>]
5239 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5240 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5241 [Richard Levitte <levitte@stacken.kth.se>]
5243 *) Fix problems with sizeof(long) == 8.
5244 [Andy Polyakov <appro@fy.chalmers.se>]
5246 *) Change functions to ANSI C.
5249 *) Fix typos in error codes.
5250 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5252 *) Remove defunct assembler files from Configure.
5255 *) SPARC v8 assembler BIGNUM implementation.
5256 [Andy Polyakov <appro@fy.chalmers.se>]
5258 *) Support for Certificate Policies extension: both print and set.
5259 Various additions to support the r2i method this uses.
5262 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5263 return a const string when you are expecting an allocated buffer.
5266 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5267 types DirectoryString and DisplayText.
5270 *) Add code to allow r2i extensions to access the configuration database,
5271 add an LHASH database driver and add several ctx helper functions.
5274 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5275 fail when they extended the size of a BIGNUM.
5278 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5279 support typesafe stack.
5282 *) Fix typo in SSL_[gs]et_options().
5283 [Nils Frostberg <nils@medcom.se>]
5285 *) Delete various functions and files that belonged to the (now obsolete)
5286 old X509V3 handling code.
5289 *) New Configure option "rsaref".
5292 *) Don't auto-generate pem.h.
5295 *) Introduce type-safe ASN.1 SETs.
5298 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5299 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5301 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5302 that links with OpenSSL (well at least cause lots of warnings), but fear
5303 not: the conversion is trivial, and it eliminates loads of evil casts. A
5304 few STACKed things have been converted already. Feel free to convert more.
5305 In the fullness of time, I'll do away with the STACK type altogether.
5308 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5309 specified in <certfile> by updating the entry in the index.txt file.
5310 This way one no longer has to edit the index.txt file manually for
5311 revoking a certificate. The -revoke option does the gory details now.
5312 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5314 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5315 `-text' option at all and this way the `-noout -text' combination was
5316 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5317 [Ralf S. Engelschall]
5319 *) Make sure a corresponding plain text error message exists for the
5320 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5321 verify callback function determined that a certificate was revoked.
5322 [Ralf S. Engelschall]
5324 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5325 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5326 all available cipers including rc5, which was forgotten until now.
5327 In order to let the testing shell script know which algorithms
5328 are available, a new (up to now undocumented) command
5329 "openssl list-cipher-commands" is used.
5332 *) Bugfix: s_client occasionally would sleep in select() when
5333 it should have checked SSL_pending() first.
5336 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5337 the raw DSA values prior to ASN.1 encoding.
5340 *) Tweaks to Configure
5341 [Niels Poppe <niels@netbox.org>]
5343 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5347 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5350 *) New config option to avoid instructions that are illegal on the 80386.
5351 The default code is faster, but requires at least a 486.
5354 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5355 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5356 same as SSL2_VERSION anyway.
5359 *) New "-showcerts" option for s_client.
5362 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5363 application. Various cleanups and fixes.
5366 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5367 modify error routines to work internally. Add error codes and PBE init
5368 to library startup routines.
5371 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5372 packing functions to asn1 and evp. Changed function names and error
5373 codes along the way.
5376 *) PKCS12 integration: and so it begins... First of several patches to
5377 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5378 objects to objects.h
5381 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5382 and display support for Thawte strong extranet extension.
5385 *) Add LinuxPPC support.
5386 [Jeff Dubrule <igor@pobox.org>]
5388 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5389 bn_div_words in alpha.s.
5390 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5392 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5393 OAEP isn't supported when OpenSSL is built with RSAref.
5394 [Ulf Moeller <ulf@fitug.de>]
5396 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5397 so they no longer are missing under -DNOPROTO.
5398 [Soren S. Jorvang <soren@t.dk>]
5401 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5403 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5404 doesn't work when the session is reused. Coming soon!
5407 *) Fix a security hole, that allows sessions to be reused in the wrong
5408 context thus bypassing client cert protection! All software that uses
5409 client certs and session caches in multiple contexts NEEDS PATCHING to
5410 allow session reuse! A fuller solution is in the works.
5411 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5413 *) Some more source tree cleanups (removed obsolete files
5414 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5415 permission on "config" script to be executable) and a fix for the INSTALL
5417 [Ulf Moeller <ulf@fitug.de>]
5419 *) Remove some legacy and erroneous uses of malloc, free instead of
5421 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5423 *) Make rsa_oaep_test return non-zero on error.
5424 [Ulf Moeller <ulf@fitug.de>]
5426 *) Add support for native Solaris shared libraries. Configure
5427 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5428 if someone would make that last step automatic.
5429 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5431 *) ctx_size was not built with the right compiler during "make links". Fixed.
5434 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5435 except NULL ciphers". This means the default cipher list will no longer
5436 enable NULL ciphers. They need to be specifically enabled e.g. with
5437 the string "DEFAULT:eNULL".
5440 *) Fix to RSA private encryption routines: if p < q then it would
5441 occasionally produce an invalid result. This will only happen with
5442 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5445 *) Be less restrictive and allow also `perl util/perlpath.pl
5446 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5447 because this way one can also use an interpreter named `perl5' (which is
5448 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5449 installed as `perl').
5450 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5452 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5453 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5455 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5456 advapi32.lib to Win32 build and change the pem test comparision
5457 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5458 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5459 and crypto/des/ede_cbcm_enc.c.
5462 *) DES quad checksum was broken on big-endian architectures. Fixed.
5465 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5466 Win32 test batch file so it (might) work again. The Win32 test batch file
5467 is horrible: I feel ill....
5470 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5471 in e_os.h. Audit of header files to check ANSI and non ANSI
5472 sections: 10 functions were absent from non ANSI section and not exported
5473 from Windows DLLs. Fixed up libeay.num for new functions.
5476 *) Make `openssl version' output lines consistent.
5477 [Ralf S. Engelschall]
5479 *) Fix Win32 symbol export lists for BIO functions: Added
5480 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5481 to ms/libeay{16,32}.def.
5482 [Ralf S. Engelschall]
5484 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5485 fine under Unix and passes some trivial tests I've now added. But the
5486 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5487 added to make sure no one expects that this stuff really works in the
5488 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5489 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5491 [Ralf S. Engelschall]
5493 *) Fix the generation of two part addresses in perl.
5494 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5496 *) Add config entry for Linux on MIPS.
5497 [John Tobey <jtobey@channel1.com>]
5499 *) Make links whenever Configure is run, unless we are on Windoze.
5502 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5503 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5507 *) Add a useful kludge to allow package maintainers to specify compiler and
5508 other platforms details on the command line without having to patch the
5509 Configure script everytime: One now can use ``perl Configure
5510 <id>:<details>'', i.e. platform ids are allowed to have details appended
5511 to them (seperated by colons). This is treated as there would be a static
5512 pre-configured entry in Configure's %table under key <id> with value
5513 <details> and ``perl Configure <id>'' is called. So, when you want to
5514 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5515 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5516 now, which overrides the FreeBSD-elf entry on-the-fly.
5517 [Ralf S. Engelschall]
5519 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5522 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5523 on the `perl Configure ...' command line. This way one can compile
5524 OpenSSL libraries with Position Independent Code (PIC) which is needed
5525 for linking it into DSOs.
5526 [Ralf S. Engelschall]
5528 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5532 *) Cleaned up the LICENSE document: The official contact for any license
5533 questions now is the OpenSSL core team under openssl-core@openssl.org.
5534 And add a paragraph about the dual-license situation to make sure people
5535 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5536 to the OpenSSL toolkit.
5537 [Ralf S. Engelschall]
5539 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5540 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5541 Additonally cleaned up the `make links' target: Remove unnecessary
5542 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5543 to speed processing and no longer clutter the display with confusing
5544 stuff. Instead only the actually done links are displayed.
5545 [Ralf S. Engelschall]
5547 *) Permit null encryption ciphersuites, used for authentication only. It used
5548 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5549 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5553 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5554 signed attributes when verifying signatures (this would break them),
5555 the detached data encoding was wrong and public keys obtained using
5556 X509_get_pubkey() weren't freed.
5559 *) Add text documentation for the BUFFER functions. Also added a work around
5560 to a Win95 console bug. This was triggered by the password read stuff: the
5561 last character typed gets carried over to the next fread(). If you were
5562 generating a new cert request using 'req' for example then the last
5563 character of the passphrase would be CR which would then enter the first
5567 *) Added the new `Includes OpenSSL Cryptography Software' button as
5568 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5569 button and can be used by applications based on OpenSSL to show the
5570 relationship to the OpenSSL project.
5571 [Ralf S. Engelschall]
5573 *) Remove confusing variables in function signatures in files
5574 ssl/ssl_lib.c and ssl/ssl.h.
5575 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5577 *) Don't install bss_file.c under PREFIX/include/
5578 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5580 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5581 functions that return function pointers and has support for NT specific
5582 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5583 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5584 unsigned to signed types: this was killing the Win32 compile.
5587 *) Add new certificate file to stack functions,
5588 SSL_add_dir_cert_subjects_to_stack() and
5589 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5590 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5591 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5592 This means that Apache-SSL and similar packages don't have to mess around
5593 to add as many CAs as they want to the preferred list.
5596 *) Experiment with doxygen documentation. Currently only partially applied to
5598 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5599 openssl.doxy as the configuration file.
5602 *) Get rid of remaining C++-style comments which strict C compilers hate.
5603 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5605 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5606 compiled in by default: it has problems with large keys.
5609 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5610 DH private keys and/or callback functions which directly correspond to
5611 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5612 is needed for applications which have to configure certificates on a
5613 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5615 For the RSA certificate situation is makes no difference, but
5616 for the DSA certificate situation this fixes the "no shared cipher"
5617 problem where the OpenSSL cipher selection procedure failed because the
5618 temporary keys were not overtaken from the context and the API provided
5619 no way to reconfigure them.
5620 The new functions now let applications reconfigure the stuff and they
5621 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5622 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5623 non-public-API function ssl_cert_instantiate() is used as a helper
5624 function and also to reduce code redundancy inside ssl_rsa.c.
5625 [Ralf S. Engelschall]
5627 *) Move s_server -dcert and -dkey options out of the undocumented feature
5628 area because they are useful for the DSA situation and should be
5629 recognized by the users.
5630 [Ralf S. Engelschall]
5632 *) Fix the cipher decision scheme for export ciphers: the export bits are
5633 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5634 SSL_EXP_MASK. So, the original variable has to be used instead of the
5635 already masked variable.
5636 [Richard Levitte <levitte@stacken.kth.se>]
5638 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5639 [Richard Levitte <levitte@stacken.kth.se>]
5641 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5642 from `int' to `unsigned int' because it's a length and initialized by
5643 EVP_DigestFinal() which expects an `unsigned int *'.
5644 [Richard Levitte <levitte@stacken.kth.se>]
5646 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5647 script. Instead use the usual Shell->Perl transition trick.
5648 [Ralf S. Engelschall]
5650 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5651 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5652 -noout -modulus' as it's already the case for `openssl rsa -noout
5653 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5654 currently the public key is printed (a decision which was already done by
5655 `openssl dsa -modulus' in the past) which serves a similar purpose.
5656 Additionally the NO_RSA no longer completely removes the whole -modulus
5657 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5659 [Ralf S. Engelschall]
5661 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5662 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5663 [Arne Ansper <arne@ats.cyber.ee>]
5665 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5666 to be added. Now both 'req' and 'ca' can use new objects defined in the
5670 *) Add cool BIO that does syslog (or event log on NT).
5671 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5673 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5674 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5675 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5676 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5679 *) Add preliminary config info for new extension code.
5682 *) Make RSA_NO_PADDING really use no padding.
5683 [Ulf Moeller <ulf@fitug.de>]
5685 *) Generate errors when private/public key check is done.
5688 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5689 for some CRL extensions and new objects added.
5692 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5693 key usage extension and fuller support for authority key id.
5696 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5697 padding method for RSA, which is recommended for new applications in PKCS
5698 #1 v2.0 (RFC 2437, October 1998).
5699 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5700 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5701 against Bleichbacher's attack on RSA.
5702 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5705 *) Updates to the new SSL compression code
5706 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5708 *) Fix so that the version number in the master secret, when passed
5709 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5710 (because the server will not accept higher), that the version number
5711 is 0x03,0x01, not 0x03,0x00
5712 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5714 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5715 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5716 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5719 *) Support for RAW extensions where an arbitrary extension can be
5720 created by including its DER encoding. See apps/openssl.cnf for
5724 *) Make sure latest Perl versions don't interpret some generated C array
5725 code as Perl array code in the crypto/err/err_genc.pl script.
5726 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5728 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5729 not many people have the assembler. Various Win32 compilation fixes and
5730 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5734 *) Modify configure script 'Configure' to automatically create crypto/date.h
5735 file under Win32 and also build pem.h from pem.org. New script
5736 util/mkfiles.pl to create the MINFO file on environments that can't do a
5737 'make files': perl util/mkfiles.pl >MINFO should work.
5740 *) Major rework of DES function declarations, in the pursuit of correctness
5741 and purity. As a result, many evil casts evaporated, and some weirdness,
5742 too. You may find this causes warnings in your code. Zapping your evil
5743 casts will probably fix them. Mostly.
5746 *) Fix for a typo in asn1.h. Bug fix to object creation script
5747 obj_dat.pl. It considered a zero in an object definition to mean
5748 "end of object": none of the objects in objects.h have any zeros
5749 so it wasn't spotted.
5750 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5752 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5753 Masking (CBCM). In the absence of test vectors, the best I have been able
5754 to do is check that the decrypt undoes the encrypt, so far. Send me test
5755 vectors if you have them.
5758 *) Correct calculation of key length for export ciphers (too much space was
5759 allocated for null ciphers). This has not been tested!
5762 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5763 message is now correct (it understands "crypto" and "ssl" on its
5764 command line). There is also now an "update" option. This will update
5765 the util/ssleay.num and util/libeay.num files with any new functions.
5767 perl util/mkdef.pl crypto ssl update
5768 it will update them.
5771 *) Overhauled the Perl interface (perl/*):
5772 - ported BN stuff to OpenSSL's different BN library
5773 - made the perl/ source tree CVS-aware
5774 - renamed the package from SSLeay to OpenSSL (the files still contain
5775 their history because I've copied them in the repository)
5776 - removed obsolete files (the test scripts will be replaced
5777 by better Test::Harness variants in the future)
5778 [Ralf S. Engelschall]
5780 *) First cut for a very conservative source tree cleanup:
5781 1. merge various obsolete readme texts into doc/ssleay.txt
5782 where we collect the old documents and readme texts.
5783 2. remove the first part of files where I'm already sure that we no
5784 longer need them because of three reasons: either they are just temporary
5785 files which were left by Eric or they are preserved original files where
5786 I've verified that the diff is also available in the CVS via "cvs diff
5787 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5788 the crypto/md/ stuff).
5789 [Ralf S. Engelschall]
5791 *) More extension code. Incomplete support for subject and issuer alt
5792 name, issuer and authority key id. Change the i2v function parameters
5793 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5794 what that's for :-) Fix to ASN1 macro which messed up
5795 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5798 *) Preliminary support for ENUMERATED type. This is largely copied from the
5802 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5803 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5805 *) Make sure `make rehash' target really finds the `openssl' program.
5806 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5808 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5809 like to hear about it if this slows down other processors.
5812 *) Add CygWin32 platform information to Configure script.
5813 [Alan Batie <batie@aahz.jf.intel.com>]
5815 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5816 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5818 *) New program nseq to manipulate netscape certificate sequences
5821 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5825 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5826 but the BN code had some problems that would cause failures when
5827 doing certificate verification and some other functions.
5828 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5830 *) Add ASN1 and PEM code to support netscape certificate sequences.
5833 *) Add ASN1 and PEM code to support netscape certificate sequences.
5836 *) Add several PKIX and private extended key usage OIDs.
5839 *) Modify the 'ca' program to handle the new extension code. Modify
5840 openssl.cnf for new extension format, add comments.
5843 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5844 and add a sample to openssl.cnf so req -x509 now adds appropriate
5848 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5849 error code, add initial support to X509_print() and x509 application.
5852 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5853 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5854 stuff is currently isolated and isn't even compiled yet.
5857 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5858 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5859 Removed the versions check from X509 routines when loading extensions:
5860 this allows certain broken certificates that don't set the version
5861 properly to be processed.
5864 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5865 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5866 can still be regenerated with "make depend".
5869 *) Spelling mistake in C version of CAST-128.
5870 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5872 *) Changes to the error generation code. The perl script err-code.pl
5873 now reads in the old error codes and retains the old numbers, only
5874 adding new ones if necessary. It also only changes the .err files if new
5875 codes are added. The makefiles have been modified to only insert errors
5876 when needed (to avoid needlessly modifying header files). This is done
5877 by only inserting errors if the .err file is newer than the auto generated
5878 C file. To rebuild all the error codes from scratch (the old behaviour)
5879 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5880 or delete all the .err files.
5883 *) CAST-128 was incorrectly implemented for short keys. The C version has
5884 been fixed, but is untested. The assembler versions are also fixed, but
5885 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5886 to regenerate it if needed.
5887 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5888 Hagino <itojun@kame.net>]
5890 *) File was opened incorrectly in randfile.c.
5891 [Ulf Möller <ulf@fitug.de>]
5893 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5894 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5895 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5896 al: it's just almost always a UTCTime. Note this patch adds new error
5897 codes so do a "make errors" if there are problems.
5900 *) Correct Linux 1 recognition in config.
5901 [Ulf Möller <ulf@fitug.de>]
5903 *) Remove pointless MD5 hash when using DSA keys in ca.
5904 [Anonymous <nobody@replay.com>]
5906 *) Generate an error if given an empty string as a cert directory. Also
5907 generate an error if handed NULL (previously returned 0 to indicate an
5908 error, but didn't set one).
5909 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5911 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5914 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5915 parameters. This was causing a warning which killed off the Win32 compile.
5918 *) Remove C++ style comments from crypto/bn/bn_local.h.
5919 [Neil Costigan <neil.costigan@celocom.com>]
5921 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5922 based on a text string, looking up short and long names and finally
5923 "dot" format. The "dot" format stuff didn't work. Added new function
5924 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5925 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5926 OID is not part of the table.
5929 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5930 X509_LOOKUP_by_alias().
5933 *) Sort openssl functions by name.
5936 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5937 encryption from sample DSA keys (in case anyone is interested the password
5941 *) Make _all_ *_free functions accept a NULL pointer.
5942 [Frans Heymans <fheymans@isaserver.be>]
5944 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5946 [Anonymous <nobody@replay.com>]
5948 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5949 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5951 *) Don't blow it for numeric -newkey arguments to apps/req.
5952 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5954 *) Temp key "for export" tests were wrong in s3_srvr.c.
5955 [Anonymous <nobody@replay.com>]
5957 *) Add prototype for temp key callback functions
5958 SSL_CTX_set_tmp_{rsa,dh}_callback().
5961 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5962 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5965 *) X509_name_add_entry() freed the wrong thing after an error.
5966 [Arne Ansper <arne@ats.cyber.ee>]
5968 *) rsa_eay.c would attempt to free a NULL context.
5969 [Arne Ansper <arne@ats.cyber.ee>]
5971 *) BIO_s_socket() had a broken should_retry() on Windoze.
5972 [Arne Ansper <arne@ats.cyber.ee>]
5974 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5975 [Arne Ansper <arne@ats.cyber.ee>]
5977 *) Make sure the already existing X509_STORE->depth variable is initialized
5978 in X509_STORE_new(), but document the fact that this variable is still
5979 unused in the certificate verification process.
5980 [Ralf S. Engelschall]
5982 *) Fix the various library and apps files to free up pkeys obtained from
5983 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5986 *) Fix reference counting in X509_PUBKEY_get(). This makes
5987 demos/maurice/example2.c work, amongst others, probably.
5988 [Steve Henson and Ben Laurie]
5990 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5991 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5992 are no longer created. This way we have a single and consistent command
5993 line interface `openssl <command>', similar to `cvs <command>'.
5994 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5996 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5997 BIT STRING wrapper always have zero unused bits.
6000 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6003 *) Make the top-level INSTALL documentation easier to understand.
6006 *) Makefiles updated to exit if an error occurs in a sub-directory
6007 make (including if user presses ^C) [Paul Sutton]
6009 *) Make Montgomery context stuff explicit in RSA data structure.
6012 *) Fix build order of pem and err to allow for generated pem.h.
6015 *) Fix renumbering bug in X509_NAME_delete_entry().
6018 *) Enhanced the err-ins.pl script so it makes the error library number
6019 global and can add a library name. This is needed for external ASN1 and
6020 other error libraries.
6023 *) Fixed sk_insert which never worked properly.
6026 *) Fix ASN1 macros so they can handle indefinite length construted
6027 EXPLICIT tags. Some non standard certificates use these: they can now
6031 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6032 into a single doc/ssleay.txt bundle. This way the information is still
6033 preserved but no longer messes up this directory. Now it's new room for
6034 the new set of documenation files.
6035 [Ralf S. Engelschall]
6037 *) SETs were incorrectly DER encoded. This was a major pain, because they
6038 shared code with SEQUENCEs, which aren't coded the same. This means that
6039 almost everything to do with SETs or SEQUENCEs has either changed name or
6040 number of arguments.
6041 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6043 *) Fix test data to work with the above.
6046 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6047 was already fixed by Eric for 0.9.1 it seems.
6048 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6050 *) Autodetect FreeBSD3.
6053 *) Fix various bugs in Configure. This affects the following platforms:
6057 unixware-2.0-pentium
6061 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6062 before they are needed.
6065 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6069 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6071 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6072 changed SSLeay to OpenSSL in version strings.
6073 [Ralf S. Engelschall]
6075 *) Some fixups to the top-level documents.
6078 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6079 because the symlink to include/ was missing.
6080 [Ralf S. Engelschall]
6082 *) Incorporated the popular no-RSA/DSA-only patches
6083 which allow to compile a RSA-free SSLeay.
6084 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6086 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6087 when "ssleay" is still not found.
6088 [Ralf S. Engelschall]
6090 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6091 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6093 *) Updated the README file.
6094 [Ralf S. Engelschall]
6096 *) Added various .cvsignore files in the CVS repository subdirs
6097 to make a "cvs update" really silent.
6098 [Ralf S. Engelschall]
6100 *) Recompiled the error-definition header files and added
6101 missing symbols to the Win32 linker tables.
6102 [Ralf S. Engelschall]
6104 *) Cleaned up the top-level documents;
6105 o new files: CHANGES and LICENSE
6106 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6107 o merged COPYRIGHT into LICENSE
6108 o removed obsolete TODO file
6109 o renamed MICROSOFT to INSTALL.W32
6110 [Ralf S. Engelschall]
6112 *) Removed dummy files from the 0.9.1b source tree:
6113 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6114 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6115 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6116 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6117 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6118 [Ralf S. Engelschall]
6120 *) Added various platform portability fixes.
6123 *) The Genesis of the OpenSSL rpject:
6124 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6125 Young and Tim J. Hudson created while they were working for C2Net until
6127 [The OpenSSL Project]
6130 Changes between 0.9.0b and 0.9.1b [not released]
6132 *) Updated a few CA certificates under certs/
6135 *) Changed some BIGNUM api stuff.
6138 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6139 DGUX x86, Linux Alpha, etc.
6142 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6143 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6147 *) Add -strparse option to asn1pars program which parses nested
6149 [Dr Stephen Henson <shenson@bigfoot.com>]
6151 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6154 *) DSA fix for "ca" program.
6157 *) Added "-genkey" option to "dsaparam" program.
6160 *) Added RIPE MD160 (rmd160) message digest.
6163 *) Added -a (all) option to "ssleay version" command.
6166 *) Added PLATFORM define which is the id given to Configure.
6169 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6172 *) Extended the ASN.1 parser routines.
6175 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6178 *) Added a BN_CTX to the BN library.
6181 *) Fixed the weak key values in DES library
6184 *) Changed API in EVP library for cipher aliases.
6187 *) Added support for RC2/64bit cipher.
6190 *) Converted the lhash library to the crypto/mem.c functions.
6193 *) Added more recognized ASN.1 object ids.
6196 *) Added more RSA padding checks for SSL/TLS.
6199 *) Added BIO proxy/filter functionality.
6202 *) Added extra_certs to SSL_CTX which can be used
6203 send extra CA certificates to the client in the CA cert chain sending
6204 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6207 *) Now Fortezza is denied in the authentication phase because
6208 this is key exchange mechanism is not supported by SSLeay at all.
6211 *) Additional PKCS1 checks.
6214 *) Support the string "TLSv1" for all TLS v1 ciphers.
6217 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6218 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6221 *) Fixed a few memory leaks.
6224 *) Fixed various code and comment typos.
6227 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6228 bytes sent in the client random.
6229 [Edward Bishop <ebishop@spyglass.com>]