5 Changes between 0.9.7e and 0.9.7f [XX xxx XXXX]
7 *) Make an explicit check during certificate validation to see that
8 the CA setting in each certificate on the chain is correct. As a
9 side effect always do the following basic checks on extensions,
10 not just when there's an associated purpose to the check:
12 - if there is an unhandled critical extension (unless the user
13 has chosen to ignore this fault)
14 - if the path length has been exceeded (if one is set at all)
15 - that certain extensions fit the associated purpose (if one has
19 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
21 *) Avoid a race condition when CRLs are checked in a multi threaded
22 environment. This would happen due to the reordering of the revoked
23 entries during signature checking and serial number lookup. Now the
24 encoding is cached and the serial number sort performed under a lock.
25 Add new STACK function sk_is_sorted().
28 *) Add Delta CRL to the extension code.
31 *) Various fixes to s3_pkt.c so alerts are sent properly.
32 [David Holmes <d.holmes@f5.com>]
34 *) Reduce the chances of duplicate issuer name and serial numbers (in
35 violation of RFC3280) using the OpenSSL certificate creation utilities.
36 This is done by creating a random 64 bit value for the initial serial
37 number when a serial number file is created or when a self signed
38 certificate is created using 'openssl req -x509'. The initial serial
39 number file is created using 'openssl x509 -next_serial' in CA.pl
40 rather than being initialized to 1.
43 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
45 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
46 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
47 [Joe Orton, Steve Henson]
49 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
51 [Joe Orton, Steve Henson]
53 *) Make it possible to have multiple active certificates with the same
54 subject in the CA index file. This is done only if the keyword
55 'unique_subject' is set to 'no' in the main CA section (default
56 if 'CA_default') of the configuration file. The value is saved
57 with the database itself in a separate index attribute file,
58 named like the index file with '.attr' appended to the name.
61 *) X509 verify fixes. Disable broken certificate workarounds when
62 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
63 keyUsage extension present. Don't accept CRLs with unhandled critical
64 extensions: since verify currently doesn't process CRL extensions this
65 rejects a CRL with *any* critical extensions. Add new verify error codes
69 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
70 A clarification of RFC2560 will require the use of OCTET STRINGs and
71 some implementations cannot handle the current raw format. Since OpenSSL
72 copies and compares OCSP nonces as opaque blobs without any attempt at
73 parsing them this should not create any compatibility issues.
76 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
77 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
78 this HMAC (and other) operations are several times slower than OpenSSL
82 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
83 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
85 *) Use the correct content when signing type "other".
88 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
90 *) Fix various bugs revealed by running the NISCC test suite:
92 Stop out of bounds reads in the ASN1 code when presented with
93 invalid tags (CAN-2003-0543 and CAN-2003-0544).
95 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
97 If verify callback ignores invalid public key errors don't try to check
98 certificate signature with the NULL public key.
102 *) New -ignore_err option in ocsp application to stop the server
103 exiting on the first error in a request.
106 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
107 if the server requested one: as stated in TLS 1.0 and SSL 3.0
111 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
112 extra data after the compression methods not only for TLS 1.0
113 but also for SSL 3.0 (as required by the specification).
114 [Bodo Moeller; problem pointed out by Matthias Loepfe]
116 *) Change X509_certificate_type() to mark the key as exported/exportable
117 when it's 512 *bits* long, not 512 bytes.
120 *) Change AES_cbc_encrypt() so it outputs exact multiple of
121 blocks during encryption.
124 *) Various fixes to base64 BIO and non blocking I/O. On write
125 flushes were not handled properly if the BIO retried. On read
126 data was not being buffered properly and had various logic bugs.
127 This also affects blocking I/O when the data being decoded is a
131 *) Various S/MIME bugfixes and compatibility changes:
132 output correct application/pkcs7 MIME type if
133 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
134 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
135 of files as .eml work). Correctly handle very long lines in MIME
139 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
141 *) Countermeasure against the Klima-Pokorny-Rosa extension of
142 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
143 a protocol version number mismatch like a decryption error
144 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
147 *) Turn on RSA blinding by default in the default implementation
148 to avoid a timing attack. Applications that don't want it can call
149 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
150 They would be ill-advised to do so in most cases.
151 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
153 *) Change RSA blinding code so that it works when the PRNG is not
154 seeded (in this case, the secret RSA exponent is abused as
155 an unpredictable seed -- if it is not unpredictable, there
156 is no point in blinding anyway). Make RSA blinding thread-safe
157 by remembering the creator's thread ID in rsa->blinding and
158 having all other threads use local one-time blinding factors
159 (this requires more computation than sharing rsa->blinding, but
160 avoids excessive locking; and if an RSA object is not shared
161 between threads, blinding will still be very fast).
164 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
165 ENGINE as defaults for all supported algorithms irrespective of
166 the 'flags' parameter. 'flags' is now honoured, so applications
167 should make sure they are passing it correctly.
170 *) Target "mingw" now allows native Windows code to be generated in
171 the Cygwin environment as well as with the MinGW compiler.
174 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
176 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
177 via timing by performing a MAC computation even if incorrrect
178 block cipher padding has been found. This is a countermeasure
179 against active attacks where the attacker has to distinguish
180 between bad padding and a MAC verification error. (CAN-2003-0078)
182 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
183 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
184 Martin Vuagnoux (EPFL, Ilion)]
186 *) Make the no-err option work as intended. The intention with no-err
187 is not to have the whole error stack handling routines removed from
188 libcrypto, it's only intended to remove all the function name and
189 reason texts, thereby removing some of the footprint that may not
190 be interesting if those errors aren't displayed anyway.
192 NOTE: it's still possible for any application or module to have it's
193 own set of error texts inserted. The routines are there, just not
194 used by default when no-err is given.
197 *) Add support for FreeBSD on IA64.
198 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
200 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
201 Kerberos function mit_des_cbc_cksum(). Before this change,
202 the value returned by DES_cbc_cksum() was like the one from
203 mit_des_cbc_cksum(), except the bytes were swapped.
204 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
206 *) Allow an application to disable the automatic SSL chain building.
207 Before this a rather primitive chain build was always performed in
208 ssl3_output_cert_chain(): an application had no way to send the
209 correct chain if the automatic operation produced an incorrect result.
211 Now the chain builder is disabled if either:
213 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
215 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
217 The reasoning behind this is that an application would not want the
218 auto chain building to take place if extra chain certificates are
219 present and it might also want a means of sending no additional
220 certificates (for example the chain has two certificates and the
224 *) Add the possibility to build without the ENGINE framework.
225 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
227 *) Under Win32 gmtime() can return NULL: check return value in
228 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
231 *) DSA routines: under certain error conditions uninitialized BN objects
232 could be freed. Solution: make sure initialization is performed early
233 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
234 Nils Larsch <nla@trustcenter.de> via PR#459)
237 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
238 checked on reconnect on the client side, therefore session resumption
239 could still fail with a "ssl session id is different" error. This
240 behaviour is masked when SSL_OP_ALL is used due to
241 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
242 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
246 *) IA-32 assembler support enhancements: unified ELF targets, support
247 for SCO/Caldera platforms, fix for Cygwin shared build.
250 *) Add support for FreeBSD on sparc64. As a consequence, support for
251 FreeBSD on non-x86 processors is separate from x86 processors on
252 the config script, much like the NetBSD support.
253 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
255 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
257 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
260 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
261 code (06) was taken as the first octet of the session ID and the last
262 octet was ignored consequently. As a result SSLv2 client side session
263 caching could not have worked due to the session ID mismatch between
265 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
269 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
270 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
274 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
275 seems that in spite of existing for more than a year, many application
276 author have done nothing to provide the necessary callbacks, which
277 means that this particular engine will not work properly anywhere.
278 This is a very unfortunate situation which forces us, in the name
279 of usability, to give the hw_ncipher.c a static lock, which is part
281 NOTE: This is for the 0.9.7 series ONLY. This hack will never
282 appear in 0.9.8 or later. We EXPECT application authors to have
283 dealt properly with this when 0.9.8 is released (unless we actually
284 make such changes in the libcrypto locking code that changes will
285 have to be made anyway).
288 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
289 octets have been read, EOF or an error occurs. Without this change
290 some truncated ASN1 structures will not produce an error.
293 *) Disable Heimdal support, since it hasn't been fully implemented.
294 Still give the possibility to force the use of Heimdal, but with
295 warnings and a request that patches get sent to openssl-dev.
298 *) Add the VC-CE target, introduce the WINCE sysname, and add
299 INSTALL.WCE and appropriate conditionals to make it build.
300 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
302 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
303 cygssl-x.y.z.dll, where x, y and z are the major, minor and
304 edit numbers of the version.
305 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
307 *) Introduce safe string copy and catenation functions
308 (BUF_strlcpy() and BUF_strlcat()).
309 [Ben Laurie (CHATS) and Richard Levitte]
311 *) Avoid using fixed-size buffers for one-line DNs.
314 *) Add BUF_MEM_grow_clean() to avoid information leakage when
315 resizing buffers containing secrets, and use where appropriate.
318 *) Avoid using fixed size buffers for configuration file location.
321 *) Avoid filename truncation for various CA files.
324 *) Use sizeof in preference to magic numbers.
327 *) Avoid filename truncation in cert requests.
330 *) Add assertions to check for (supposedly impossible) buffer
334 *) Don't cache truncated DNS entries in the local cache (this could
335 potentially lead to a spoofing attack).
338 *) Fix various buffers to be large enough for hex/decimal
339 representations in a platform independent manner.
342 *) Add CRYPTO_realloc_clean() to avoid information leakage when
343 resizing buffers containing secrets, and use where appropriate.
346 *) Add BIO_indent() to avoid much slightly worrying code to do
350 *) Convert sprintf()/BIO_puts() to BIO_printf().
353 *) buffer_gets() could terminate with the buffer only half
357 *) Add assertions to prevent user-supplied crypto functions from
358 overflowing internal buffers by having large block sizes, etc.
361 *) New OPENSSL_assert() macro (similar to assert(), but enabled
365 *) Eliminate unused copy of key in RC4.
368 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
371 *) Fix off-by-one error in EGD path.
374 *) If RANDFILE path is too long, ignore instead of truncating.
377 *) Eliminate unused and incorrectly sized X.509 structure
381 *) Eliminate unused and dangerous function knumber().
384 *) Eliminate unused and dangerous structure, KSSL_ERR.
387 *) Protect against overlong session ID context length in an encoded
388 session object. Since these are local, this does not appear to be
392 *) Change from security patch (see 0.9.6e below) that did not affect
393 the 0.9.6 release series:
395 Remote buffer overflow in SSL3 protocol - an attacker could
396 supply an oversized master key in Kerberos-enabled versions.
400 *) Change the SSL kerb5 codes to match RFC 2712.
403 *) Make -nameopt work fully for req and add -reqopt switch.
404 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
406 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
407 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
409 *) Make sure tests can be performed even if the corresponding algorithms
410 have been removed entirely. This was also the last step to make
411 OpenSSL compilable with DJGPP under all reasonable conditions.
412 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
414 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
415 to allow version independent disabling of normally unselected ciphers,
416 which may be activated as a side-effect of selecting a single cipher.
418 (E.g., cipher list string "RSA" enables ciphersuites that are left
419 out of "ALL" because they do not provide symmetric encryption.
420 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
421 [Lutz Jaenicke, Bodo Moeller]
423 *) Add appropriate support for separate platform-dependent build
424 directories. The recommended way to make a platform-dependent
425 build directory is the following (tested on Linux), maybe with
428 # Place yourself outside of the OpenSSL source tree. In
429 # this example, the environment variable OPENSSL_SOURCE
430 # is assumed to contain the absolute OpenSSL source directory.
431 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
432 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
433 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
434 mkdir -p `dirname $F`
435 ln -s $OPENSSL_SOURCE/$F $F
438 To be absolutely sure not to disturb the source tree, a "make clean"
439 is a good thing. If it isn't successfull, don't worry about it,
440 it probably means the source directory is very clean.
443 *) Make sure any ENGINE control commands make local copies of string
444 pointers passed to them whenever necessary. Otherwise it is possible
445 the caller may have overwritten (or deallocated) the original string
446 data when a later ENGINE operation tries to use the stored values.
447 [Götz Babin-Ebell <babinebell@trustcenter.de>]
449 *) Improve diagnostics in file reading and command-line digests.
450 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
452 *) Add AES modes CFB and OFB to the object database. Correct an
453 error in AES-CFB decryption.
456 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
457 allows existing EVP_CIPHER_CTX structures to be reused after
458 calling EVP_*Final(). This behaviour is used by encryption
459 BIOs and some applications. This has the side effect that
460 applications must explicitly clean up cipher contexts with
461 EVP_CIPHER_CTX_cleanup() or they will leak memory.
464 *) Check the values of dna and dnb in bn_mul_recursive before calling
465 bn_mul_comba (a non zero value means the a or b arrays do not contain
466 n2 elements) and fallback to bn_mul_normal if either is not zero.
469 *) Fix escaping of non-ASCII characters when using the -subj option
470 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
473 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
474 form for "surname", serialNumber has no short form.
475 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
476 therefore remove "mail" short name for "internet 7".
477 The OID for unique identifiers in X509 certificates is
478 x500UniqueIdentifier, not uniqueIdentifier.
479 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
482 *) Add an "init" command to the ENGINE config module and auto initialize
483 ENGINEs. Without any "init" command the ENGINE will be initialized
484 after all ctrl commands have been executed on it. If init=1 the
485 ENGINE is initailized at that point (ctrls before that point are run
486 on the uninitialized ENGINE and after on the initialized one). If
487 init=0 then the ENGINE will not be iniatialized at all.
490 *) Fix the 'app_verify_callback' interface so that the user-defined
491 argument is actually passed to the callback: In the
492 SSL_CTX_set_cert_verify_callback() prototype, the callback
493 declaration has been changed from
496 int (*cb)(X509_STORE_CTX *,void *);
497 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
498 i=s->ctx->app_verify_callback(&ctx)
499 has been changed into
500 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
502 To update applications using SSL_CTX_set_cert_verify_callback(),
503 a dummy argument can be added to their callback functions.
504 [D. K. Smetters <smetters@parc.xerox.com>]
506 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
507 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
509 *) Add and OPENSSL_LOAD_CONF define which will cause
510 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
511 This allows older applications to transparently support certain
512 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
513 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
514 load the config file and OPENSSL_add_all_algorithms_conf() which will
515 always load it have also been added.
518 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
519 Adjust NIDs and EVP layer.
520 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
522 *) Config modules support in openssl utility.
524 Most commands now load modules from the config file,
525 though in a few (such as version) this isn't done
526 because it couldn't be used for anything.
528 In the case of ca and req the config file used is
529 the same as the utility itself: that is the -config
530 command line option can be used to specify an
534 *) Move default behaviour from OPENSSL_config(). If appname is NULL
535 use "openssl_conf" if filename is NULL use default openssl config file.
538 *) Add an argument to OPENSSL_config() to allow the use of an alternative
539 config section name. Add a new flag to tolerate a missing config file
540 and move code to CONF_modules_load_file().
543 *) Support for crypto accelerator cards from Accelerated Encryption
544 Processing, www.aep.ie. (Use engine 'aep')
545 The support was copied from 0.9.6c [engine] and adapted/corrected
546 to work with the new engine framework.
547 [AEP Inc. and Richard Levitte]
549 *) Support for SureWare crypto accelerator cards from Baltimore
550 Technologies. (Use engine 'sureware')
551 The support was copied from 0.9.6c [engine] and adapted
552 to work with the new engine framework.
555 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
556 make the newer ENGINE framework commands for the CHIL engine work.
557 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
559 *) Make it possible to produce shared libraries on ReliantUNIX.
560 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
562 *) Add the configuration target debug-linux-ppro.
563 Make 'openssl rsa' use the general key loading routines
564 implemented in apps.c, and make those routines able to
565 handle the key format FORMAT_NETSCAPE and the variant
567 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
569 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
570 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
572 *) Add -keyform to rsautl, and document -engine.
573 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
575 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
576 BIO_R_NO_SUCH_FILE error code rather than the generic
577 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
582 ERR_peek_last_error_line
583 ERR_peek_last_error_line_data.
587 ERR_peek_error_line_data,
588 but report on the latest error recorded rather than the first one
589 still in the error queue.
590 [Ben Laurie, Bodo Moeller]
592 *) default_algorithms option in ENGINE config module. This allows things
594 default_algorithms = ALL
595 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
598 *) Prelminary ENGINE config module.
601 *) New experimental application configuration code.
604 *) Change the AES code to follow the same name structure as all other
605 symmetric ciphers, and behave the same way. Move everything to
606 the directory crypto/aes, thereby obsoleting crypto/rijndael.
607 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
609 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
610 [Ben Laurie and Theo de Raadt]
612 *) Add option to output public keys in req command.
613 [Massimiliano Pala madwolf@openca.org]
615 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
616 (up to about 10% better than before for P-192 and P-224).
619 *) New functions/macros
621 SSL_CTX_set_msg_callback(ctx, cb)
622 SSL_CTX_set_msg_callback_arg(ctx, arg)
623 SSL_set_msg_callback(ssl, cb)
624 SSL_set_msg_callback_arg(ssl, arg)
626 to request calling a callback function
628 void cb(int write_p, int version, int content_type,
629 const void *buf, size_t len, SSL *ssl, void *arg)
631 whenever a protocol message has been completely received
632 (write_p == 0) or sent (write_p == 1). Here 'version' is the
633 protocol version according to which the SSL library interprets
634 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
635 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
636 the content type as defined in the SSL 3.0/TLS 1.0 protocol
637 specification (change_cipher_spec(20), alert(21), handshake(22)).
638 'buf' and 'len' point to the actual message, 'ssl' to the
639 SSL object, and 'arg' is the application-defined value set by
640 SSL[_CTX]_set_msg_callback_arg().
642 'openssl s_client' and 'openssl s_server' have new '-msg' options
643 to enable a callback that displays all protocol messages.
646 *) Change the shared library support so shared libraries are built as
647 soon as the corresponding static library is finished, and thereby get
648 openssl and the test programs linked against the shared library.
649 This still only happens when the keyword "shard" has been given to
650 the configuration scripts.
652 NOTE: shared library support is still an experimental thing, and
653 backward binary compatibility is still not guaranteed.
654 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
656 *) Add support for Subject Information Access extension.
657 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
659 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
660 additional bytes when new memory had to be allocated, not just
661 when reusing an existing buffer.
664 *) New command line and configuration option 'utf8' for the req command.
665 This allows field values to be specified as UTF8 strings.
668 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
669 runs for the former and machine-readable output for the latter.
672 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
673 of the e-mail address in the DN (i.e., it will go into a certificate
674 extension only). The new configuration file option 'email_in_dn = no'
676 [Massimiliano Pala madwolf@openca.org]
678 *) Change all functions with names starting with des_ to be starting
679 with DES_ instead. Add wrappers that are compatible with libdes,
680 but are named _ossl_old_des_*. Finally, add macros that map the
681 des_* symbols to the corresponding _ossl_old_des_* if libdes
682 compatibility is desired. If OpenSSL 0.9.6c compatibility is
683 desired, the des_* symbols will be mapped to DES_*, with one
686 Since we provide two compatibility mappings, the user needs to
687 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
688 compatibility is desired. The default (i.e., when that macro
689 isn't defined) is OpenSSL 0.9.6c compatibility.
691 There are also macros that enable and disable the support of old
692 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
693 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
694 are defined, the default will apply: to support the old des routines.
696 In either case, one must include openssl/des.h to get the correct
697 definitions. Do not try to just include openssl/des_old.h, that
700 NOTE: This is a major break of an old API into a new one. Software
701 authors are encouraged to switch to the DES_ style functions. Some
702 time in the future, des_old.h and the libdes compatibility functions
703 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
704 default), and then completely removed.
707 *) Test for certificates which contain unsupported critical extensions.
708 If such a certificate is found during a verify operation it is
709 rejected by default: this behaviour can be overridden by either
710 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
711 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
712 X509_supported_extension() has also been added which returns 1 if a
713 particular extension is supported.
716 *) Modify the behaviour of EVP cipher functions in similar way to digests
717 to retain compatibility with existing code.
720 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
721 compatibility with existing code. In particular the 'ctx' parameter does
722 not have to be to be initialized before the call to EVP_DigestInit() and
723 it is tidied up after a call to EVP_DigestFinal(). New function
724 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
725 EVP_MD_CTX_copy() changed to not require the destination to be
726 initialized valid and new function EVP_MD_CTX_copy_ex() added which
727 requires the destination to be valid.
729 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
730 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
733 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
734 so that complete 'Handshake' protocol structures are kept in memory
735 instead of overwriting 'msg_type' and 'length' with 'body' data.
738 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
739 [Massimo Santin via Richard Levitte]
741 *) Major restructuring to the underlying ENGINE code. This includes
742 reduction of linker bloat, separation of pure "ENGINE" manipulation
743 (initialisation, etc) from functionality dealing with implementations
744 of specific crypto iterfaces. This change also introduces integrated
745 support for symmetric ciphers and digest implementations - so ENGINEs
746 can now accelerate these by providing EVP_CIPHER and EVP_MD
747 implementations of their own. This is detailed in crypto/engine/README
748 as it couldn't be adequately described here. However, there are a few
749 API changes worth noting - some RSA, DSA, DH, and RAND functions that
750 were changed in the original introduction of ENGINE code have now
751 reverted back - the hooking from this code to ENGINE is now a good
752 deal more passive and at run-time, operations deal directly with
753 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
754 dereferencing through an ENGINE pointer any more. Also, the ENGINE
755 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
756 they were not being used by the framework as there is no concept of a
757 BIGNUM_METHOD and they could not be generalised to the new
758 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
759 ENGINE_cpy() has been removed as it cannot be consistently defined in
763 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
766 *) Change mkdef.pl to sort symbols that get the same entry number,
767 and make sure the automatically generated functions ERR_load_*
768 become part of libeay.num as well.
771 *) New function SSL_renegotiate_pending(). This returns true once
772 renegotiation has been requested (either SSL_renegotiate() call
773 or HelloRequest/ClientHello receveived from the peer) and becomes
774 false once a handshake has been completed.
775 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
776 sends a HelloRequest, but does not ensure that a handshake takes
777 place. SSL_renegotiate_pending() is useful for checking if the
778 client has followed the request.)
781 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
782 By default, clients may request session resumption even during
783 renegotiation (if session ID contexts permit); with this option,
784 session resumption is possible only in the first handshake.
786 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
787 more bits available for options that should not be part of
788 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
791 *) Add some demos for certificate and certificate request creation.
794 *) Make maximum certificate chain size accepted from the peer application
795 settable (SSL*_get/set_max_cert_list()), as proposed by
796 "Douglas E. Engert" <deengert@anl.gov>.
799 *) Add support for shared libraries for Unixware-7
800 (Boyd Lynn Gerber <gerberb@zenez.com>).
803 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
804 be done prior to destruction. Use this to unload error strings from
805 ENGINEs that load their own error strings. NB: This adds two new API
806 functions to "get" and "set" this destroy handler in an ENGINE.
809 *) Alter all existing ENGINE implementations (except "openssl" and
810 "openbsd") to dynamically instantiate their own error strings. This
811 makes them more flexible to be built both as statically-linked ENGINEs
812 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
813 Also, add stub code to each that makes building them as self-contained
814 shared-libraries easier (see README.ENGINE).
817 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
818 implementations into applications that are completely implemented in
819 self-contained shared-libraries. The "dynamic" ENGINE exposes control
820 commands that can be used to configure what shared-library to load and
821 to control aspects of the way it is handled. Also, made an update to
822 the README.ENGINE file that brings its information up-to-date and
823 provides some information and instructions on the "dynamic" ENGINE
824 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
827 *) Make it possible to unload ranges of ERR strings with a new
828 "ERR_unload_strings" function.
831 *) Add a copy() function to EVP_MD.
834 *) Make EVP_MD routines take a context pointer instead of just the
835 md_data void pointer.
838 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
839 that the digest can only process a single chunk of data
840 (typically because it is provided by a piece of
841 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
842 is only going to provide a single chunk of data, and hence the
843 framework needn't accumulate the data for oneshot drivers.
846 *) As with "ERR", make it possible to replace the underlying "ex_data"
847 functions. This change also alters the storage and management of global
848 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
849 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
850 index counters. The API functions that use this state have been changed
851 to take a "class_index" rather than pointers to the class's local STACK
852 and counter, and there is now an API function to dynamically create new
853 classes. This centralisation allows us to (a) plug a lot of the
854 thread-safety problems that existed, and (b) makes it possible to clean
855 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
856 such data would previously have always leaked in application code and
857 workarounds were in place to make the memory debugging turn a blind eye
858 to it. Application code that doesn't use this new function will still
859 leak as before, but their memory debugging output will announce it now
860 rather than letting it slide.
862 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
863 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
864 has a return value to indicate success or failure.
867 *) Make it possible to replace the underlying "ERR" functions such that the
868 global state (2 LHASH tables and 2 locks) is only used by the "default"
869 implementation. This change also adds two functions to "get" and "set"
870 the implementation prior to it being automatically set the first time
871 any other ERR function takes place. Ie. an application can call "get",
872 pass the return value to a module it has just loaded, and that module
873 can call its own "set" function using that value. This means the
874 module's "ERR" operations will use (and modify) the error state in the
875 application and not in its own statically linked copy of OpenSSL code.
878 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
879 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
880 the operation, and provides a more encapsulated way for external code
881 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
882 to use these functions rather than manually incrementing the counts.
884 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
887 *) Add EVP test program.
890 *) Add symmetric cipher support to ENGINE. Expect the API to change!
893 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
894 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
895 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
896 These allow a CRL to be built without having to access X509_CRL fields
897 directly. Modify 'ca' application to use new functions.
900 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
901 bug workarounds. Rollback attack detection is a security feature.
902 The problem will only arise on OpenSSL servers when TLSv1 is not
903 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
904 Software authors not wanting to support TLSv1 will have special reasons
905 for their choice and can explicitly enable this option.
906 [Bodo Moeller, Lutz Jaenicke]
908 *) Rationalise EVP so it can be extended: don't include a union of
909 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
910 (similar to those existing for EVP_CIPHER_CTX).
915 EVP_MD_CTX_init(&md); /* new function call */
916 EVP_DigestInit(&md, EVP_sha1());
917 EVP_DigestUpdate(&md, in, len);
918 EVP_DigestFinal(&md, out, NULL);
919 EVP_MD_CTX_cleanup(&md); /* new function call */
923 *) Make DES key schedule conform to the usual scheme, as well as
924 correcting its structure. This means that calls to DES functions
925 now have to pass a pointer to a des_key_schedule instead of a
926 plain des_key_schedule (which was actually always a pointer
931 des_set_key_checked(..., &ks);
932 des_ncbc_encrypt(..., &ks, ...);
934 (Note that a later change renames 'des_...' into 'DES_...'.)
937 *) Initial reduction of linker bloat: the use of some functions, such as
938 PEM causes large amounts of unused functions to be linked in due to
939 poor organisation. For example pem_all.c contains every PEM function
940 which has a knock on effect of linking in large amounts of (unused)
941 ASN1 code. Grouping together similar functions and splitting unrelated
942 functions prevents this.
945 *) Cleanup of EVP macros.
948 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
952 *) Add initial OCSP responder support to ocsp application. The
953 revocation information is handled using the text based index
954 use by the ca application. The responder can either handle
955 requests generated internally, supplied in files (for example
956 via a CGI script) or using an internal minimal server.
959 *) Add configuration choices to get zlib compression for TLS.
962 *) Changes to Kerberos SSL for RFC 2712 compliance:
963 1. Implemented real KerberosWrapper, instead of just using
964 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
965 2. Implemented optional authenticator field of KerberosWrapper.
967 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
968 and authenticator structs; see crypto/krb5/.
970 Generalized Kerberos calls to support multiple Kerberos libraries.
971 [Vern Staats <staatsvr@asc.hpc.mil>,
972 Jeffrey Altman <jaltman@columbia.edu>
975 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
976 already does with RSA. testdsa.h now has 'priv_key/pub_key'
977 values for each of the key sizes rather than having just
978 parameters (and 'speed' generating keys each time).
981 *) Speed up EVP routines.
984 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
985 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
986 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
987 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
989 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
990 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
991 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
994 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
996 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
999 *) Added the OS2-EMX target.
1000 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1002 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1003 to support NCONF routines in extension code. New function CONF_set_nconf()
1004 to allow functions which take an NCONF to also handle the old LHASH
1005 structure: this means that the old CONF compatible routines can be
1006 retained (in particular wrt extensions) without having to duplicate the
1007 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1010 *) Enhance the general user interface with mechanisms for inner control
1011 and with possibilities to have yes/no kind of prompts.
1014 *) Change all calls to low level digest routines in the library and
1015 applications to use EVP. Add missing calls to HMAC_cleanup() and
1016 don't assume HMAC_CTX can be copied using memcpy().
1017 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1019 *) Add the possibility to control engines through control names but with
1020 arbitrary arguments instead of just a string.
1021 Change the key loaders to take a UI_METHOD instead of a callback
1022 function pointer. NOTE: this breaks binary compatibility with earlier
1023 versions of OpenSSL [engine].
1024 Adapt the nCipher code for these new conditions and add a card insertion
1028 *) Enhance the general user interface with mechanisms to better support
1029 dialog box interfaces, application-defined prompts, the possibility
1030 to use defaults (for example default passwords from somewhere else)
1031 and interrupts/cancellations.
1034 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1035 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1038 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1039 tidy up some unnecessarily weird code in 'sk_new()').
1040 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1042 *) Change the key loading routines for ENGINEs to use the same kind
1043 callback (pem_password_cb) as all other routines that need this
1047 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1048 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1049 than this minimum value is recommended.
1052 *) New random seeder for OpenVMS, using the system process statistics
1053 that are easily reachable.
1056 *) Windows apparently can't transparently handle global
1057 variables defined in DLLs. Initialisations such as:
1059 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1061 wont compile. This is used by the any applications that need to
1062 declare their own ASN1 modules. This was fixed by adding the option
1063 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1064 needed for static libraries under Win32.
1067 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1068 setting of purpose and trust fields. New X509_STORE trust and
1069 purpose functions and tidy up setting in other SSL functions.
1072 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1073 structure. These are inherited by X509_STORE_CTX when it is
1074 initialised. This allows various defaults to be set in the
1075 X509_STORE structure (such as flags for CRL checking and custom
1076 purpose or trust settings) for functions which only use X509_STORE_CTX
1077 internally such as S/MIME.
1079 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1080 trust settings if they are not set in X509_STORE. This allows X509_STORE
1081 purposes and trust (in S/MIME for example) to override any set by default.
1083 Add command line options for CRL checking to smime, s_client and s_server
1087 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1088 are set then the CRL is looked up in the X509_STORE structure and
1089 its validity and signature checked, then if the certificate is found
1090 in the CRL the verify fails with a revoked error.
1092 Various new CRL related callbacks added to X509_STORE_CTX structure.
1094 Command line options added to 'verify' application to support this.
1096 This needs some additional work, such as being able to handle multiple
1097 CRLs with different times, extension based lookup (rather than just
1098 by subject name) and ultimately more complete V2 CRL extension
1102 *) Add a general user interface API (crypto/ui/). This is designed
1103 to replace things like des_read_password and friends (backward
1104 compatibility functions using this new API are provided).
1105 The purpose is to remove prompting functions from the DES code
1106 section as well as provide for prompting through dialog boxes in
1107 a window system and the like.
1110 *) Add "ex_data" support to ENGINE so implementations can add state at a
1111 per-structure level rather than having to store it globally.
1114 *) Make it possible for ENGINE structures to be copied when retrieved by
1115 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1116 This causes the "original" ENGINE structure to act like a template,
1117 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1118 operational state can be localised to each ENGINE structure, despite the
1119 fact they all share the same "methods". New ENGINE structures returned in
1120 this case have no functional references and the return value is the single
1121 structural reference. This matches the single structural reference returned
1122 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1126 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1127 needs to match any other type at all we need to manually clear the
1131 *) Changes to the "openssl engine" utility to include;
1132 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1133 about an ENGINE's available control commands.
1134 - executing control commands from command line arguments using the
1135 '-pre' and '-post' switches. '-post' is only used if '-t' is
1136 specified and the ENGINE is successfully initialised. The syntax for
1137 the individual commands are colon-separated, for example;
1138 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1141 *) New dynamic control command support for ENGINEs. ENGINEs can now
1142 declare their own commands (numbers), names (strings), descriptions,
1143 and input types for run-time discovery by calling applications. A
1144 subset of these commands are implicitly classed as "executable"
1145 depending on their input type, and only these can be invoked through
1146 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1147 can be based on user input, config files, etc). The distinction is
1148 that "executable" commands cannot return anything other than a boolean
1149 result and can only support numeric or string input, whereas some
1150 discoverable commands may only be for direct use through
1151 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1152 pointers, or other custom uses. The "executable" commands are to
1153 support parameterisations of ENGINE behaviour that can be
1154 unambiguously defined by ENGINEs and used consistently across any
1155 OpenSSL-based application. Commands have been added to all the
1156 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1157 control over shared-library paths without source code alterations.
1160 *) Changed all ENGINE implementations to dynamically allocate their
1161 ENGINEs rather than declaring them statically. Apart from this being
1162 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1163 this also allows the implementations to compile without using the
1164 internal engine_int.h header.
1167 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1168 'const' value. Any code that should be able to modify a RAND_METHOD
1169 should already have non-const pointers to it (ie. they should only
1170 modify their own ones).
1173 *) Made a variety of little tweaks to the ENGINE code.
1174 - "atalla" and "ubsec" string definitions were moved from header files
1175 to C code. "nuron" string definitions were placed in variables
1176 rather than hard-coded - allowing parameterisation of these values
1177 later on via ctrl() commands.
1178 - Removed unused "#if 0"'d code.
1179 - Fixed engine list iteration code so it uses ENGINE_free() to release
1180 structural references.
1181 - Constified the RAND_METHOD element of ENGINE structures.
1182 - Constified various get/set functions as appropriate and added
1183 missing functions (including a catch-all ENGINE_cpy that duplicates
1184 all ENGINE values onto a new ENGINE except reference counts/state).
1185 - Removed NULL parameter checks in get/set functions. Setting a method
1186 or function to NULL is a way of cancelling out a previously set
1187 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1188 and doesn't justify the extra error symbols and code.
1189 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1190 flags from engine_int.h to engine.h.
1191 - Changed prototypes for ENGINE handler functions (init(), finish(),
1192 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1195 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1196 to the algorithm using long division. The binary algorithm can be
1197 used only if the modulus is odd. On 32-bit systems, it is faster
1198 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1199 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1200 up to 450 bits. In 64-bit environments, the binary algorithm
1201 appears to be advantageous for much longer moduli; here we use it
1202 for moduli up to 2048 bits.
1205 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1206 could not support the combine flag in choice fields.
1209 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1210 extensions from a certificate request to the certificate.
1213 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1214 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1215 file: this allows the display of the certificate about to be
1216 signed to be customised, to allow certain fields to be included
1217 or excluded and extension details. The old system didn't display
1218 multicharacter strings properly, omitted fields not in the policy
1219 and couldn't display additional details such as extensions.
1222 *) Function EC_POINTs_mul for multiple scalar multiplication
1223 of an arbitrary number of elliptic curve points
1224 \sum scalars[i]*points[i],
1225 optionally including the generator defined for the EC_GROUP:
1226 scalar*generator + \sum scalars[i]*points[i].
1228 EC_POINT_mul is a simple wrapper function for the typical case
1229 that the point list has just one item (besides the optional
1233 *) First EC_METHODs for curves over GF(p):
1235 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1236 operations and provides various method functions that can also
1237 operate with faster implementations of modular arithmetic.
1239 EC_GFp_mont_method() reuses most functions that are part of
1240 EC_GFp_simple_method, but uses Montgomery arithmetic.
1242 [Bodo Moeller; point addition and point doubling
1243 implementation directly derived from source code provided by
1244 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1246 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1247 crypto/ec/ec_lib.c):
1249 Curves are EC_GROUP objects (with an optional group generator)
1250 based on EC_METHODs that are built into the library.
1252 Points are EC_POINT objects based on EC_GROUP objects.
1254 Most of the framework would be able to handle curves over arbitrary
1255 finite fields, but as there are no obvious types for fields other
1256 than GF(p), some functions are limited to that for now.
1259 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1260 that the file contains a complete HTTP response.
1263 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1264 change the def and num file printf format specifier from "%-40sXXX"
1265 to "%-39s XXX". The latter will always guarantee a space after the
1266 field while the former will cause them to run together if the field
1267 is 40 of more characters long.
1270 *) Constify the cipher and digest 'method' functions and structures
1271 and modify related functions to take constant EVP_MD and EVP_CIPHER
1275 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1276 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1279 *) Modify EVP_Digest*() routines so they now return values. Although the
1280 internal software routines can never fail additional hardware versions
1284 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1286 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1287 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1290 ERR_R_NESTED_ASN1_ERROR
1292 ERR_R_MISSING_ASN1_EOS
1293 were 4 .. 9, conflicting with
1294 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1296 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1297 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1299 Add new error code 'ERR_R_INTERNAL_ERROR'.
1302 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1306 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1307 sets the subject name for a new request or supersedes the
1308 subject name in a given request. Formats that can be parsed are
1309 'CN=Some Name, OU=myOU, C=IT'
1311 'CN=Some Name/OU=myOU/C=IT'.
1313 Add options '-batch' and '-verbose' to 'openssl req'.
1314 [Massimiliano Pala <madwolf@hackmasters.net>]
1316 *) Introduce the possibility to access global variables through
1317 functions on platform were that's the best way to handle exporting
1318 global variables in shared libraries. To enable this functionality,
1319 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1320 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1321 is normally done by Configure or something similar).
1323 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1324 in the source file (foo.c) like this:
1326 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1327 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1329 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1330 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1332 OPENSSL_DECLARE_GLOBAL(int,foo);
1333 #define foo OPENSSL_GLOBAL_REF(foo)
1334 OPENSSL_DECLARE_GLOBAL(double,bar);
1335 #define bar OPENSSL_GLOBAL_REF(bar)
1337 The #defines are very important, and therefore so is including the
1338 header file everywhere where the defined globals are used.
1340 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1341 of ASN.1 items, but that structure is a bit different.
1343 The largest change is in util/mkdef.pl which has been enhanced with
1344 better and easier to understand logic to choose which symbols should
1345 go into the Windows .def files as well as a number of fixes and code
1346 cleanup (among others, algorithm keywords are now sorted
1347 lexicographically to avoid constant rewrites).
1350 *) In BN_div() keep a copy of the sign of 'num' before writing the
1351 result to 'rm' because if rm==num the value will be overwritten
1352 and produce the wrong result if 'num' is negative: this caused
1353 problems with BN_mod() and BN_nnmod().
1356 *) Function OCSP_request_verify(). This checks the signature on an
1357 OCSP request and verifies the signer certificate. The signer
1358 certificate is just checked for a generic purpose and OCSP request
1362 *) Add OCSP_check_validity() function to check the validity of OCSP
1363 responses. OCSP responses are prepared in real time and may only
1364 be a few seconds old. Simply checking that the current time lies
1365 between thisUpdate and nextUpdate max reject otherwise valid responses
1366 caused by either OCSP responder or client clock inaccuracy. Instead
1367 we allow thisUpdate and nextUpdate to fall within a certain period of
1368 the current time. The age of the response can also optionally be
1369 checked. Two new options -validity_period and -status_age added to
1373 *) If signature or public key algorithm is unrecognized print out its
1374 OID rather that just UNKNOWN.
1377 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1378 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1379 ID to be generated from the issuer certificate alone which can then be
1380 passed to OCSP_id_issuer_cmp().
1383 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1384 ASN1 modules to export functions returning ASN1_ITEM pointers
1385 instead of the ASN1_ITEM structures themselves. This adds several
1386 new macros which allow the underlying ASN1 function/structure to
1387 be accessed transparently. As a result code should not use ASN1_ITEM
1388 references directly (such as &X509_it) but instead use the relevant
1389 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1390 use of the new ASN1 code on platforms where exporting structures
1391 is problematical (for example in shared libraries) but exporting
1392 functions returning pointers to structures is not.
1395 *) Add support for overriding the generation of SSL/TLS session IDs.
1396 These callbacks can be registered either in an SSL_CTX or per SSL.
1397 The purpose of this is to allow applications to control, if they wish,
1398 the arbitrary values chosen for use as session IDs, particularly as it
1399 can be useful for session caching in multiple-server environments. A
1400 command-line switch for testing this (and any client code that wishes
1401 to use such a feature) has been added to "s_server".
1402 [Geoff Thorpe, Lutz Jaenicke]
1404 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1405 of the form '#if defined(...) || defined(...) || ...' and
1406 '#if !defined(...) && !defined(...) && ...'. This also avoids
1407 the growing number of special cases it was previously handling.
1410 *) Make all configuration macros available for application by making
1411 sure they are available in opensslconf.h, by giving them names starting
1412 with "OPENSSL_" to avoid conflicts with other packages and by making
1413 sure e_os2.h will cover all platform-specific cases together with
1415 Additionally, it is now possible to define configuration/platform-
1416 specific names (called "system identities"). In the C code, these
1417 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1418 macro with the name beginning with "OPENSSL_SYS_", which is determined
1419 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1423 *) New option -set_serial to 'req' and 'x509' this allows the serial
1424 number to use to be specified on the command line. Previously self
1425 signed certificates were hard coded with serial number 0 and the
1426 CA options of 'x509' had to use a serial number in a file which was
1430 *) New options to 'ca' utility to support V2 CRL entry extensions.
1431 Currently CRL reason, invalidity date and hold instruction are
1432 supported. Add new CRL extensions to V3 code and some new objects.
1435 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1436 disable standard block padding (aka PKCS#5 padding) in the EVP
1437 API, which was previously mandatory. This means that the data is
1438 not padded in any way and so the total length much be a multiple
1439 of the block size, otherwise an error occurs.
1442 *) Initial (incomplete) OCSP SSL support.
1445 *) New function OCSP_parse_url(). This splits up a URL into its host,
1446 port and path components: primarily to parse OCSP URLs. New -url
1447 option to ocsp utility.
1450 *) New nonce behavior. The return value of OCSP_check_nonce() now
1451 reflects the various checks performed. Applications can decide
1452 whether to tolerate certain situations such as an absent nonce
1453 in a response when one was present in a request: the ocsp application
1454 just prints out a warning. New function OCSP_add1_basic_nonce()
1455 this is to allow responders to include a nonce in a response even if
1456 the request is nonce-less.
1459 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1460 skipped when using openssl x509 multiple times on a single input file,
1461 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1464 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1465 set string type: to handle setting ASN1_TIME structures. Fix ca
1466 utility to correctly initialize revocation date of CRLs.
1469 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1470 the clients preferred ciphersuites and rather use its own preferences.
1471 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1472 Internet Explorer by ensuring unchanged hash method during stepup.
1473 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1476 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1477 to aes and add a new 'exist' option to print out symbols that don't
1481 *) Additional options to ocsp utility to allow flags to be set and
1482 additional certificates supplied.
1485 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1486 OCSP client a number of certificate to only verify the response
1490 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1491 handle the new API. Currently only ECB, CBC modes supported. Add new
1494 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1495 Encryption Standard (AES) Ciphersuites for Transport Layer
1496 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1497 not enabled by default and were not part of the "ALL" ciphersuite
1498 alias because they were not yet official; they could be
1499 explicitly requested by specifying the "AESdraft" ciphersuite
1500 group alias. In the final release of OpenSSL 0.9.7, the group
1501 alias is called "AES" and is part of "ALL".)
1502 [Ben Laurie, Steve Henson, Bodo Moeller]
1504 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1505 request to response.
1508 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1509 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1510 extract information from a certificate request. OCSP_response_create()
1511 creates a response and optionally adds a basic response structure.
1512 OCSP_basic_add1_status() adds a complete single response to a basic
1513 response and returns the OCSP_SINGLERESP structure just added (to allow
1514 extensions to be included for example). OCSP_basic_add1_cert() adds a
1515 certificate to a basic response and OCSP_basic_sign() signs a basic
1516 response with various flags. New helper functions ASN1_TIME_check()
1517 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1518 (converts ASN1_TIME to GeneralizedTime).
1521 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1522 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1523 structure from a certificate. X509_pubkey_digest() digests the public_key
1524 contents: this is used in various key identifiers.
1527 *) Make sk_sort() tolerate a NULL argument.
1528 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1530 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1531 passed by the function are trusted implicitly. If any of them signed the
1532 response then it is assumed to be valid and is not verified.
1535 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1536 to data. This was previously part of the PKCS7 ASN1 code. This
1537 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1538 [Steve Henson, reported by Kenneth R. Robinette
1539 <support@securenetterm.com>]
1541 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1542 routines: without these tracing memory leaks is very painful.
1543 Fix leaks in PKCS12 and PKCS7 routines.
1546 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1547 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1548 effectively meant GeneralizedTime would never be used. Now it
1549 is initialised to -1 but X509_time_adj() now has to check the value
1550 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1551 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1552 [Steve Henson, reported by Kenneth R. Robinette
1553 <support@securenetterm.com>]
1555 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1556 result in a zero length in the ASN1_INTEGER structure which was
1557 not consistent with the structure when d2i_ASN1_INTEGER() was used
1558 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1559 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1560 where it did not print out a minus for negative ASN1_INTEGER.
1563 *) Add summary printout to ocsp utility. The various functions which
1564 convert status values to strings have been renamed to:
1565 OCSP_response_status_str(), OCSP_cert_status_str() and
1566 OCSP_crl_reason_str() and are no longer static. New options
1567 to verify nonce values and to disable verification. OCSP response
1568 printout format cleaned up.
1571 *) Add additional OCSP certificate checks. These are those specified
1572 in RFC2560. This consists of two separate checks: the CA of the
1573 certificate being checked must either be the OCSP signer certificate
1574 or the issuer of the OCSP signer certificate. In the latter case the
1575 OCSP signer certificate must contain the OCSP signing extended key
1576 usage. This check is performed by attempting to match the OCSP
1577 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1578 in the OCSP_CERTID structures of the response.
1581 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1582 and related routines. This uses the standard OpenSSL certificate
1583 verify routines to perform initial checks (just CA validity) and
1584 to obtain the certificate chain. Then additional checks will be
1585 performed on the chain. Currently the root CA is checked to see
1586 if it is explicitly trusted for OCSP signing. This is used to set
1587 a root CA as a global signing root: that is any certificate that
1588 chains to that CA is an acceptable OCSP signing certificate.
1591 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1592 extensions from a separate configuration file.
1593 As when reading extensions from the main configuration file,
1594 the '-extensions ...' option may be used for specifying the
1596 [Massimiliano Pala <madwolf@comune.modena.it>]
1598 *) New OCSP utility. Allows OCSP requests to be generated or
1599 read. The request can be sent to a responder and the output
1600 parsed, outputed or printed in text form. Not complete yet:
1601 still needs to check the OCSP response validity.
1604 *) New subcommands for 'openssl ca':
1605 'openssl ca -status <serial>' prints the status of the cert with
1606 the given serial number (according to the index file).
1607 'openssl ca -updatedb' updates the expiry status of certificates
1609 [Massimiliano Pala <madwolf@comune.modena.it>]
1611 *) New '-newreq-nodes' command option to CA.pl. This is like
1612 '-newreq', but calls 'openssl req' with the '-nodes' option
1613 so that the resulting key is not encrypted.
1614 [Damien Miller <djm@mindrot.org>]
1616 *) New configuration for the GNU Hurd.
1617 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1619 *) Initial code to implement OCSP basic response verify. This
1620 is currently incomplete. Currently just finds the signer's
1621 certificate and verifies the signature on the response.
1624 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1625 value of OPENSSLDIR. This is available via the new '-d' option
1626 to 'openssl version', and is also included in 'openssl version -a'.
1629 *) Allowing defining memory allocation callbacks that will be given
1630 file name and line number information in additional arguments
1631 (a const char* and an int). The basic functionality remains, as
1632 well as the original possibility to just replace malloc(),
1633 realloc() and free() by functions that do not know about these
1634 additional arguments. To register and find out the current
1635 settings for extended allocation functions, the following
1636 functions are provided:
1638 CRYPTO_set_mem_ex_functions
1639 CRYPTO_set_locked_mem_ex_functions
1640 CRYPTO_get_mem_ex_functions
1641 CRYPTO_get_locked_mem_ex_functions
1643 These work the same way as CRYPTO_set_mem_functions and friends.
1644 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1645 extended allocation function is enabled.
1646 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1647 a conventional allocation function is enabled.
1648 [Richard Levitte, Bodo Moeller]
1650 *) Finish off removing the remaining LHASH function pointer casts.
1651 There should no longer be any prototype-casting required when using
1652 the LHASH abstraction, and any casts that remain are "bugs". See
1653 the callback types and macros at the head of lhash.h for details
1654 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1657 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1658 If /dev/[u]random devices are not available or do not return enough
1659 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1661 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1662 /etc/entropy will be queried once each in this sequence, quering stops
1663 when enough entropy was collected without querying more sockets.
1666 *) Change the Unix RAND_poll() variant to be able to poll several
1667 random devices, as specified by DEVRANDOM, until a sufficient amount
1668 of data has been collected. We spend at most 10 ms on each file
1669 (select timeout) and read in non-blocking mode. DEVRANDOM now
1670 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1671 (previously it was just the string "/dev/urandom"), so on typical
1672 platforms the 10 ms delay will never occur.
1673 Also separate out the Unix variant to its own file, rand_unix.c.
1674 For VMS, there's a currently-empty rand_vms.c.
1677 *) Move OCSP client related routines to ocsp_cl.c. These
1678 provide utility functions which an application needing
1679 to issue a request to an OCSP responder and analyse the
1680 response will typically need: as opposed to those which an
1681 OCSP responder itself would need which will be added later.
1683 OCSP_request_sign() signs an OCSP request with an API similar
1684 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1685 response. OCSP_response_get1_basic() extracts basic response
1686 from response. OCSP_resp_find_status(): finds and extracts status
1687 information from an OCSP_CERTID structure (which will be created
1688 when the request structure is built). These are built from lower
1689 level functions which work on OCSP_SINGLERESP structures but
1690 wont normally be used unless the application wishes to examine
1691 extensions in the OCSP response for example.
1693 Replace nonce routines with a pair of functions.
1694 OCSP_request_add1_nonce() adds a nonce value and optionally
1695 generates a random value. OCSP_check_nonce() checks the
1696 validity of the nonce in an OCSP response.
1699 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1700 This doesn't copy the supplied OCSP_CERTID and avoids the
1701 need to free up the newly created id. Change return type
1702 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1703 This can then be used to add extensions to the request.
1704 Deleted OCSP_request_new(), since most of its functionality
1705 is now in OCSP_REQUEST_new() (and the case insensitive name
1706 clash) apart from the ability to set the request name which
1707 will be added elsewhere.
1710 *) Update OCSP API. Remove obsolete extensions argument from
1711 various functions. Extensions are now handled using the new
1712 OCSP extension code. New simple OCSP HTTP function which
1713 can be used to send requests and parse the response.
1716 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1717 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1718 uses the special reorder version of SET OF to sort the attributes
1719 and reorder them to match the encoded order. This resolves a long
1720 standing problem: a verify on a PKCS7 structure just after signing
1721 it used to fail because the attribute order did not match the
1722 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1723 it uses the received order. This is necessary to tolerate some broken
1724 software that does not order SET OF. This is handled by encoding
1725 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1726 to produce the required SET OF.
1729 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1730 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1731 files to get correct declarations of the ASN.1 item variables.
1734 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1735 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1736 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1737 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1738 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1739 ASN1_ITEM and no wrapper functions.
1742 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1743 replace the old function pointer based I/O routines. Change most of
1744 the *_d2i_bio() and *_d2i_fp() functions to use these.
1747 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1748 lines, recognice more "algorithms" that can be deselected, and make
1749 it complain about algorithm deselection that isn't recognised.
1752 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1753 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1754 to use new functions. Add NO_ASN1_OLD which can be set to remove
1755 some old style ASN1 functions: this can be used to determine if old
1756 code will still work when these eventually go away.
1759 *) New extension functions for OCSP structures, these follow the
1760 same conventions as certificates and CRLs.
1763 *) New function X509V3_add1_i2d(). This automatically encodes and
1764 adds an extension. Its behaviour can be customised with various
1765 flags to append, replace or delete. Various wrappers added for
1766 certifcates and CRLs.
1769 *) Fix to avoid calling the underlying ASN1 print routine when
1770 an extension cannot be parsed. Correct a typo in the
1771 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1774 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1775 entries for variables.
1778 *) Add functionality to apps/openssl.c for detecting locking
1779 problems: As the program is single-threaded, all we have
1780 to do is register a locking callback using an array for
1781 storing which locks are currently held by the program.
1784 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1785 SSL_get_ex_data_X509_STORE_idx(), which is used in
1786 ssl_verify_cert_chain() and thus can be called at any time
1787 during TLS/SSL handshakes so that thread-safety is essential.
1788 Unfortunately, the ex_data design is not at all suited
1789 for multi-threaded use, so it probably should be abolished.
1792 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1793 [Broadcom, tweaked and integrated by Geoff Thorpe]
1795 *) Move common extension printing code to new function
1796 X509V3_print_extensions(). Reorganise OCSP print routines and
1797 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1800 *) New function X509_signature_print() to remove duplication in some
1804 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1805 set (this was treated exactly the same as SET OF previously). This
1806 is used to reorder the STACK representing the structure to match the
1807 encoding. This will be used to get round a problem where a PKCS7
1808 structure which was signed could not be verified because the STACK
1809 order did not reflect the encoded order.
1812 *) Reimplement the OCSP ASN1 module using the new code.
1815 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1816 for its ASN1 operations. The old style function pointers still exist
1817 for now but they will eventually go away.
1820 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1821 completely replaces the old ASN1 functionality with a table driven
1822 encoder and decoder which interprets an ASN1_ITEM structure describing
1823 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1824 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1825 has also been converted to the new form.
1828 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1829 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1830 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1831 for negative moduli.
1834 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1835 of not touching the result's sign bit.
1838 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1842 *) Changed the LHASH code to use prototypes for callbacks, and created
1843 macros to declare and implement thin (optionally static) functions
1844 that provide type-safety and avoid function pointer casting for the
1845 type-specific callbacks.
1848 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1850 [Veers Staats <staatsvr@asc.hpc.mil>,
1851 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1853 *) Reformat the FAQ so the different questions and answers can be divided
1854 in sections depending on the subject.
1857 *) Have the zlib compression code load ZLIB.DLL dynamically under
1861 *) New function BN_mod_sqrt for computing square roots modulo a prime
1862 (using the probabilistic Tonelli-Shanks algorithm unless
1863 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1864 be handled deterministically).
1865 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1867 *) Make BN_mod_inverse faster by explicitly handling small quotients
1868 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1869 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1872 *) New function BN_kronecker.
1875 *) Fix BN_gcd so that it works on negative inputs; the result is
1876 positive unless both parameters are zero.
1877 Previously something reasonably close to an infinite loop was
1878 possible because numbers could be growing instead of shrinking
1879 in the implementation of Euclid's algorithm.
1882 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1883 sign of the number in question.
1885 Fix BN_is_word(a,w) to work correctly for w == 0.
1887 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1888 because its test if the absolute value of 'a' equals 'w'.
1889 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1890 it exists mostly for use in the implementations of BN_is_zero(),
1891 BN_is_one(), and BN_is_word().
1894 *) New function BN_swap.
1897 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1898 the exponentiation functions are more likely to produce reasonable
1899 results on negative inputs.
1902 *) Change BN_mod_mul so that the result is always non-negative.
1903 Previously, it could be negative if one of the factors was negative;
1904 I don't think anyone really wanted that behaviour.
1907 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1908 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1909 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1910 and add new functions:
1919 BN_mod_lshift1_quick
1923 These functions always generate non-negative results.
1925 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1926 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1928 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1929 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1930 be reduced modulo m.
1931 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1934 The following entry accidentily appeared in the CHANGES file
1935 distributed with OpenSSL 0.9.7. The modifications described in
1936 it do *not* apply to OpenSSL 0.9.7.
1938 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1939 was actually never needed) and in BN_mul(). The removal in BN_mul()
1940 required a small change in bn_mul_part_recursive() and the addition
1941 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1942 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1943 bn_sub_words() and bn_add_words() except they take arrays with
1948 *) In 'openssl passwd', verify passwords read from the terminal
1949 unless the '-salt' option is used (which usually means that
1950 verification would just waste user's time since the resulting
1951 hash is going to be compared with some given password hash)
1952 or the new '-noverify' option is used.
1954 This is an incompatible change, but it does not affect
1955 non-interactive use of 'openssl passwd' (passwords on the command
1956 line, '-stdin' option, '-in ...' option) and thus should not
1960 *) Remove all references to RSAref, since there's no more need for it.
1963 *) Make DSO load along a path given through an environment variable
1964 (SHLIB_PATH) with shl_load().
1967 *) Constify the ENGINE code as a result of BIGNUM constification.
1968 Also constify the RSA code and most things related to it. In a
1969 few places, most notable in the depth of the ASN.1 code, ugly
1970 casts back to non-const were required (to be solved at a later
1974 *) Make it so the openssl application has all engines loaded by default.
1977 *) Constify the BIGNUM routines a little more.
1980 *) Add the following functions:
1982 ENGINE_load_cswift()
1984 ENGINE_load_atalla()
1986 ENGINE_load_builtin_engines()
1988 That way, an application can itself choose if external engines that
1989 are built-in in OpenSSL shall ever be used or not. The benefit is
1990 that applications won't have to be linked with libdl or other dso
1991 libraries unless it's really needed.
1993 Changed 'openssl engine' to load all engines on demand.
1994 Changed the engine header files to avoid the duplication of some
1995 declarations (they differed!).
1998 *) 'openssl engine' can now list capabilities.
2001 *) Better error reporting in 'openssl engine'.
2004 *) Never call load_dh_param(NULL) in s_server.
2007 *) Add engine application. It can currently list engines by name and
2008 identity, and test if they are actually available.
2011 *) Improve RPM specification file by forcing symbolic linking and making
2012 sure the installed documentation is also owned by root.root.
2013 [Damien Miller <djm@mindrot.org>]
2015 *) Give the OpenSSL applications more possibilities to make use of
2016 keys (public as well as private) handled by engines.
2019 *) Add OCSP code that comes from CertCo.
2022 *) Add VMS support for the Rijndael code.
2025 *) Added untested support for Nuron crypto accelerator.
2028 *) Add support for external cryptographic devices. This code was
2029 previously distributed separately as the "engine" branch.
2030 [Geoff Thorpe, Richard Levitte]
2032 *) Rework the filename-translation in the DSO code. It is now possible to
2033 have far greater control over how a "name" is turned into a filename
2034 depending on the operating environment and any oddities about the
2035 different shared library filenames on each system.
2038 *) Support threads on FreeBSD-elf in Configure.
2041 *) Fix for SHA1 assembly problem with MASM: it produces
2042 warnings about corrupt line number information when assembling
2043 with debugging information. This is caused by the overlapping
2045 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2048 NCONF_get_number() has no error checking at all. As a replacement,
2049 NCONF_get_number_e() is defined (_e for "error checking") and is
2050 promoted strongly. The old NCONF_get_number is kept around for
2051 binary backward compatibility.
2052 Make it possible for methods to load from something other than a BIO,
2053 by providing a function pointer that is given a name instead of a BIO.
2054 For example, this could be used to load configuration data from an
2058 *) Fix for non blocking accept BIOs. Added new I/O special reason
2059 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2060 with non blocking I/O was not possible because no retry code was
2061 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2065 *) Added the beginnings of Rijndael support.
2068 *) Fix for bug in DirectoryString mask setting. Add support for
2069 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2070 to allow certificate printing to more controllable, additional
2071 'certopt' option to 'x509' to allow new printing options to be
2075 *) Clean old EAY MD5 hack from e_os.h.
2078 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2080 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2081 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2082 [Joe Orton, Steve Henson]
2084 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2086 *) Fix additional bug revealed by the NISCC test suite:
2088 Stop bug triggering large recursion when presented with
2089 certain ASN.1 tags (CAN-2003-0851)
2092 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2094 *) Fix various bugs revealed by running the NISCC test suite:
2096 Stop out of bounds reads in the ASN1 code when presented with
2097 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2099 If verify callback ignores invalid public key errors don't try to check
2100 certificate signature with the NULL public key.
2104 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2105 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2109 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2110 extra data after the compression methods not only for TLS 1.0
2111 but also for SSL 3.0 (as required by the specification).
2112 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2114 *) Change X509_certificate_type() to mark the key as exported/exportable
2115 when it's 512 *bits* long, not 512 bytes.
2118 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2120 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2121 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2122 a protocol version number mismatch like a decryption error
2123 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2126 *) Turn on RSA blinding by default in the default implementation
2127 to avoid a timing attack. Applications that don't want it can call
2128 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2129 They would be ill-advised to do so in most cases.
2130 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2132 *) Change RSA blinding code so that it works when the PRNG is not
2133 seeded (in this case, the secret RSA exponent is abused as
2134 an unpredictable seed -- if it is not unpredictable, there
2135 is no point in blinding anyway). Make RSA blinding thread-safe
2136 by remembering the creator's thread ID in rsa->blinding and
2137 having all other threads use local one-time blinding factors
2138 (this requires more computation than sharing rsa->blinding, but
2139 avoids excessive locking; and if an RSA object is not shared
2140 between threads, blinding will still be very fast).
2143 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2145 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2146 via timing by performing a MAC computation even if incorrrect
2147 block cipher padding has been found. This is a countermeasure
2148 against active attacks where the attacker has to distinguish
2149 between bad padding and a MAC verification error. (CAN-2003-0078)
2151 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2152 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2153 Martin Vuagnoux (EPFL, Ilion)]
2155 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2157 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2158 memory from it's contents. This is done with a counter that will
2159 place alternating values in each byte. This can be used to solve
2160 two issues: 1) the removal of calls to memset() by highly optimizing
2161 compilers, and 2) cleansing with other values than 0, since those can
2162 be read through on certain media, for example a swap space on disk.
2165 *) Bugfix: client side session caching did not work with external caching,
2166 because the session->cipher setting was not restored when reloading
2167 from the external cache. This problem was masked, when
2168 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2169 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2172 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2173 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2174 [Zeev Lieber <zeev-l@yahoo.com>]
2176 *) Undo an undocumented change introduced in 0.9.6e which caused
2177 repeated calls to OpenSSL_add_all_ciphers() and
2178 OpenSSL_add_all_digests() to be ignored, even after calling
2182 *) Change the default configuration reader to deal with last line not
2183 being properly terminated.
2186 *) Change X509_NAME_cmp() so it applies the special rules on handling
2187 DN values that are of type PrintableString, as well as RDNs of type
2188 emailAddress where the value has the type ia5String.
2189 [stefank@valicert.com via Richard Levitte]
2191 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2192 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2193 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2194 the bitwise-OR of the two for use by the majority of applications
2195 wanting this behaviour, and update the docs. The documented
2196 behaviour and actual behaviour were inconsistent and had been
2197 changing anyway, so this is more a bug-fix than a behavioural
2199 [Geoff Thorpe, diagnosed by Nadav Har'El]
2201 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2202 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2205 *) Fix initialization code race conditions in
2206 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2207 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2208 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2209 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2210 ssl2_get_cipher_by_char(),
2211 ssl3_get_cipher_by_char().
2212 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2214 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2215 the cached sessions are flushed, as the remove_cb() might use ex_data
2216 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2217 (see [openssl.org #212]).
2218 [Geoff Thorpe, Lutz Jaenicke]
2220 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2221 length, instead of the encoding length to d2i_ASN1_OBJECT.
2224 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2226 *) [In 0.9.6g-engine release:]
2227 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2228 [Lynn Gazis <lgazis@rainbow.com>]
2230 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2232 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2233 and get fix the header length calculation.
2234 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2235 Alon Kantor <alonk@checkpoint.com> (and others),
2238 *) Use proper error handling instead of 'assertions' in buffer
2239 overflow checks added in 0.9.6e. This prevents DoS (the
2240 assertions could call abort()).
2241 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2243 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2245 *) Add various sanity checks to asn1_get_length() to reject
2246 the ASN1 length bytes if they exceed sizeof(long), will appear
2247 negative or the content length exceeds the length of the
2249 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2251 *) Fix cipher selection routines: ciphers without encryption had no flags
2252 for the cipher strength set and where therefore not handled correctly
2253 by the selection routines (PR #130).
2256 *) Fix EVP_dsa_sha macro.
2260 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2261 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2262 that was added in OpenSSL 0.9.6d.
2264 As the countermeasure turned out to be incompatible with some
2265 broken SSL implementations, the new option is part of SSL_OP_ALL.
2266 SSL_OP_ALL is usually employed when compatibility with weird SSL
2267 implementations is desired (e.g. '-bugs' option to 's_client' and
2268 's_server'), so the new option is automatically set in many
2272 *) Changes in security patch:
2274 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2275 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2276 Air Force Materiel Command, USAF, under agreement number
2279 *) Add various sanity checks to asn1_get_length() to reject
2280 the ASN1 length bytes if they exceed sizeof(long), will appear
2281 negative or the content length exceeds the length of the
2282 supplied buffer. (CAN-2002-0659)
2283 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2285 *) Assertions for various potential buffer overflows, not known to
2287 [Ben Laurie (CHATS)]
2289 *) Various temporary buffers to hold ASCII versions of integers were
2290 too small for 64 bit platforms. (CAN-2002-0655)
2291 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2293 *) Remote buffer overflow in SSL3 protocol - an attacker could
2294 supply an oversized session ID to a client. (CAN-2002-0656)
2295 [Ben Laurie (CHATS)]
2297 *) Remote buffer overflow in SSL2 protocol - an attacker could
2298 supply an oversized client master key. (CAN-2002-0656)
2299 [Ben Laurie (CHATS)]
2301 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2303 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2304 encoded as NULL) with id-dsa-with-sha1.
2305 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2307 *) Check various X509_...() return values in apps/req.c.
2308 [Nils Larsch <nla@trustcenter.de>]
2310 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2311 an end-of-file condition would erronously be flagged, when the CRLF
2312 was just at the end of a processed block. The bug was discovered when
2313 processing data through a buffering memory BIO handing the data to a
2314 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2315 <ptsekov@syntrex.com> and Nedelcho Stanev.
2318 *) Implement a countermeasure against a vulnerability recently found
2319 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2320 before application data chunks to avoid the use of known IVs
2321 with data potentially chosen by the attacker.
2324 *) Fix length checks in ssl3_get_client_hello().
2327 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2328 to prevent ssl3_read_internal() from incorrectly assuming that
2329 ssl3_read_bytes() found application data while handshake
2330 processing was enabled when in fact s->s3->in_read_app_data was
2331 merely automatically cleared during the initial handshake.
2332 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2334 *) Fix object definitions for Private and Enterprise: they were not
2335 recognized in their shortname (=lowercase) representation. Extend
2336 obj_dat.pl to issue an error when using undefined keywords instead
2337 of silently ignoring the problem (Svenning Sorensen
2338 <sss@sss.dnsalias.net>).
2341 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2342 generators, i.e. generators other than 2 and 5. (Previously, the
2343 code did not properly initialise the 'add' and 'rem' values to
2344 BN_generate_prime().)
2346 In the new general case, we do not insist that 'generator' is
2347 actually a primitive root: This requirement is rather pointless;
2348 a generator of the order-q subgroup is just as good, if not
2352 *) Map new X509 verification errors to alerts. Discovered and submitted by
2353 Tom Wu <tom@arcot.com>.
2356 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2357 returning non-zero before the data has been completely received
2358 when using non-blocking I/O.
2359 [Bodo Moeller; problem pointed out by John Hughes]
2361 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2362 [Ben Laurie, Lutz Jaenicke]
2364 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2365 Yoram Zahavi <YoramZ@gilian.com>).
2368 *) Add information about CygWin 1.3 and on, and preserve proper
2369 configuration for the versions before that.
2370 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2372 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2373 check whether we deal with a copy of a session and do not delete from
2374 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2375 <izhar@checkpoint.com>.
2378 *) Do not store session data into the internal session cache, if it
2379 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2380 flag is set). Proposed by Aslam <aslam@funk.com>.
2383 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2387 *) [In 0.9.6d-engine release:]
2388 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2389 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2391 *) Add the configuration target linux-s390x.
2392 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2394 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2395 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2396 variable as an indication that a ClientHello message has been
2397 received. As the flag value will be lost between multiple
2398 invocations of ssl3_accept when using non-blocking I/O, the
2399 function may not be aware that a handshake has actually taken
2400 place, thus preventing a new session from being added to the
2403 To avoid this problem, we now set s->new_session to 2 instead of
2404 using a local variable.
2405 [Lutz Jaenicke, Bodo Moeller]
2407 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2408 if the SSL_R_LENGTH_MISMATCH error is detected.
2409 [Geoff Thorpe, Bodo Moeller]
2411 *) New 'shared_ldflag' column in Configure platform table.
2414 *) Fix EVP_CIPHER_mode macro.
2415 ["Dan S. Camper" <dan@bti.net>]
2417 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2418 type, we must throw them away by setting rr->length to 0.
2419 [D P Chang <dpc@qualys.com>]
2421 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2423 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2424 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2425 worked incorrectly for those cases where range = 10..._2 and
2426 3*range is two bits longer than range.)
2429 *) Only add signing time to PKCS7 structures if it is not already
2433 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2434 OBJ_ld_ce should be OBJ_id_ce.
2435 Also some ip-pda OIDs in crypto/objects/objects.txt were
2436 incorrect (cf. RFC 3039).
2437 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2439 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2440 returns early because it has nothing to do.
2441 [Andy Schneider <andy.schneider@bjss.co.uk>]
2443 *) [In 0.9.6c-engine release:]
2444 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2445 [Andy Schneider <andy.schneider@bjss.co.uk>]
2447 *) [In 0.9.6c-engine release:]
2448 Add support for Cryptographic Appliance's keyserver technology.
2449 (Use engine 'keyclient')
2450 [Cryptographic Appliances and Geoff Thorpe]
2452 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2453 is called via tools/c89.sh because arguments have to be
2454 rearranged (all '-L' options must appear before the first object
2456 [Richard Shapiro <rshapiro@abinitio.com>]
2458 *) [In 0.9.6c-engine release:]
2459 Add support for Broadcom crypto accelerator cards, backported
2461 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2463 *) [In 0.9.6c-engine release:]
2464 Add support for SureWare crypto accelerator cards from
2465 Baltimore Technologies. (Use engine 'sureware')
2466 [Baltimore Technologies and Mark Cox]
2468 *) [In 0.9.6c-engine release:]
2469 Add support for crypto accelerator cards from Accelerated
2470 Encryption Processing, www.aep.ie. (Use engine 'aep')
2471 [AEP Inc. and Mark Cox]
2473 *) Add a configuration entry for gcc on UnixWare.
2474 [Gary Benson <gbenson@redhat.com>]
2476 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2477 messages are stored in a single piece (fixed-length part and
2478 variable-length part combined) and fix various bugs found on the way.
2481 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2482 instead. BIO_gethostbyname() does not know what timeouts are
2483 appropriate, so entries would stay in cache even when they have
2485 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2487 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2488 faced with a pathologically small ClientHello fragment that does
2489 not contain client_version: Instead of aborting with an error,
2490 simply choose the highest available protocol version (i.e.,
2491 TLS 1.0 unless it is disabled). In practice, ClientHello
2492 messages are never sent like this, but this change gives us
2493 strictly correct behaviour at least for TLS.
2496 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2497 never resets s->method to s->ctx->method when called from within
2498 one of the SSL handshake functions.
2499 [Bodo Moeller; problem pointed out by Niko Baric]
2501 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2502 (sent using the client's version number) if client_version is
2503 smaller than the protocol version in use. Also change
2504 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2505 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2506 the client will at least see that alert.
2509 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2513 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2514 client receives HelloRequest while in a handshake.
2515 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2517 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2518 should end in 'break', not 'goto end' which circuments various
2519 cleanups done in state SSL_ST_OK. But session related stuff
2520 must be disabled for SSL_ST_OK in the case that we just sent a
2523 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2524 before just sending a HelloRequest.
2525 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2527 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2528 reveal whether illegal block cipher padding was found or a MAC
2529 verification error occured. (Neither SSLerr() codes nor alerts
2530 are directly visible to potential attackers, but the information
2531 may leak via logfiles.)
2533 Similar changes are not required for the SSL 2.0 implementation
2534 because the number of padding bytes is sent in clear for SSL 2.0,
2535 and the extra bytes are just ignored. However ssl/s2_pkt.c
2536 failed to verify that the purported number of padding bytes is in
2540 *) Add OpenUNIX-8 support including shared libraries
2541 (Boyd Lynn Gerber <gerberb@zenez.com>).
2544 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2545 'wristwatch attack' using huge encoding parameters (cf.
2546 James H. Manger's CRYPTO 2001 paper). Note that the
2547 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2548 encoding parameters and hence was not vulnerable.
2551 *) BN_sqr() bug fix.
2552 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2554 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2555 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2556 followed by modular reduction.
2557 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2559 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2560 equivalent based on BN_pseudo_rand() instead of BN_rand().
2563 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2564 This function was broken, as the check for a new client hello message
2565 to handle SGC did not allow these large messages.
2566 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2569 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2572 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2573 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2576 *) Rework the configuration and shared library support for Tru64 Unix.
2577 The configuration part makes use of modern compiler features and
2578 still retains old compiler behavior for those that run older versions
2579 of the OS. The shared library support part includes a variant that
2580 uses the RPATH feature, and is available through the special
2581 configuration target "alpha-cc-rpath", which will never be selected
2583 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2585 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2586 with the same message size as in ssl3_get_certificate_request().
2587 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2588 messages might inadvertently be reject as too long.
2589 [Petr Lampa <lampa@fee.vutbr.cz>]
2591 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2594 *) Modified SSL library such that the verify_callback that has been set
2595 specificly for an SSL object with SSL_set_verify() is actually being
2596 used. Before the change, a verify_callback set with this function was
2597 ignored and the verify_callback() set in the SSL_CTX at the time of
2598 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2599 to allow the necessary settings.
2602 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2603 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2604 done automatically (in contradiction to the requirements of the C
2605 standard). This made problems when used from OpenSSH.
2608 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2609 dh->length and always used
2611 BN_rand_range(priv_key, dh->p).
2613 BN_rand_range() is not necessary for Diffie-Hellman, and this
2614 specific range makes Diffie-Hellman unnecessarily inefficient if
2615 dh->length (recommended exponent length) is much smaller than the
2616 length of dh->p. We could use BN_rand_range() if the order of
2617 the subgroup was stored in the DH structure, but we only have
2622 BN_rand(priv_key, l, ...)
2624 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2630 RSA_eay_public_encrypt
2631 RSA_eay_private_decrypt
2632 RSA_eay_private_encrypt (signing)
2633 RSA_eay_public_decrypt (signature verification)
2635 (default implementations for RSA_public_encrypt,
2636 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2637 always reject numbers >= n.
2640 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2641 to synchronize access to 'locking_thread'. This is necessary on
2642 systems where access to 'locking_thread' (an 'unsigned long'
2643 variable) is not atomic.
2646 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2647 *before* setting the 'crypto_lock_rand' flag. The previous code had
2648 a race condition if 0 is a valid thread ID.
2649 [Travis Vitek <vitek@roguewave.com>]
2651 *) Add support for shared libraries under Irix.
2652 [Albert Chin-A-Young <china@thewrittenword.com>]
2654 *) Add configuration option to build on Linux on both big-endian and
2656 [Ralf Baechle <ralf@uni-koblenz.de>]
2658 *) Add the possibility to create shared libraries on HP-UX.
2661 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2663 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2664 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2665 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2666 PRNG state recovery was possible based on the output of
2667 one PRNG request appropriately sized to gain knowledge on
2668 'md' followed by enough consecutive 1-byte PRNG requests
2669 to traverse all of 'state'.
2671 1. When updating 'md_local' (the current thread's copy of 'md')
2672 during PRNG output generation, hash all of the previous
2673 'md_local' value, not just the half used for PRNG output.
2675 2. Make the number of bytes from 'state' included into the hash
2676 independent from the number of PRNG bytes requested.
2678 The first measure alone would be sufficient to avoid
2679 Markku-Juhani's attack. (Actually it had never occurred
2680 to me that the half of 'md_local' used for chaining was the
2681 half from which PRNG output bytes were taken -- I had always
2682 assumed that the secret half would be used.) The second
2683 measure makes sure that additional data from 'state' is never
2684 mixed into 'md_local' in small portions; this heuristically
2685 further strengthens the PRNG.
2688 *) Fix crypto/bn/asm/mips3.s.
2691 *) When only the key is given to "enc", the IV is undefined. Print out
2692 an error message in this case.
2695 *) Handle special case when X509_NAME is empty in X509 printing routines.
2698 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2699 positive and less than q.
2702 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2703 used: it isn't thread safe and the add_lock_callback should handle
2705 [Paul Rose <Paul.Rose@bridge.com>]
2707 *) Verify that incoming data obeys the block size in
2708 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2712 [Ulf Möller, Bodo Möller]
2714 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2715 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2716 when fixing the server behaviour for backwards-compatible 'client
2717 hello' messages. (Note that the attack is impractical against
2718 SSL 3.0 and TLS 1.0 anyway because length and version checking
2719 means that the probability of guessing a valid ciphertext is
2720 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2723 Before 0.9.5, the countermeasure (hide the error by generating a
2724 random 'decryption result') did not work properly because
2725 ERR_clear_error() was missing, meaning that SSL_get_error() would
2726 detect the supposedly ignored error.
2728 Both problems are now fixed.
2731 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2732 (previously it was 1024).
2735 *) Fix for compatibility mode trust settings: ignore trust settings
2736 unless some valid trust or reject settings are present.
2739 *) Fix for blowfish EVP: its a variable length cipher.
2742 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2743 parameters in DSA public key structures and return an error in the
2744 DSA routines if parameters are absent.
2747 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2748 in the current directory if neither $RANDFILE nor $HOME was set.
2749 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2750 caused some confusion to Windows users who haven't defined $HOME.
2751 Thus RAND_file_name() is changed again: e_os.h can define a
2752 DEFAULT_HOME, which will be used if $HOME is not set.
2753 For Windows, we use "C:"; on other platforms, we still require
2754 environment variables.
2756 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2757 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2758 having multiple threads call RAND_poll() concurrently.
2761 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2762 combination of a flag and a thread ID variable.
2763 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2764 flag), *other* threads can enter ssleay_add_bytes without obeying
2765 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2766 that they do not hold after the first thread unsets add_do_not_lock).
2769 *) Change bctest again: '-x' expressions are not available in all
2773 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2775 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2776 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2778 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2779 the default extension for executables, if any. Also, make the perl
2780 scripts that use symlink() to test if it really exists and use "cp"
2781 if it doesn't. All this made OpenSSL compilable and installable in
2785 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2786 If SEQUENCE is length is indefinite just set c->slen to the total
2787 amount of data available.
2788 [Steve Henson, reported by shige@FreeBSD.org]
2789 [This change does not apply to 0.9.7.]
2791 *) Change bctest to avoid here-documents inside command substitution
2792 (workaround for FreeBSD /bin/sh bug).
2793 For compatibility with Ultrix, avoid shell functions (introduced
2794 in the bctest version that searches along $PATH).
2797 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2798 with des_encrypt() defined on some operating systems, like Solaris
2802 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2803 On the Importance of Eliminating Errors in Cryptographic
2804 Computations, J. Cryptology 14 (2001) 2, 101-119,
2805 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2808 *) MIPS assembler BIGNUM division bug fix.
2811 *) Disabled incorrect Alpha assembler code.
2814 *) Fix PKCS#7 decode routines so they correctly update the length
2815 after reading an EOC for the EXPLICIT tag.
2817 [This change does not apply to 0.9.7.]
2819 *) Fix bug in PKCS#12 key generation routines. This was triggered
2820 if a 3DES key was generated with a 0 initial byte. Include
2821 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2822 (but broken) behaviour.
2825 *) Enhance bctest to search for a working bc along $PATH and print
2827 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2829 *) Fix memory leaks in err.c: free err_data string if necessary;
2830 don't write to the wrong index in ERR_set_error_data.
2833 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2837 *) Replace rdtsc with _emit statements for VC++ version 5.
2838 [Jeremy Cooper <jeremy@baymoo.org>]
2840 *) Make it possible to reuse SSLv2 sessions.
2843 *) In copy_email() check for >= 0 as a return value for
2844 X509_NAME_get_index_by_NID() since 0 is a valid index.
2845 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2847 *) Avoid coredump with unsupported or invalid public keys by checking if
2848 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2849 PKCS7_verify() fails with non detached data.
2852 *) Don't use getenv in library functions when run as setuid/setgid.
2853 New function OPENSSL_issetugid().
2856 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2857 due to incorrect handling of multi-threading:
2859 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2861 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2863 3. Count how many times MemCheck_off() has been called so that
2864 nested use can be treated correctly. This also avoids
2865 inband-signalling in the previous code (which relied on the
2866 assumption that thread ID 0 is impossible).
2869 *) Add "-rand" option also to s_client and s_server.
2872 *) Fix CPU detection on Irix 6.x.
2873 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2874 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2876 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2879 [This change does not apply to 0.9.7.]
2881 *) Use the cached encoding of an X509_NAME structure rather than
2882 copying it. This is apparently the reason for the libsafe "errors"
2883 but the code is actually correct.
2886 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2887 Bleichenbacher's DSA attack.
2888 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2889 to be set and top=0 forces the highest bit to be set; top=-1 is new
2890 and leaves the highest bit random.
2891 [Ulf Moeller, Bodo Moeller]
2893 *) In the NCONF_...-based implementations for CONF_... queries
2894 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2895 a temporary CONF structure with the data component set to NULL
2896 (which gives segmentation faults in lh_retrieve).
2897 Instead, use NULL for the CONF pointer in CONF_get_string and
2898 CONF_get_number (which may use environment variables) and directly
2899 return NULL from CONF_get_section.
2902 *) Fix potential buffer overrun for EBCDIC.
2905 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2906 keyUsage if basicConstraints absent for a CA.
2909 *) Make SMIME_write_PKCS7() write mail header values with a format that
2910 is more generally accepted (no spaces before the semicolon), since
2911 some programs can't parse those values properly otherwise. Also make
2912 sure BIO's that break lines after each write do not create invalid
2916 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2917 macros previously used would not encode an empty SEQUENCE OF
2918 and break the signature.
2920 [This change does not apply to 0.9.7.]
2922 *) Zero the premaster secret after deriving the master secret in
2926 *) Add some EVP_add_digest_alias registrations (as found in
2927 OpenSSL_add_all_digests()) to SSL_library_init()
2928 aka OpenSSL_add_ssl_algorithms(). This provides improved
2929 compatibility with peers using X.509 certificates
2930 with unconventional AlgorithmIdentifier OIDs.
2933 *) Fix for Irix with NO_ASM.
2934 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2936 *) ./config script fixes.
2937 [Ulf Moeller, Richard Levitte]
2939 *) Fix 'openssl passwd -1'.
2942 *) Change PKCS12_key_gen_asc() so it can cope with non null
2943 terminated strings whose length is passed in the passlen
2944 parameter, for example from PEM callbacks. This was done
2945 by adding an extra length parameter to asc2uni().
2946 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2948 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2949 call failed, free the DSA structure.
2952 *) Fix to uni2asc() to cope with zero length Unicode strings.
2953 These are present in some PKCS#12 files.
2956 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2957 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2958 when writing a 32767 byte record.
2959 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2961 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2962 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2964 (RSA objects have a reference count access to which is protected
2965 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2966 so they are meant to be shared between threads.)
2967 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2968 "Reddie, Steven" <Steven.Reddie@ca.com>]
2970 *) Fix a deadlock in CRYPTO_mem_leaks().
2973 *) Use better test patterns in bntest.
2976 *) rand_win.c fix for Borland C.
2979 *) BN_rshift bugfix for n == 0.
2982 *) Add a 'bctest' script that checks for some known 'bc' bugs
2983 so that 'make test' does not abort just because 'bc' is broken.
2986 *) Store verify_result within SSL_SESSION also for client side to
2987 avoid potential security hole. (Re-used sessions on the client side
2988 always resulted in verify_result==X509_V_OK, not using the original
2989 result of the server certificate verification.)
2992 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2993 SSL3_RT_APPLICATION_DATA, return 0.
2994 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2998 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2999 releases, have been re-implemented by renaming the previous
3000 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3001 and ssl3_read_internal, respectively, and adding 'peek' parameters
3002 to them. The new ssl[23]_{read,peek} functions are calls to
3003 ssl[23]_read_internal with the 'peek' flag set appropriately.
3004 A 'peek' parameter has also been added to ssl3_read_bytes, which
3005 does the actual work for ssl3_read_internal.
3008 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3009 the method-specific "init()" handler. Also clean up ex_data after
3010 calling the method-specific "finish()" handler. Previously, this was
3011 happening the other way round.
3014 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3015 The previous value, 12, was not always sufficient for BN_mod_exp().
3018 *) Make sure that shared libraries get the internal name engine with
3019 the full version number and not just 0. This should mark the
3020 shared libraries as not backward compatible. Of course, this should
3021 be changed again when we can guarantee backward binary compatibility.
3024 *) Fix typo in get_cert_by_subject() in by_dir.c
3025 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3027 *) Rework the system to generate shared libraries:
3029 - Make note of the expected extension for the shared libraries and
3030 if there is a need for symbolic links from for example libcrypto.so.0
3031 to libcrypto.so.0.9.7. There is extended info in Configure for
3034 - Make as few rebuilds of the shared libraries as possible.
3036 - Still avoid linking the OpenSSL programs with the shared libraries.
3038 - When installing, install the shared libraries separately from the
3042 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3044 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3045 and not in SSL_clear because the latter is also used by the
3046 accept/connect functions; previously, the settings made by
3047 SSL_set_read_ahead would be lost during the handshake.
3048 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3050 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3051 Previously, it would create entries for disableed algorithms no
3055 *) Added several new manual pages for SSL_* function.
3058 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3060 *) In ssl23_get_client_hello, generate an error message when faced
3061 with an initial SSL 3.0/TLS record that is too small to contain the
3062 first two bytes of the ClientHello message, i.e. client_version.
3063 (Note that this is a pathologic case that probably has never happened
3064 in real life.) The previous approach was to use the version number
3065 from the record header as a substitute; but our protocol choice
3066 should not depend on that one because it is not authenticated
3067 by the Finished messages.
3070 *) More robust randomness gathering functions for Windows.
3071 [Jeffrey Altman <jaltman@columbia.edu>]
3073 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3074 not set then we don't setup the error code for issuer check errors
3075 to avoid possibly overwriting other errors which the callback does
3076 handle. If an application does set the flag then we assume it knows
3077 what it is doing and can handle the new informational codes
3081 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3082 a general "ANY" type, as such it should be able to decode anything
3083 including tagged types. However it didn't check the class so it would
3084 wrongly interpret tagged types in the same way as their universal
3085 counterpart and unknown types were just rejected. Changed so that the
3086 tagged and unknown types are handled in the same way as a SEQUENCE:
3087 that is the encoding is stored intact. There is also a new type
3088 "V_ASN1_OTHER" which is used when the class is not universal, in this
3089 case we have no idea what the actual type is so we just lump them all
3093 *) On VMS, stdout may very well lead to a file that is written to
3094 in a record-oriented fashion. That means that every write() will
3095 write a separate record, which will be read separately by the
3096 programs trying to read from it. This can be very confusing.
3098 The solution is to put a BIO filter in the way that will buffer
3099 text until a linefeed is reached, and then write everything a
3100 line at a time, so every record written will be an actual line,
3101 not chunks of lines and not (usually doesn't happen, but I've
3102 seen it once) several lines in one record. BIO_f_linebuffer() is
3105 Currently, it's a VMS-only method, because that's where it has
3106 been tested well enough.
3109 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3110 it can return incorrect results.
3111 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3112 but it was in 0.9.6-beta[12].)
3115 *) Disable the check for content being present when verifying detached
3116 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3117 include zero length content when signing messages.
3120 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3121 BIO_ctrl (for BIO pairs).
3124 *) Add DSO method for VMS.
3127 *) Bug fix: Montgomery multiplication could produce results with the
3131 *) Add RPM specification openssl.spec and modify it to build three
3132 packages. The default package contains applications, application
3133 documentation and run-time libraries. The devel package contains
3134 include files, static libraries and function documentation. The
3135 doc package contains the contents of the doc directory. The original
3136 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3139 *) Add a large number of documentation files for many SSL routines.
3140 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3142 *) Add a configuration entry for Sony News 4.
3143 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3145 *) Don't set the two most significant bits to one when generating a
3146 random number < q in the DSA library.
3149 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3150 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3151 the underlying transport is blocking) if a handshake took place.
3152 (The default behaviour is needed by applications such as s_client
3153 and s_server that use select() to determine when to use SSL_read;
3154 but for applications that know in advance when to expect data, it
3155 just makes things more complicated.)
3158 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3162 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3163 work better on such systems.
3164 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3166 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3167 Update PKCS12_parse() so it copies the friendlyName and the
3168 keyid to the certificates aux info.
3171 *) Fix bug in PKCS7_verify() which caused an infinite loop
3172 if there was more than one signature.
3173 [Sven Uszpelkat <su@celocom.de>]
3175 *) Major change in util/mkdef.pl to include extra information
3176 about each symbol, as well as presentig variables as well
3177 as functions. This change means that there's n more need
3178 to rebuild the .num files when some algorithms are excluded.
3181 *) Allow the verify time to be set by an application,
3182 rather than always using the current time.
3185 *) Phase 2 verify code reorganisation. The certificate
3186 verify code now looks up an issuer certificate by a
3187 number of criteria: subject name, authority key id
3188 and key usage. It also verifies self signed certificates
3189 by the same criteria. The main comparison function is
3190 X509_check_issued() which performs these checks.
3192 Lot of changes were necessary in order to support this
3193 without completely rewriting the lookup code.
3195 Authority and subject key identifier are now cached.
3197 The LHASH 'certs' is X509_STORE has now been replaced
3198 by a STACK_OF(X509_OBJECT). This is mainly because an
3199 LHASH can't store or retrieve multiple objects with
3200 the same hash value.
3202 As a result various functions (which were all internal
3203 use only) have changed to handle the new X509_STORE
3204 structure. This will break anything that messed round
3205 with X509_STORE internally.
3207 The functions X509_STORE_add_cert() now checks for an
3208 exact match, rather than just subject name.
3210 The X509_STORE API doesn't directly support the retrieval
3211 of multiple certificates matching a given criteria, however
3212 this can be worked round by performing a lookup first
3213 (which will fill the cache with candidate certificates)
3214 and then examining the cache for matches. This is probably
3215 the best we can do without throwing out X509_LOOKUP
3216 entirely (maybe later...).
3218 The X509_VERIFY_CTX structure has been enhanced considerably.
3220 All certificate lookup operations now go via a get_issuer()
3221 callback. Although this currently uses an X509_STORE it
3222 can be replaced by custom lookups. This is a simple way
3223 to bypass the X509_STORE hackery necessary to make this
3224 work and makes it possible to use more efficient techniques
3225 in future. A very simple version which uses a simple
3226 STACK for its trusted certificate store is also provided
3227 using X509_STORE_CTX_trusted_stack().
3229 The verify_cb() and verify() callbacks now have equivalents
3230 in the X509_STORE_CTX structure.
3232 X509_STORE_CTX also has a 'flags' field which can be used
3233 to customise the verify behaviour.
3236 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3237 excludes S/MIME capabilities.
3240 *) When a certificate request is read in keep a copy of the
3241 original encoding of the signed data and use it when outputing
3242 again. Signatures then use the original encoding rather than
3243 a decoded, encoded version which may cause problems if the
3244 request is improperly encoded.
3247 *) For consistency with other BIO_puts implementations, call
3248 buffer_write(b, ...) directly in buffer_puts instead of calling
3251 In BIO_puts, increment b->num_write as in BIO_write.
3252 [Peter.Sylvester@EdelWeb.fr]
3254 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3255 BN_zero, we may not return a BIGNUM with an array consisting of
3259 *) Avoid calling abort() from within the library when problems are
3260 detected, except if preprocessor symbols have been defined
3261 (such as REF_CHECK, BN_DEBUG etc.).
3264 *) New openssl application 'rsautl'. This utility can be
3265 used for low level RSA operations. DER public key
3266 BIO/fp routines also added.
3269 *) New Configure entry and patches for compiling on QNX 4.
3270 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3272 *) A demo state-machine implementation was sponsored by
3273 Nuron (http://www.nuron.com/) and is now available in
3274 demos/state_machine.
3277 *) New options added to the 'dgst' utility for signature
3278 generation and verification.
3281 *) Unrecognized PKCS#7 content types are now handled via a
3282 catch all ASN1_TYPE structure. This allows unsupported
3283 types to be stored as a "blob" and an application can
3284 encode and decode it manually.
3287 *) Fix various signed/unsigned issues to make a_strex.c
3289 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3291 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3292 length if passed a buffer. ASN1_INTEGER_to_BN failed
3293 if passed a NULL BN and its argument was negative.
3294 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3296 *) Modification to PKCS#7 encoding routines to output definite
3297 length encoding. Since currently the whole structures are in
3298 memory there's not real point in using indefinite length
3299 constructed encoding. However if OpenSSL is compiled with
3300 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3303 *) Added BIO_vprintf() and BIO_vsnprintf().
3306 *) Added more prefixes to parse for in the the strings written
3307 through a logging bio, to cover all the levels that are available
3308 through syslog. The prefixes are now:
3310 PANIC, EMERG, EMR => LOG_EMERG
3311 ALERT, ALR => LOG_ALERT
3312 CRIT, CRI => LOG_CRIT
3313 ERROR, ERR => LOG_ERR
3314 WARNING, WARN, WAR => LOG_WARNING
3315 NOTICE, NOTE, NOT => LOG_NOTICE
3316 INFO, INF => LOG_INFO
3317 DEBUG, DBG => LOG_DEBUG
3319 and as before, if none of those prefixes are present at the
3320 beginning of the string, LOG_ERR is chosen.
3322 On Win32, the LOG_* levels are mapped according to this:
3324 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3325 LOG_WARNING => EVENTLOG_WARNING_TYPE
3326 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3330 *) Made it possible to reconfigure with just the configuration
3331 argument "reconf" or "reconfigure". The command line arguments
3332 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3333 and are retrieved from there when reconfiguring.
3337 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3339 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3342 *) The obj_dat.pl script was messing up the sorting of object
3343 names. The reason was that it compared the quoted version
3344 of strings as a result "OCSP" > "OCSP Signing" because
3345 " > SPACE. Changed script to store unquoted versions of
3346 names and add quotes on output. It was also omitting some
3347 names from the lookup table if they were given a default
3348 value (that is if SN is missing it is given the same
3349 value as LN and vice versa), these are now added on the
3350 grounds that if an object has a name we should be able to
3351 look it up. Finally added warning output when duplicate
3352 short or long names are found.
3355 *) Changes needed for Tandem NSK.
3356 [Scott Uroff <scott@xypro.com>]
3358 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3359 RSA_padding_check_SSLv23(), special padding was never detected
3360 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3361 version rollback attacks was not effective.
3363 In s23_clnt.c, don't use special rollback-attack detection padding
3364 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3365 client; similarly, in s23_srvr.c, don't do the rollback check if
3366 SSL 2.0 is the only protocol enabled in the server.
3369 *) Make it possible to get hexdumps of unprintable data with 'openssl
3370 asn1parse'. By implication, the functions ASN1_parse_dump() and
3371 BIO_dump_indent() are added.
3374 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3375 these print out strings and name structures based on various
3376 flags including RFC2253 support and proper handling of
3377 multibyte characters. Added options to the 'x509' utility
3378 to allow the various flags to be set.
3381 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3382 Also change the functions X509_cmp_current_time() and
3383 X509_gmtime_adj() work with an ASN1_TIME structure,
3384 this will enable certificates using GeneralizedTime in validity
3385 dates to be checked.
3388 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3389 negative public key encodings) on by default,
3390 NO_NEG_PUBKEY_BUG can be set to disable it.
3393 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3394 content octets. An i2c_ASN1_OBJECT is unnecessary because
3395 the encoding can be trivially obtained from the structure.
3398 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3399 not read locks (CRYPTO_r_[un]lock).
3402 *) A first attempt at creating official support for shared
3403 libraries through configuration. I've kept it so the
3404 default is static libraries only, and the OpenSSL programs
3405 are always statically linked for now, but there are
3406 preparations for dynamic linking in place.
3407 This has been tested on Linux and Tru64.
3410 *) Randomness polling function for Win9x, as described in:
3411 Peter Gutmann, Software Generation of Practically Strong
3415 *) Fix so PRNG is seeded in req if using an already existing
3419 *) New options to smime application. -inform and -outform
3420 allow alternative formats for the S/MIME message including
3421 PEM and DER. The -content option allows the content to be
3422 specified separately. This should allow things like Netscape
3423 form signing output easier to verify.
3426 *) Fix the ASN1 encoding of tags using the 'long form'.
3429 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3430 STRING types. These convert content octets to and from the
3431 underlying type. The actual tag and length octets are
3432 already assumed to have been read in and checked. These
3433 are needed because all other string types have virtually
3434 identical handling apart from the tag. By having versions
3435 of the ASN1 functions that just operate on content octets
3436 IMPLICIT tagging can be handled properly. It also allows
3437 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3438 and ASN1_INTEGER are identical apart from the tag.
3441 *) Change the handling of OID objects as follows:
3443 - New object identifiers are inserted in objects.txt, following
3444 the syntax given in objects.README.
3445 - objects.pl is used to process obj_mac.num and create a new
3447 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3450 This is currently kind of a hack, and the perl code in objects.pl
3451 isn't very elegant, but it works as I intended. The simplest way
3452 to check that it worked correctly is to look in obj_dat.h and
3453 check the array nid_objs and make sure the objects haven't moved
3454 around (this is important!). Additions are OK, as well as
3455 consistent name changes.
3458 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3461 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3462 The given file adds to whatever has already been seeded into the
3463 random pool through the RANDFILE configuration file option or
3464 environment variable, or the default random state file.
3467 *) mkstack.pl now sorts each macro group into lexical order.
3468 Previously the output order depended on the order the files
3469 appeared in the directory, resulting in needless rewriting
3473 *) Patches to make OpenSSL compile under Win32 again. Mostly
3474 work arounds for the VC++ problem that it treats func() as
3475 func(void). Also stripped out the parts of mkdef.pl that
3476 added extra typesafe functions: these no longer exist.
3479 *) Reorganisation of the stack code. The macros are now all
3480 collected in safestack.h . Each macro is defined in terms of
3481 a "stack macro" of the form SKM_<name>(type, a, b). The
3482 DEBUG_SAFESTACK is now handled in terms of function casts,
3483 this has the advantage of retaining type safety without the
3484 use of additional functions. If DEBUG_SAFESTACK is not defined
3485 then the non typesafe macros are used instead. Also modified the
3486 mkstack.pl script to handle the new form. Needs testing to see
3487 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3488 the default if no major problems. Similar behaviour for ASN1_SET_OF
3489 and PKCS12_STACK_OF.
3492 *) When some versions of IIS use the 'NET' form of private key the
3493 key derivation algorithm is different. Normally MD5(password) is
3494 used as a 128 bit RC4 key. In the modified case
3495 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3496 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3497 as the old Netscape_RSA functions except they have an additional
3498 'sgckey' parameter which uses the modified algorithm. Also added
3499 an -sgckey command line option to the rsa utility. Thanks to
3500 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3501 algorithm to openssl-dev.
3504 *) The evp_local.h macros were using 'c.##kname' which resulted in
3505 invalid expansion on some systems (SCO 5.0.5 for example).
3506 Corrected to 'c.kname'.
3507 [Phillip Porch <root@theporch.com>]
3509 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3510 a STACK of email addresses from a certificate or request, these look
3511 in the subject name and the subject alternative name extensions and
3512 omit any duplicate addresses.
3515 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3516 This makes DSA verification about 2 % faster.
3519 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3520 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3521 plus overhead for 1024 bit moduli).
3522 This makes exponentiations about 0.5 % faster for 1024 bit
3523 exponents (as measured by "openssl speed rsa2048").
3526 *) Rename memory handling macros to avoid conflicts with other
3528 Malloc => OPENSSL_malloc
3529 Malloc_locked => OPENSSL_malloc_locked
3530 Realloc => OPENSSL_realloc
3531 Free => OPENSSL_free
3534 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3535 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3538 *) CygWin32 support.
3539 [John Jarvie <jjarvie@newsguy.com>]
3541 *) The type-safe stack code has been rejigged. It is now only compiled
3542 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3543 by default all type-specific stack functions are "#define"d back to
3544 standard stack functions. This results in more streamlined output
3545 but retains the type-safety checking possibilities of the original
3549 *) The STACK code has been cleaned up, and certain type declarations
3550 that didn't make a lot of sense have been brought in line. This has
3551 also involved a cleanup of sorts in safestack.h to more correctly
3552 map type-safe stack functions onto their plain stack counterparts.
3553 This work has also resulted in a variety of "const"ifications of
3554 lots of the code, especially "_cmp" operations which should normally
3555 be prototyped with "const" parameters anyway.
3558 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3559 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3560 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3561 where all of 'md' is used each time the PRNG is used, but 'state'
3562 is used only indexed by a cyclic counter. As entropy may not be
3563 well distributed from the beginning, 'md' is important as a
3564 chaining variable. However, the output function chains only half
3565 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3566 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3567 in all of 'state' being rewritten, with the new values depending
3568 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3571 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3572 the handshake is continued after ssl_verify_cert_chain();
3573 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3574 can lead to 'unexplainable' connection aborts later.
3575 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3577 *) Major EVP API cipher revision.
3578 Add hooks for extra EVP features. This allows various cipher
3579 parameters to be set in the EVP interface. Support added for variable
3580 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3581 setting of RC2 and RC5 parameters.
3583 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3586 Remove lots of duplicated code from the EVP library. For example *every*
3587 cipher init() function handles the 'iv' in the same way according to the
3588 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3589 for CFB and OFB modes they zero ctx->num.
3591 New functionality allows removal of S/MIME code RC2 hack.
3593 Most of the routines have the same form and so can be declared in terms
3596 By shifting this to the top level EVP_CipherInit() it can be removed from
3597 all individual ciphers. If the cipher wants to handle IVs or keys
3598 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3601 Change lots of functions like EVP_EncryptUpdate() to now return a
3602 value: although software versions of the algorithms cannot fail
3603 any installed hardware versions can.
3606 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3607 this option is set, tolerate broken clients that send the negotiated
3608 protocol version number instead of the requested protocol version
3612 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3613 i.e. non-zero for export ciphersuites, zero otherwise.
3614 Previous versions had this flag inverted, inconsistent with
3615 rsa_tmp_cb (..._TMP_RSA_CB).
3616 [Bodo Moeller; problem reported by Amit Chopra]
3618 *) Add missing DSA library text string. Work around for some IIS
3619 key files with invalid SEQUENCE encoding.
3622 *) Add a document (doc/standards.txt) that list all kinds of standards
3623 and so on that are implemented in OpenSSL.
3626 *) Enhance c_rehash script. Old version would mishandle certificates
3627 with the same subject name hash and wouldn't handle CRLs at all.
3628 Added -fingerprint option to crl utility, to support new c_rehash
3632 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3635 *) Fix for SSL server purpose checking. Server checking was
3636 rejecting certificates which had extended key usage present
3637 but no ssl client purpose.
3638 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3640 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3641 is a little unclear about how a blank password is handled.
3642 Since the password in encoded as a BMPString with terminating
3643 double NULL a zero length password would end up as just the
3644 double NULL. However no password at all is different and is
3645 handled differently in the PKCS#12 key generation code. NS
3646 treats a blank password as zero length. MSIE treats it as no
3647 password on export: but it will try both on import. We now do
3648 the same: PKCS12_parse() tries zero length and no password if
3649 the password is set to "" or NULL (NULL is now a valid password:
3650 it wasn't before) as does the pkcs12 application.
3653 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3654 perror when PEM_read_bio_X509_REQ fails, the error message must
3655 be obtained from the error queue.
3658 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3659 it in ERR_remove_state if appropriate, and change ERR_get_state
3660 accordingly to avoid race conditions (this is necessary because
3661 thread_hash is no longer constant once set).
3664 *) Bugfix for linux-elf makefile.one.
3667 *) RSA_get_default_method() will now cause a default
3668 RSA_METHOD to be chosen if one doesn't exist already.
3669 Previously this was only set during a call to RSA_new()
3670 or RSA_new_method(NULL) meaning it was possible for
3671 RSA_get_default_method() to return NULL.
3674 *) Added native name translation to the existing DSO code
3675 that will convert (if the flag to do so is set) filenames
3676 that are sufficiently small and have no path information
3677 into a canonical native form. Eg. "blah" converted to
3678 "libblah.so" or "blah.dll" etc.
3681 *) New function ERR_error_string_n(e, buf, len) which is like
3682 ERR_error_string(e, buf), but writes at most 'len' bytes
3683 including the 0 terminator. For ERR_error_string_n, 'buf'
3685 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3687 *) CONF library reworked to become more general. A new CONF
3688 configuration file reader "class" is implemented as well as a
3689 new functions (NCONF_*, for "New CONF") to handle it. The now
3690 old CONF_* functions are still there, but are reimplemented to
3691 work in terms of the new functions. Also, a set of functions
3692 to handle the internal storage of the configuration data is
3693 provided to make it easier to write new configuration file
3694 reader "classes" (I can definitely see something reading a
3695 configuration file in XML format, for example), called _CONF_*,
3696 or "the configuration storage API"...
3698 The new configuration file reading functions are:
3700 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3701 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3703 NCONF_default, NCONF_WIN32
3705 NCONF_dump_fp, NCONF_dump_bio
3707 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3708 NCONF_new creates a new CONF object. This works in the same way
3709 as other interfaces in OpenSSL, like the BIO interface.
3710 NCONF_dump_* dump the internal storage of the configuration file,
3711 which is useful for debugging. All other functions take the same
3712 arguments as the old CONF_* functions wth the exception of the
3713 first that must be a `CONF *' instead of a `LHASH *'.
3715 To make it easer to use the new classes with the old CONF_* functions,
3716 the function CONF_set_default_method is provided.
3719 *) Add '-tls1' option to 'openssl ciphers', which was already
3720 mentioned in the documentation but had not been implemented.
3721 (This option is not yet really useful because even the additional
3722 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3725 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3726 OpenSSL-based applications) load shared libraries and bind to
3727 them in a portable way.
3728 [Geoff Thorpe, with contributions from Richard Levitte]
3730 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3732 *) Make sure _lrotl and _lrotr are only used with MSVC.
3734 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3735 (the default implementation of RAND_status).
3737 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3738 to '-clrext' (= clear extensions), as intended and documented.
3739 [Bodo Moeller; inconsistency pointed out by Michael Attili
3742 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3743 was larger than the MD block size.
3744 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3746 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3747 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3748 using the passed key: if the passed key was a private key the result
3749 of X509_print(), for example, would be to print out all the private key
3753 *) des_quad_cksum() byte order bug fix.
3754 [Ulf Möller, using the problem description in krb4-0.9.7, where
3755 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3757 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3759 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3761 *) For easily testing in shell scripts whether some command
3762 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3763 returns with exit code 0 iff no command of the given name is available.
3764 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3765 the output goes to stdout and nothing is printed to stderr.
3766 Additional arguments are always ignored.
3768 Since for each cipher there is a command of the same name,
3769 the 'no-cipher' compilation switches can be tested this way.
3771 ('openssl no-XXX' is not able to detect pseudo-commands such
3772 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3775 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3778 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3779 is set; it will be thrown away anyway because each handshake creates
3781 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3782 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3783 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3784 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3787 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3788 'Q' and 'R' lose their special meanings (quit/renegotiate).
3789 This is part of what -quiet does; unlike -quiet, -ign_eof
3790 does not suppress any output.
3793 *) Add compatibility options to the purpose and trust code. The
3794 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3795 accepts a certificate or CA, this was the previous behaviour,
3796 with all the associated security issues.
3798 X509_TRUST_COMPAT is the old trust behaviour: only and
3799 automatically trust self signed roots in certificate store. A
3800 new trust setting X509_TRUST_DEFAULT is used to specify that
3801 a purpose has no associated trust setting and it should instead
3802 use the value in the default purpose.
3805 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3806 and fix a memory leak.
3809 *) In util/mkerr.pl (which implements 'make errors'), preserve
3810 reason strings from the previous version of the .c file, as
3811 the default to have only downcase letters (and digits) in
3812 automatically generated reasons codes is not always appropriate.
3815 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3816 using strerror. Previously, ERR_reason_error_string() returned
3817 library names as reason strings for SYSerr; but SYSerr is a special
3818 case where small numbers are errno values, not library numbers.
3821 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3822 converts DSA parameters into DH parameters. (When creating parameters,
3823 DSA_generate_parameters is used.)
3826 *) Include 'length' (recommended exponent length) in C code generated
3827 by 'openssl dhparam -C'.
3830 *) The second argument to set_label in perlasm was already being used
3831 so couldn't be used as a "file scope" flag. Moved to third argument
3835 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3836 instead of RAND_bytes for encryption IVs and salts.
3839 *) Include RAND_status() into RAND_METHOD instead of implementing
3840 it only for md_rand.c Otherwise replacing the PRNG by calling
3841 RAND_set_rand_method would be impossible.
3844 *) Don't let DSA_generate_key() enter an infinite loop if the random
3845 number generation fails.
3848 *) New 'rand' application for creating pseudo-random output.
3851 *) Added configuration support for Linux/IA64
3852 [Rolf Haberrecker <rolf@suse.de>]
3854 *) Assembler module support for Mingw32.
3857 *) Shared library support for HPUX (in shlib/).
3858 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3860 *) Shared library support for Solaris gcc.
3861 [Lutz Behnke <behnke@trustcenter.de>]
3863 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3865 *) PKCS7_encrypt() was adding text MIME headers twice because they
3866 were added manually and by SMIME_crlf_copy().
3869 *) In bntest.c don't call BN_rand with zero bits argument.
3870 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3872 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3873 case was implemented. This caused BN_div_recp() to fail occasionally.
3876 *) Add an optional second argument to the set_label() in the perl
3877 assembly language builder. If this argument exists and is set
3878 to 1 it signals that the assembler should use a symbol whose
3879 scope is the entire file, not just the current function. This
3880 is needed with MASM which uses the format label:: for this scope.
3881 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3883 *) Change the ASN1 types so they are typedefs by default. Before
3884 almost all types were #define'd to ASN1_STRING which was causing
3885 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3889 *) Change names of new functions to the new get1/get0 naming
3890 convention: After 'get1', the caller owns a reference count
3891 and has to call ..._free; 'get0' returns a pointer to some
3892 data structure without incrementing reference counters.
3893 (Some of the existing 'get' functions increment a reference
3894 counter, some don't.)
3895 Similarly, 'set1' and 'add1' functions increase reference
3896 counters or duplicate objects.
3899 *) Allow for the possibility of temp RSA key generation failure:
3900 the code used to assume it always worked and crashed on failure.
3903 *) Fix potential buffer overrun problem in BIO_printf().
3904 [Ulf Möller, using public domain code by Patrick Powell; problem
3905 pointed out by David Sacerdote <das33@cornell.edu>]
3907 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3908 RAND_egd() and RAND_status(). In the command line application,
3909 the EGD socket can be specified like a seed file using RANDFILE
3913 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3914 Some CAs (e.g. Verisign) distribute certificates in this form.
3917 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3918 list to exclude them. This means that no special compilation option
3919 is needed to use anonymous DH: it just needs to be included in the
3923 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3924 EVP_MD_type. The old functionality is available in a new macro called
3925 EVP_MD_md(). Change code that uses it and update docs.
3928 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3929 where the 'void *' argument is replaced by a function pointer argument.
3930 Previously 'void *' was abused to point to functions, which works on
3931 many platforms, but is not correct. As these functions are usually
3932 called by macros defined in OpenSSL header files, most source code
3933 should work without changes.
3936 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3937 sections with information on -D... compiler switches used for
3938 compiling the library so that applications can see them. To enable
3939 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3940 must be defined. E.g.,
3941 #define OPENSSL_ALGORITHM_DEFINES
3942 #include <openssl/opensslconf.h>
3943 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3944 [Richard Levitte, Ulf and Bodo Möller]
3946 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3950 *) Change the 'other' type in certificate aux info to a STACK_OF
3951 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3952 the required ASN1 format: arbitrary types determined by an OID.
3955 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3956 argument to 'req'. This is not because the function is newer or
3957 better than others it just uses the work 'NEW' in the certificate
3958 request header lines. Some software needs this.
3961 *) Reorganise password command line arguments: now passwords can be
3962 obtained from various sources. Delete the PEM_cb function and make
3963 it the default behaviour: i.e. if the callback is NULL and the
3964 usrdata argument is not NULL interpret it as a null terminated pass
3965 phrase. If usrdata and the callback are NULL then the pass phrase
3966 is prompted for as usual.
3969 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3970 the support is automatically enabled. The resulting binaries will
3971 autodetect the card and use it if present.
3972 [Ben Laurie and Compaq Inc.]
3974 *) Work around for Netscape hang bug. This sends certificate request
3975 and server done in one record. Since this is perfectly legal in the
3976 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3977 the bugs/SSLv3 entry for more info.
3980 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3983 *) Add -rand argument to smime and pkcs12 applications and read/write
3987 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3990 *) Add command line password options to the remaining applications.
3993 *) Bug fix for BN_div_recp() for numerators with an even number of
3997 *) More tests in bntest.c, and changed test_bn output.
4000 *) ./config recognizes MacOS X now.
4003 *) Bug fix for BN_div() when the first words of num and divsor are
4004 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4007 *) Add support for various broken PKCS#8 formats, and command line
4008 options to produce them.
4011 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4012 get temporary BIGNUMs from a BN_CTX.
4015 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4019 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4020 include a #define from the old name to the new. The original intent
4021 was that statically linked binaries could for example just call
4022 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4023 link with digests. This never worked becayse SSLeay_add_all_digests()
4024 and SSLeay_add_all_ciphers() were in the same source file so calling
4025 one would link with the other. They are now in separate source files.
4028 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4031 *) Use a less unusual form of the Miller-Rabin primality test (it used
4032 a binary algorithm for exponentiation integrated into the Miller-Rabin
4033 loop, our standard modexp algorithms are faster).
4036 *) Support for the EBCDIC character set completed.
4037 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4039 *) Source code cleanups: use const where appropriate, eliminate casts,
4040 use void * instead of char * in lhash.
4043 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4044 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4045 this the server could overwrite ephemeral keys that the client
4049 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4050 using 50 iterations of the Rabin-Miller test.
4052 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4053 iterations of the Rabin-Miller test as required by the appendix
4054 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4055 As BN_is_prime_fasttest includes trial division, DSA parameter
4056 generation becomes much faster.
4058 This implies a change for the callback functions in DSA_is_prime
4059 and DSA_generate_parameters: The callback function is called once
4060 for each positive witness in the Rabin-Miller test, not just
4061 occasionally in the inner loop; and the parameters to the
4062 callback function now provide an iteration count for the outer
4063 loop rather than for the current invocation of the inner loop.
4064 DSA_generate_parameters additionally can call the callback
4065 function with an 'iteration count' of -1, meaning that a
4066 candidate has passed the trial division test (when q is generated
4067 from an application-provided seed, trial division is skipped).
4070 *) New function BN_is_prime_fasttest that optionally does trial
4071 division before starting the Rabin-Miller test and has
4072 an additional BN_CTX * argument (whereas BN_is_prime always
4073 has to allocate at least one BN_CTX).
4074 'callback(1, -1, cb_arg)' is called when a number has passed the
4075 trial division stage.
4078 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4082 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4085 *) New function BN_pseudo_rand().
4088 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4089 bignum version of BN_from_montgomery() with the working code from
4090 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4094 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4095 made it impossible to use the same SSL_SESSION data structure in
4096 SSL2 clients in multiple threads.
4099 *) The return value of RAND_load_file() no longer counts bytes obtained
4100 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4101 to seed the PRNG (previously an explicit byte count was required).
4102 [Ulf Möller, Bodo Möller]
4104 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4105 used (char *) instead of (void *) and had casts all over the place.
4108 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4111 *) Retain source code compatibility for BN_prime_checks macro:
4112 BN_is_prime(..., BN_prime_checks, ...) now uses
4113 BN_prime_checks_for_size to determine the appropriate number of
4114 Rabin-Miller iterations.
4117 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4118 DH_CHECK_P_NOT_SAFE_PRIME.
4119 (Check if this is true? OpenPGP calls them "strong".)
4122 *) Merge the functionality of "dh" and "gendh" programs into a new program
4123 "dhparam". The old programs are retained for now but will handle DH keys
4124 (instead of parameters) in future.
4127 *) Make the ciphers, s_server and s_client programs check the return values
4128 when a new cipher list is set.
4131 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4132 ciphers. Before when the 56bit ciphers were enabled the sorting was
4135 The syntax for the cipher sorting has been extended to support sorting by
4136 cipher-strength (using the strength_bits hard coded in the tables).
4137 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4139 Fix a bug in the cipher-command parser: when supplying a cipher command
4140 string with an "undefined" symbol (neither command nor alphanumeric
4141 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4142 an error is flagged.
4144 Due to the strength-sorting extension, the code of the
4145 ssl_create_cipher_list() function was completely rearranged. I hope that
4146 the readability was also increased :-)
4147 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4149 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4150 for the first serial number and places 2 in the serial number file. This
4151 avoids problems when the root CA is created with serial number zero and
4152 the first user certificate has the same issuer name and serial number
4156 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4157 the new code. Add documentation for this stuff.
4160 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4161 X509_*() to X509at_*() on the grounds that they don't handle X509
4162 structures and behave in an analagous way to the X509v3 functions:
4163 they shouldn't be called directly but wrapper functions should be used
4166 So we also now have some wrapper functions that call the X509at functions
4167 when passed certificate requests. (TO DO: similar things can be done with
4168 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4169 things. Some of these need some d2i or i2d and print functionality
4170 because they handle more complex structures.)
4173 *) Add missing #ifndefs that caused missing symbols when building libssl
4174 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4175 NO_RSA in ssl/s2*.c.
4176 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4178 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4179 has a return value which indicates the quality of the random data
4180 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4181 error queue. New function RAND_pseudo_bytes() generates output that is
4182 guaranteed to be unique but not unpredictable. RAND_add is like
4183 RAND_seed, but takes an extra argument for an entropy estimate
4184 (RAND_seed always assumes full entropy).
4187 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4188 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4189 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4190 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4191 false-positive rate of at most 2^-80 for random input.
4194 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4197 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4198 in the 0.9.5 release), this returns the chain
4199 from an X509_CTX structure with a dup of the stack and all
4200 the X509 reference counts upped: so the stack will exist
4201 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4204 Also make SSL_SESSION_print() print out the verify return
4208 *) Add manpage for the pkcs12 command. Also change the default
4209 behaviour so MAC iteration counts are used unless the new
4210 -nomaciter option is used. This improves file security and
4211 only older versions of MSIE (4.0 for example) need it.
4214 *) Honor the no-xxx Configure options when creating .DEF files.
4217 *) Add PKCS#10 attributes to field table: challengePassword,
4218 unstructuredName and unstructuredAddress. These are taken from
4219 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4220 international characters are used.
4222 More changes to X509_ATTRIBUTE code: allow the setting of types
4223 based on strings. Remove the 'loc' parameter when adding
4224 attributes because these will be a SET OF encoding which is sorted
4228 *) Initial changes to the 'req' utility to allow request generation
4229 automation. This will allow an application to just generate a template
4230 file containing all the field values and have req construct the
4233 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4234 used all over the place including certificate requests and PKCS#7
4235 structures. They are currently handled manually where necessary with
4236 some primitive wrappers for PKCS#7. The new functions behave in a
4237 manner analogous to the X509 extension functions: they allow
4238 attributes to be looked up by NID and added.
4240 Later something similar to the X509V3 code would be desirable to
4241 automatically handle the encoding, decoding and printing of the
4242 more complex types. The string types like challengePassword can
4243 be handled by the string table functions.
4245 Also modified the multi byte string table handling. Now there is
4246 a 'global mask' which masks out certain types. The table itself
4247 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4248 is useful when for example there is only one permissible type
4249 (as in countryName) and using the mask might result in no valid
4253 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4254 SSL_get_peer_finished to allow applications to obtain the latest
4255 Finished messages sent to the peer or expected from the peer,
4256 respectively. (SSL_get_peer_finished is usually the Finished message
4257 actually received from the peer, otherwise the protocol will be aborted.)
4259 As the Finished message are message digests of the complete handshake
4260 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4261 be used for external authentication procedures when the authentication
4262 provided by SSL/TLS is not desired or is not enough.
4265 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4266 the host supports BWX extension and if Compaq C is present on the
4267 $PATH. Just exploiting of the BWX extension results in 20-30%
4268 performance kick for some algorithms, e.g. DES and RC4 to mention
4269 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4273 *) Add support for MS "fast SGC". This is arguably a violation of the
4274 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4275 weak crypto and after checking the certificate is SGC a second one
4276 with strong crypto. MS SGC stops the first handshake after receiving
4277 the server certificate message and sends a second client hello. Since
4278 a server will typically do all the time consuming operations before
4279 expecting any further messages from the client (server key exchange
4280 is the most expensive) there is little difference between the two.
4282 To get OpenSSL to support MS SGC we have to permit a second client
4283 hello message after we have sent server done. In addition we have to
4284 reset the MAC if we do get this second client hello.
4287 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4288 if a DER encoded private key is RSA or DSA traditional format. Changed
4289 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4290 format DER encoded private key. Newer code should use PKCS#8 format which
4291 has the key type encoded in the ASN1 structure. Added DER private key
4292 support to pkcs8 application.
4295 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4296 ciphersuites has been selected (as required by the SSL 3/TLS 1
4297 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4298 is set, we interpret this as a request to violate the specification
4299 (the worst that can happen is a handshake failure, and 'correct'
4300 behaviour would result in a handshake failure anyway).
4303 *) In SSL_CTX_add_session, take into account that there might be multiple
4304 SSL_SESSION structures with the same session ID (e.g. when two threads
4305 concurrently obtain them from an external cache).
4306 The internal cache can handle only one SSL_SESSION with a given ID,
4307 so if there's a conflict, we now throw out the old one to achieve
4311 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4312 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4313 some routines that use cipher OIDs: some ciphers do not have OIDs
4314 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4318 *) Simplify the trust setting structure and code. Now we just have
4319 two sequences of OIDs for trusted and rejected settings. These will
4320 typically have values the same as the extended key usage extension
4321 and any application specific purposes.
4323 The trust checking code now has a default behaviour: it will just
4324 check for an object with the same NID as the passed id. Functions can
4325 be provided to override either the default behaviour or the behaviour
4326 for a given id. SSL client, server and email already have functions
4327 in place for compatibility: they check the NID and also return "trusted"
4328 if the certificate is self signed.
4331 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4332 traditional format into an EVP_PKEY structure.
4335 *) Add a password callback function PEM_cb() which either prompts for
4336 a password if usr_data is NULL or otherwise assumes it is a null
4337 terminated password. Allow passwords to be passed on command line
4338 environment or config files in a few more utilities.
4341 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4342 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4343 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4344 Update documentation.
4347 *) Support for ASN1 "NULL" type. This could be handled before by using
4348 ASN1_TYPE but there wasn't any function that would try to read a NULL
4349 and produce an error if it couldn't. For compatibility we also have
4350 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4351 don't allocate anything because they don't need to.
4354 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4356 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4358 *) Rebuild of the memory allocation routines used by OpenSSL code and
4359 possibly others as well. The purpose is to make an interface that
4360 provide hooks so anyone can build a separate set of allocation and
4361 deallocation routines to be used by OpenSSL, for example memory
4362 pool implementations, or something else, which was previously hard
4363 since Malloc(), Realloc() and Free() were defined as macros having
4364 the values malloc, realloc and free, respectively (except for Win32
4365 compilations). The same is provided for memory debugging code.
4366 OpenSSL already comes with functionality to find memory leaks, but
4367 this gives people a chance to debug other memory problems.
4369 With these changes, a new set of functions and macros have appeared:
4371 CRYPTO_set_mem_debug_functions() [F]
4372 CRYPTO_get_mem_debug_functions() [F]
4373 CRYPTO_dbg_set_options() [F]
4374 CRYPTO_dbg_get_options() [F]
4375 CRYPTO_malloc_debug_init() [M]
4377 The memory debug functions are NULL by default, unless the library
4378 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4379 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4380 gives the standard debugging functions that come with OpenSSL) or
4381 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4382 provided by the library user) must be used. When the standard
4383 debugging functions are used, CRYPTO_dbg_set_options can be used to
4384 request additional information:
4385 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4386 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4388 Also, things like CRYPTO_set_mem_functions will always give the
4389 expected result (the new set of functions is used for allocation
4390 and deallocation) at all times, regardless of platform and compiler
4393 To finish it up, some functions that were never use in any other
4394 way than through macros have a new API and new semantic:
4397 CRYPTO_dbg_realloc()
4400 All macros of value have retained their old syntax.
4401 [Richard Levitte and Bodo Moeller]
4403 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4404 ordering of SMIMECapabilities wasn't in "strength order" and there
4405 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4409 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4410 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4411 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4413 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4414 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4415 functionality to handle multipart/signed properly) and a utility
4416 called 'smime' to call all this stuff. This is based on code I
4417 originally wrote for Celo who have kindly allowed it to be
4418 included in OpenSSL.
4421 *) Add variants des_set_key_checked and des_set_key_unchecked of
4422 des_set_key (aka des_key_sched). Global variable des_check_key
4423 decides which of these is called by des_set_key; this way
4424 des_check_key behaves as it always did, but applications and
4425 the library itself, which was buggy for des_check_key == 1,
4426 have a cleaner way to pick the version they need.
4429 *) New function PKCS12_newpass() which changes the password of a
4433 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4434 dynamic mix. In both cases the ids can be used as an index into the
4435 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4436 functions so they accept a list of the field values and the
4437 application doesn't need to directly manipulate the X509_TRUST
4441 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4445 *) Modify the way the V3 extension code looks up extensions. This now
4446 works in a similar way to the object code: we have some "standard"
4447 extensions in a static table which is searched with OBJ_bsearch()
4448 and the application can add dynamic ones if needed. The file
4449 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4450 updated whenever a new extension is added to the core code and kept
4451 in ext_nid order. There is a simple program 'tabtest.c' which checks
4452 this. New extensions are not added too often so this file can readily
4453 be maintained manually.
4455 There are two big advantages in doing things this way. The extensions
4456 can be looked up immediately and no longer need to be "added" using
4457 X509V3_add_standard_extensions(): this function now does nothing.
4458 [Side note: I get *lots* of email saying the extension code doesn't
4459 work because people forget to call this function]
4460 Also no dynamic allocation is done unless new extensions are added:
4461 so if we don't add custom extensions there is no need to call
4462 X509V3_EXT_cleanup().
4465 *) Modify enc utility's salting as follows: make salting the default. Add a
4466 magic header, so unsalted files fail gracefully instead of just decrypting
4467 to garbage. This is because not salting is a big security hole, so people
4468 should be discouraged from doing it.
4471 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4472 digest to be passed on the command line but it only used this
4473 parameter when signing a certificate. Modified so all relevant
4474 operations are affected by the digest parameter including the
4475 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4476 DSA key was used because it didn't fix the digest.
4479 *) Initial certificate chain verify code. Currently tests the untrusted
4480 certificates for consistency with the verify purpose (which is set
4481 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4483 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4484 this is because it will reject chains with invalid extensions whereas
4485 every previous version of OpenSSL and SSLeay made no checks at all.
4487 Trust code: checks the root CA for the relevant trust settings. Trust
4488 settings have an initial value consistent with the verify purpose: e.g.
4489 if the verify purpose is for SSL client use it expects the CA to be
4490 trusted for SSL client use. However the default value can be changed to
4491 permit custom trust settings: one example of this would be to only trust
4492 certificates from a specific "secure" set of CAs.
4494 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4495 which should be used for version portability: especially since the
4496 verify structure is likely to change more often now.
4498 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4499 to set them. If not set then assume SSL clients will verify SSL servers
4502 Two new options to the verify program: -untrusted allows a set of
4503 untrusted certificates to be passed in and -purpose which sets the
4504 intended purpose of the certificate. If a purpose is set then the
4505 new chain verify code is used to check extension consistency.
4508 *) Support for the authority information access extension.
4511 *) Modify RSA and DSA PEM read routines to transparently handle
4512 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4513 public keys in a format compatible with certificate
4514 SubjectPublicKeyInfo structures. Unfortunately there were already
4515 functions called *_PublicKey_* which used various odd formats so
4516 these are retained for compatibility: however the DSA variants were
4517 never in a public release so they have been deleted. Changed dsa/rsa
4518 utilities to handle the new format: note no releases ever handled public
4519 keys so we should be OK.
4521 The primary motivation for this change is to avoid the same fiasco
4522 that dogs private keys: there are several incompatible private key
4523 formats some of which are standard and some OpenSSL specific and
4524 require various evil hacks to allow partial transparent handling and
4525 even then it doesn't work with DER formats. Given the option anything
4526 other than PKCS#8 should be dumped: but the other formats have to
4527 stay in the name of compatibility.
4529 With public keys and the benefit of hindsight one standard format
4530 is used which works with EVP_PKEY, RSA or DSA structures: though
4531 it clearly returns an error if you try to read the wrong kind of key.
4533 Added a -pubkey option to the 'x509' utility to output the public key.
4534 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4535 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4536 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4537 that do the same as the EVP_PKEY_assign_*() except they up the
4538 reference count of the added key (they don't "swallow" the
4542 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4543 CRLs would fail if the file contained no certificates or no CRLs:
4544 added a new function to read in both types and return the number
4545 read: this means that if none are read it will be an error. The
4546 DER versions of the certificate and CRL reader would always fail
4547 because it isn't possible to mix certificates and CRLs in DER format
4548 without choking one or the other routine. Changed this to just read
4549 a certificate: this is the best we can do. Also modified the code
4550 in apps/verify.c to take notice of return codes: it was previously
4551 attempting to read in certificates from NULL pointers and ignoring
4552 any errors: this is one reason why the cert and CRL reader seemed
4553 to work. It doesn't check return codes from the default certificate
4554 routines: these may well fail if the certificates aren't installed.
4557 *) Code to support otherName option in GeneralName.
4560 *) First update to verify code. Change the verify utility
4561 so it warns if it is passed a self signed certificate:
4562 for consistency with the normal behaviour. X509_verify
4563 has been modified to it will now verify a self signed
4564 certificate if *exactly* the same certificate appears
4565 in the store: it was previously impossible to trust a
4566 single self signed certificate. This means that:
4567 openssl verify ss.pem
4568 now gives a warning about a self signed certificate but
4569 openssl verify -CAfile ss.pem ss.pem
4573 *) For servers, store verify_result in SSL_SESSION data structure
4574 (and add it to external session representation).
4575 This is needed when client certificate verifications fails,
4576 but an application-provided verification callback (set by
4577 SSL_CTX_set_cert_verify_callback) allows accepting the session
4578 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4579 but returns 1): When the session is reused, we have to set
4580 ssl->verify_result to the appropriate error code to avoid
4582 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4584 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4585 case in PKCS7_dataInit() where the signed PKCS7 structure
4586 didn't contain any existing data because it was being created.
4587 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4589 *) Add a salt to the key derivation routines in enc.c. This
4590 forms the first 8 bytes of the encrypted file. Also add a
4591 -S option to allow a salt to be input on the command line.
4594 *) New function X509_cmp(). Oddly enough there wasn't a function
4595 to compare two certificates. We do this by working out the SHA1
4596 hash and comparing that. X509_cmp() will be needed by the trust
4600 *) SSL_get1_session() is like SSL_get_session(), but increments
4601 the reference count in the SSL_SESSION returned.
4602 [Geoff Thorpe <geoff@eu.c2.net>]
4604 *) Fix for 'req': it was adding a null to request attributes.
4605 Also change the X509_LOOKUP and X509_INFO code to handle
4606 certificate auxiliary information.
4609 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4613 *) Add the possibility to add extra information to the memory leak
4614 detecting output, to form tracebacks, showing from where each
4615 allocation was originated: CRYPTO_push_info("constant string") adds
4616 the string plus current file name and line number to a per-thread
4617 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4618 is like calling CYRPTO_pop_info() until the stack is empty.
4619 Also updated memory leak detection code to be multi-thread-safe.
4622 *) Add options -text and -noout to pkcs7 utility and delete the
4623 encryption options which never did anything. Update docs.
4626 *) Add options to some of the utilities to allow the pass phrase
4627 to be included on either the command line (not recommended on
4628 OSes like Unix) or read from the environment. Update the
4629 manpages and fix a few bugs.
4632 *) Add a few manpages for some of the openssl commands.
4635 *) Fix the -revoke option in ca. It was freeing up memory twice,
4636 leaking and not finding already revoked certificates.
4639 *) Extensive changes to support certificate auxiliary information.
4640 This involves the use of X509_CERT_AUX structure and X509_AUX
4641 functions. An X509_AUX function such as PEM_read_X509_AUX()
4642 can still read in a certificate file in the usual way but it
4643 will also read in any additional "auxiliary information". By
4644 doing things this way a fair degree of compatibility can be
4645 retained: existing certificates can have this information added
4646 using the new 'x509' options.
4648 Current auxiliary information includes an "alias" and some trust
4649 settings. The trust settings will ultimately be used in enhanced
4650 certificate chain verification routines: currently a certificate
4651 can only be trusted if it is self signed and then it is trusted
4655 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4656 The problem was that one of the replacement routines had not been working
4657 since SSLeay releases. For now the offending routine has been replaced
4658 with non-optimised assembler. Even so, this now gives around 95%
4659 performance improvement for 1024 bit RSA signs.
4662 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4663 handling. Most clients have the effective key size in bits equal to
4664 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4665 A few however don't do this and instead use the size of the decrypted key
4666 to determine the RC2 key length and the AlgorithmIdentifier to determine
4667 the effective key length. In this case the effective key length can still
4668 be 40 bits but the key length can be 168 bits for example. This is fixed
4669 by manually forcing an RC2 key into the EVP_PKEY structure because the
4670 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4671 the key length and effective key length are equal.
4674 *) Add a bunch of functions that should simplify the creation of
4675 X509_NAME structures. Now you should be able to do:
4676 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4677 and have it automatically work out the correct field type and fill in
4678 the structures. The more adventurous can try:
4679 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4680 and it will (hopefully) work out the correct multibyte encoding.
4683 *) Change the 'req' utility to use the new field handling and multibyte
4684 copy routines. Before the DN field creation was handled in an ad hoc
4685 way in req, ca, and x509 which was rather broken and didn't support
4686 BMPStrings or UTF8Strings. Since some software doesn't implement
4687 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4688 using the dirstring_type option. See the new comment in the default
4689 openssl.cnf for more info.
4692 *) Make crypto/rand/md_rand.c more robust:
4693 - Assure unique random numbers after fork().
4694 - Make sure that concurrent threads access the global counter and
4695 md serializably so that we never lose entropy in them
4696 or use exactly the same state in multiple threads.
4697 Access to the large state is not always serializable because
4698 the additional locking could be a performance killer, and
4699 md should be large enough anyway.
4702 *) New file apps/app_rand.c with commonly needed functionality
4703 for handling the random seed file.
4705 Use the random seed file in some applications that previously did not:
4707 dsaparam -genkey (which also ignored its '-rand' option),
4710 x509 (when signing).
4711 Except on systems with /dev/urandom, it is crucial to have a random
4712 seed file at least for key creation, DSA signing, and for DH exchanges;
4713 for RSA signatures we could do without one.
4715 gendh and gendsa (unlike genrsa) used to read only the first byte
4716 of each file listed in the '-rand' option. The function as previously
4717 found in genrsa is now in app_rand.c and is used by all programs
4718 that support '-rand'.
4721 *) In RAND_write_file, use mode 0600 for creating files;
4722 don't just chmod when it may be too late.
4725 *) Report an error from X509_STORE_load_locations
4726 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4729 *) New function ASN1_mbstring_copy() this copies a string in either
4730 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4731 into an ASN1_STRING type. A mask of permissible types is passed
4732 and it chooses the "minimal" type to use or an error if not type
4736 *) Add function equivalents to the various macros in asn1.h. The old
4737 macros are retained with an M_ prefix. Code inside the library can
4738 use the M_ macros. External code (including the openssl utility)
4739 should *NOT* in order to be "shared library friendly".
4742 *) Add various functions that can check a certificate's extensions
4743 to see if it usable for various purposes such as SSL client,
4744 server or S/MIME and CAs of these types. This is currently
4745 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4746 verification. Also added a -purpose flag to x509 utility to
4747 print out all the purposes.
4750 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4754 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4755 for, obtain and decode and extension and obtain its critical flag.
4756 This allows all the necessary extension code to be handled in a
4757 single function call.
4760 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4761 platforms. See crypto/rc4/rc4_enc.c for further details.
4764 *) New -noout option to asn1parse. This causes no output to be produced
4765 its main use is when combined with -strparse and -out to extract data
4766 from a file (which may not be in ASN.1 format).
4769 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4770 when producing the local key id.
4771 [Richard Levitte <levitte@stacken.kth.se>]
4773 *) New option -dhparam in s_server. This allows a DH parameter file to be
4774 stated explicitly. If it is not stated then it tries the first server
4775 certificate file. The previous behaviour hard coded the filename
4779 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4780 a public key to be input or output. For example:
4781 openssl rsa -in key.pem -pubout -out pubkey.pem
4782 Also added necessary DSA public key functions to handle this.
4785 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4786 in the message. This was handled by allowing
4787 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4788 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4790 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4791 to the end of the strings whereas this didn't. This would cause problems
4792 if strings read with d2i_ASN1_bytes() were later modified.
4793 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4795 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4796 data and it contains EOF it will end up returning an error. This is
4797 caused by input 46 bytes long. The cause is due to the way base64
4798 BIOs find the start of base64 encoded data. They do this by trying a
4799 trial decode on each line until they find one that works. When they
4800 do a flag is set and it starts again knowing it can pass all the
4801 data directly through the decoder. Unfortunately it doesn't reset
4802 the context it uses. This means that if EOF is reached an attempt
4803 is made to pass two EOFs through the context and this causes the
4804 resulting error. This can also cause other problems as well. As is
4805 usual with these problems it takes *ages* to find and the fix is
4806 trivial: move one line.
4807 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4809 *) Ugly workaround to get s_client and s_server working under Windows. The
4810 old code wouldn't work because it needed to select() on sockets and the
4811 tty (for keypresses and to see if data could be written). Win32 only
4812 supports select() on sockets so we select() with a 1s timeout on the
4813 sockets and then see if any characters are waiting to be read, if none
4814 are present then we retry, we also assume we can always write data to
4815 the tty. This isn't nice because the code then blocks until we've
4816 received a complete line of data and it is effectively polling the
4817 keyboard at 1s intervals: however it's quite a bit better than not
4818 working at all :-) A dedicated Windows application might handle this
4819 with an event loop for example.
4822 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4823 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4824 will be called when RSA_sign() and RSA_verify() are used. This is useful
4825 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4826 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4827 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4828 This necessitated the support of an extra signature type NID_md5_sha1
4829 for SSL signatures and modifications to the SSL library to use it instead
4830 of calling RSA_public_decrypt() and RSA_private_encrypt().
4833 *) Add new -verify -CAfile and -CApath options to the crl program, these
4834 will lookup a CRL issuers certificate and verify the signature in a
4835 similar way to the verify program. Tidy up the crl program so it
4836 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4837 less strict. It will now permit CRL extensions even if it is not
4838 a V2 CRL: this will allow it to tolerate some broken CRLs.
4841 *) Initialize all non-automatic variables each time one of the openssl
4842 sub-programs is started (this is necessary as they may be started
4843 multiple times from the "OpenSSL>" prompt).
4844 [Lennart Bang, Bodo Moeller]
4846 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4847 removing all other RSA functionality (this is what NO_RSA does). This
4848 is so (for example) those in the US can disable those operations covered
4849 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4853 *) Non-copying interface to BIO pairs.
4854 (still largely untested)
4857 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4858 ASCII string. This was handled independently in various places before.
4861 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4862 UTF8 strings a character at a time.
4865 *) Use client_version from client hello to select the protocol
4866 (s23_srvr.c) and for RSA client key exchange verification
4867 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4870 *) Add various utility functions to handle SPKACs, these were previously
4871 handled by poking round in the structure internals. Added new function
4872 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4873 print, verify and generate SPKACs. Based on an original idea from
4874 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4877 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4880 *) Allow the config file extension section to be overwritten on the
4881 command line. Based on an original idea from Massimiliano Pala
4882 <madwolf@comune.modena.it>. The new option is called -extensions
4883 and can be applied to ca, req and x509. Also -reqexts to override
4884 the request extensions in req and -crlexts to override the crl extensions
4888 *) Add new feature to the SPKAC handling in ca. Now you can include
4889 the same field multiple times by preceding it by "XXXX." for example:
4892 this is the same syntax as used in the req config file.
4895 *) Allow certificate extensions to be added to certificate requests. These
4896 are specified in a 'req_extensions' option of the req section of the
4897 config file. They can be printed out with the -text option to req but
4898 are otherwise ignored at present.
4901 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4902 data read consists of only the final block it would not decrypted because
4903 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4904 A misplaced 'break' also meant the decrypted final block might not be
4905 copied until the next read.
4908 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4909 a few extra parameters to the DH structure: these will be useful if
4910 for example we want the value of 'q' or implement X9.42 DH.
4913 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4914 provides hooks that allow the default DSA functions or functions on a
4915 "per key" basis to be replaced. This allows hardware acceleration and
4916 hardware key storage to be handled without major modification to the
4917 library. Also added low level modexp hooks and CRYPTO_EX structure and
4918 associated functions.
4921 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4922 as "read only": it can't be written to and the buffer it points to will
4923 not be freed. Reading from a read only BIO is much more efficient than
4924 a normal memory BIO. This was added because there are several times when
4925 an area of memory needs to be read from a BIO. The previous method was
4926 to create a memory BIO and write the data to it, this results in two
4927 copies of the data and an O(n^2) reading algorithm. There is a new
4928 function BIO_new_mem_buf() which creates a read only memory BIO from
4929 an area of memory. Also modified the PKCS#7 routines to use read only
4933 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4934 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4935 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4936 but a retry condition occured while trying to read the rest.
4939 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4940 NID_pkcs7_encrypted by default: this was wrong since this should almost
4941 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4942 the encrypted data type: this is a more sensible place to put it and it
4943 allows the PKCS#12 code to be tidied up that duplicated this
4947 *) Changed obj_dat.pl script so it takes its input and output files on
4948 the command line. This should avoid shell escape redirection problems
4952 *) Initial support for certificate extension requests, these are included
4953 in things like Xenroll certificate requests. Included functions to allow
4954 extensions to be obtained and added.
4957 *) -crlf option to s_client and s_server for sending newlines as
4958 CRLF (as required by many protocols).
4961 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4963 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4964 [Ralf S. Engelschall]
4966 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4967 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4969 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4973 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4974 DH parameters/keys (q is lost during that conversion, but the resulting
4975 DH parameters contain its length).
4977 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4978 much faster than DH_generate_parameters (which creates parameters
4979 where p = 2*q + 1), and also the smaller q makes DH computations
4980 much more efficient (160-bit exponentiation instead of 1024-bit
4981 exponentiation); so this provides a convenient way to support DHE
4982 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4983 utter importance to use
4984 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4986 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4987 when such DH parameters are used, because otherwise small subgroup
4988 attacks may become possible!
4991 *) Avoid memory leak in i2d_DHparams.
4994 *) Allow the -k option to be used more than once in the enc program:
4995 this allows the same encrypted message to be read by multiple recipients.
4998 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4999 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5000 it will always use the numerical form of the OID, even if it has a short
5004 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5005 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5006 otherwise bn_mod_exp was called. In the case of hardware keys for example
5007 no private key components need be present and it might store extra data
5008 in the RSA structure, which cannot be accessed from bn_mod_exp.
5009 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5010 private key operations.
5013 *) Added support for SPARC Linux.
5016 *) pem_password_cb function type incompatibly changed from
5017 typedef int pem_password_cb(char *buf, int size, int rwflag);
5019 ....(char *buf, int size, int rwflag, void *userdata);
5020 so that applications can pass data to their callbacks:
5021 The PEM[_ASN1]_{read,write}... functions and macros now take an
5022 additional void * argument, which is just handed through whenever
5023 the password callback is called.
5024 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5026 New function SSL_CTX_set_default_passwd_cb_userdata.
5028 Compatibility note: As many C implementations push function arguments
5029 onto the stack in reverse order, the new library version is likely to
5030 interoperate with programs that have been compiled with the old
5031 pem_password_cb definition (PEM_whatever takes some data that
5032 happens to be on the stack as its last argument, and the callback
5033 just ignores this garbage); but there is no guarantee whatsoever that
5036 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5037 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5038 problems not only on Windows, but also on some Unix platforms.
5039 To avoid problematic command lines, these definitions are now in an
5040 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5041 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5044 *) MIPS III/IV assembler module is reimplemented.
5047 *) More DES library cleanups: remove references to srand/rand and
5048 delete an unused file.
5051 *) Add support for the the free Netwide assembler (NASM) under Win32,
5052 since not many people have MASM (ml) and it can be hard to obtain.
5053 This is currently experimental but it seems to work OK and pass all
5054 the tests. Check out INSTALL.W32 for info.
5057 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5058 without temporary keys kept an extra copy of the server key,
5059 and connections with temporary keys did not free everything in case
5063 *) New function RSA_check_key and new openssl rsa option -check
5064 for verifying the consistency of RSA keys.
5065 [Ulf Moeller, Bodo Moeller]
5067 *) Various changes to make Win32 compile work:
5068 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5069 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5070 comparison" warnings.
5071 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5074 *) Add a debugging option to PKCS#5 v2 key generation function: when
5075 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5076 derived keys are printed to stderr.
5079 *) Copy the flags in ASN1_STRING_dup().
5080 [Roman E. Pavlov <pre@mo.msk.ru>]
5082 *) The x509 application mishandled signing requests containing DSA
5083 keys when the signing key was also DSA and the parameters didn't match.
5085 It was supposed to omit the parameters when they matched the signing key:
5086 the verifying software was then supposed to automatically use the CA's
5087 parameters if they were absent from the end user certificate.
5089 Omitting parameters is no longer recommended. The test was also
5090 the wrong way round! This was probably due to unusual behaviour in
5091 EVP_cmp_parameters() which returns 1 if the parameters match.
5092 This meant that parameters were omitted when they *didn't* match and
5093 the certificate was useless. Certificates signed with 'ca' didn't have
5095 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5097 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5098 The interface is as follows:
5099 Applications can use
5100 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5101 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5102 "off" is now the default.
5103 The library internally uses
5104 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5105 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5106 to disable memory-checking temporarily.
5108 Some inconsistent states that previously were possible (and were
5109 even the default) are now avoided.
5111 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5112 with each memory chunk allocated; this is occasionally more helpful
5113 than just having a counter.
5115 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5117 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5121 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5122 which largely parallels "options", but is for changing API behaviour,
5123 whereas "options" are about protocol behaviour.
5124 Initial "mode" flags are:
5126 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5127 a single record has been written.
5128 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5129 retries use the same buffer location.
5130 (But all of the contents must be
5134 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5137 *) Fix problems with no-hmac etc.
5138 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5140 *) New functions RSA_get_default_method(), RSA_set_method() and
5141 RSA_get_method(). These allows replacement of RSA_METHODs without having
5142 to mess around with the internals of an RSA structure.
5145 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5146 Also really enable memory leak checks in openssl.c and in some
5148 [Chad C. Mulligan, Bodo Moeller]
5150 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5151 up the length of negative integers. This has now been simplified to just
5152 store the length when it is first determined and use it later, rather
5153 than trying to keep track of where data is copied and updating it to
5155 [Steve Henson, reported by Brien Wheeler
5156 <bwheeler@authentica-security.com>]
5158 *) Add a new function PKCS7_signatureVerify. This allows the verification
5159 of a PKCS#7 signature but with the signing certificate passed to the
5160 function itself. This contrasts with PKCS7_dataVerify which assumes the
5161 certificate is present in the PKCS#7 structure. This isn't always the
5162 case: certificates can be omitted from a PKCS#7 structure and be
5163 distributed by "out of band" means (such as a certificate database).
5166 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5167 function prototypes in pem.h, also change util/mkdef.pl to add the
5168 necessary function names.
5171 *) mk1mf.pl (used by Windows builds) did not properly read the
5172 options set by Configure in the top level Makefile, and Configure
5173 was not even able to write more than one option correctly.
5174 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5177 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5178 file to be loaded from a BIO or FILE pointer. The BIO version will
5179 for example allow memory BIOs to contain config info.
5182 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5183 Whoever hopes to achieve shared-library compatibility across versions
5184 must use this, not the compile-time macro.
5185 (Exercise 0.9.4: Which is the minimum library version required by
5187 Note: All this applies only to multi-threaded programs, others don't
5191 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5192 through a BIO pair triggered the default case, i.e.
5193 SSLerr(...,SSL_R_UNKNOWN_STATE).
5196 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5197 can use the SSL library even if none of the specific BIOs is
5201 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5202 for the encoded length.
5203 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5205 *) Add initial documentation of the X509V3 functions.
5208 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5209 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5210 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5211 secure PKCS#8 private key format with a high iteration count.
5214 *) Fix determination of Perl interpreter: A perl or perl5
5215 _directory_ in $PATH was also accepted as the interpreter.
5216 [Ralf S. Engelschall]
5218 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5219 wrong with it but it was very old and did things like calling
5220 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5224 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5225 to use the new extension code.
5228 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5229 with macros. This should make it easier to change their form, add extra
5230 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5234 *) Add to configuration table a new entry that can specify an alternative
5235 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5236 according to Mark Crispin <MRC@Panda.COM>.
5240 *) DES CBC did not update the IV. Weird.
5243 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5244 Changing the behaviour of the former might break existing programs --
5245 where IV updating is needed, des_ncbc_encrypt can be used.
5248 *) When bntest is run from "make test" it drives bc to check its
5249 calculations, as well as internally checking them. If an internal check
5250 fails, it needs to cause bc to give a non-zero result or make test carries
5251 on without noticing the failure. Fixed.
5254 *) DES library cleanups.
5257 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5258 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5259 ciphers. NOTE: although the key derivation function has been verified
5260 against some published test vectors it has not been extensively tested
5261 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5265 *) Instead of "mkdir -p", which is not fully portable, use new
5266 Perl script "util/mkdir-p.pl".
5269 *) Rewrite the way password based encryption (PBE) is handled. It used to
5270 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5271 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5272 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5273 the 'parameter' field of the AlgorithmIdentifier is passed to the
5274 underlying key generation function so it must do its own ASN1 parsing.
5275 This has also changed the EVP_PBE_CipherInit() function which now has a
5276 'parameter' argument instead of literal salt and iteration count values
5277 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5280 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5281 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5282 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5283 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5284 value was just used as a "magic string" and not used directly its
5285 value doesn't matter.
5288 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5292 *) "linux-sparc64" configuration (ultrapenguin).
5293 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5294 "linux-sparc" configuration.
5295 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5297 *) config now generates no-xxx options for missing ciphers.
5300 *) Support the EBCDIC character set (work in progress).
5301 File ebcdic.c not yet included because it has a different license.
5302 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5304 *) Support BS2000/OSD-POSIX.
5305 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5307 *) Make callbacks for key generation use void * instead of char *.
5310 *) Make S/MIME samples compile (not yet tested).
5313 *) Additional typesafe stacks.
5316 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5320 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5322 *) New configuration variant "sco5-gcc".
5324 *) Updated some demos.
5325 [Sean O Riordain, Wade Scholine]
5327 *) Add missing BIO_free at exit of pkcs12 application.
5330 *) Fix memory leak in conf.c.
5333 *) Updates for Win32 to assembler version of MD5.
5336 *) Set #! path to perl in apps/der_chop to where we found it
5337 instead of using a fixed path.
5340 *) SHA library changes for irix64-mips4-cc.
5343 *) Improvements for VMS support.
5347 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5349 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5350 This also avoids the problems with SC4.2 and unpatched SC5.
5351 [Andy Polyakov <appro@fy.chalmers.se>]
5353 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5354 These are required because of the typesafe stack would otherwise break
5355 existing code. If old code used a structure member which used to be STACK
5356 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5357 sk_num or sk_value it would produce an error because the num, data members
5358 are not present in STACK_OF. Now it just produces a warning. sk_set
5359 replaces the old method of assigning a value to sk_value
5360 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5361 that does this will no longer work (and should use sk_set instead) but
5362 this could be regarded as a "questionable" behaviour anyway.
5365 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5366 correctly handle encrypted S/MIME data.
5369 *) Change type of various DES function arguments from des_cblock
5370 (which means, in function argument declarations, pointer to char)
5371 to des_cblock * (meaning pointer to array with 8 char elements),
5372 which allows the compiler to do more typechecking; it was like
5373 that back in SSLeay, but with lots of ugly casts.
5375 Introduce new type const_des_cblock.
5378 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5379 problems: find RecipientInfo structure that matches recipient certificate
5380 and initialise the ASN1 structures properly based on passed cipher.
5383 *) Belatedly make the BN tests actually check the results.
5386 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5387 to and from BNs: it was completely broken. New compilation option
5388 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5389 key elements as negative integers.
5392 *) Reorganize and speed up MD5.
5393 [Andy Polyakov <appro@fy.chalmers.se>]
5396 [Richard Levitte <richard@levitte.org>]
5398 *) New option -out to asn1parse to allow the parsed structure to be
5399 output to a file. This is most useful when combined with the -strparse
5400 option to examine the output of things like OCTET STRINGS.
5403 *) Make SSL library a little more fool-proof by not requiring any longer
5404 that SSL_set_{accept,connect}_state be called before
5405 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5406 in many applications because usually everything *appeared* to work as
5407 intended anyway -- now it really works as intended).
5410 *) Move openssl.cnf out of lib/.
5413 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5414 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5415 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5416 [Ralf S. Engelschall]
5418 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5419 handle PKCS#7 enveloped data properly.
5420 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5422 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5423 copying pointers. The cert_st handling is changed by this in
5424 various ways (and thus what used to be known as ctx->default_cert
5425 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5426 any longer when s->cert does not give us what we need).
5427 ssl_cert_instantiate becomes obsolete by this change.
5428 As soon as we've got the new code right (possibly it already is?),
5429 we have solved a couple of bugs of the earlier code where s->cert
5430 was used as if it could not have been shared with other SSL structures.
5432 Note that using the SSL API in certain dirty ways now will result
5433 in different behaviour than observed with earlier library versions:
5434 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5435 does not influence s as it used to.
5437 In order to clean up things more thoroughly, inside SSL_SESSION
5438 we don't use CERT any longer, but a new structure SESS_CERT
5439 that holds per-session data (if available); currently, this is
5440 the peer's certificate chain and, for clients, the server's certificate
5441 and temporary key. CERT holds only those values that can have
5442 meaningful defaults in an SSL_CTX.
5445 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5446 from the internal representation. Various PKCS#7 fixes: remove some
5447 evil casts and set the enc_dig_alg field properly based on the signing
5451 *) Allow PKCS#12 password to be set from the command line or the
5452 environment. Let 'ca' get its config file name from the environment
5453 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5457 *) Allow certificate policies extension to use an IA5STRING for the
5458 organization field. This is contrary to the PKIX definition but
5459 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5463 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5464 without disallowing inline assembler and the like for non-pedantic builds.
5467 *) Support Borland C++ builder.
5468 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5473 *) SHA-1 cleanups and performance enhancements.
5474 [Andy Polyakov <appro@fy.chalmers.se>]
5476 *) Sparc v8plus assembler for the bignum library.
5477 [Andy Polyakov <appro@fy.chalmers.se>]
5479 *) Accept any -xxx and +xxx compiler options in Configure.
5482 *) Update HPUX configuration.
5485 *) Add missing sk_<type>_unshift() function to safestack.h
5486 [Ralf S. Engelschall]
5488 *) New function SSL_CTX_use_certificate_chain_file that sets the
5489 "extra_cert"s in addition to the certificate. (This makes sense
5490 only for "PEM" format files, as chains as a whole are not
5494 *) Support verify_depth from the SSL API.
5495 x509_vfy.c had what can be considered an off-by-one-error:
5496 Its depth (which was not part of the external interface)
5497 was actually counting the number of certificates in a chain;
5498 now it really counts the depth.
5501 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5502 instead of X509err, which often resulted in confusing error
5503 messages since the error codes are not globally unique
5504 (e.g. an alleged error in ssl3_accept when a certificate
5505 didn't match the private key).
5507 *) New function SSL_CTX_set_session_id_context that allows to set a default
5508 value (so that you don't need SSL_set_session_id_context for each
5509 connection using the SSL_CTX).
5512 *) OAEP decoding bug fix.
5515 *) Support INSTALL_PREFIX for package builders, as proposed by
5519 *) New Configure options "threads" and "no-threads". For systems
5520 where the proper compiler options are known (currently Solaris
5521 and Linux), "threads" is the default.
5524 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5527 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5528 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5529 such as /usr/local/bin.
5532 *) "make linux-shared" to build shared libraries.
5533 [Niels Poppe <niels@netbox.org>]
5535 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5538 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5539 extension adding in x509 utility.
5542 *) Remove NOPROTO sections and error code comments.
5545 *) Partial rewrite of the DEF file generator to now parse the ANSI
5549 *) New Configure options --prefix=DIR and --openssldir=DIR.
5552 *) Complete rewrite of the error code script(s). It is all now handled
5553 by one script at the top level which handles error code gathering,
5554 header rewriting and C source file generation. It should be much better
5555 than the old method: it now uses a modified version of Ulf's parser to
5556 read the ANSI prototypes in all header files (thus the old K&R definitions
5557 aren't needed for error creation any more) and do a better job of
5558 translating function codes into names. The old 'ASN1 error code imbedded
5559 in a comment' is no longer necessary and it doesn't use .err files which
5560 have now been deleted. Also the error code call doesn't have to appear all
5561 on one line (which resulted in some large lines...).
5564 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5567 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5568 0 (which usually indicates a closed connection), but continue reading.
5571 *) Fix some race conditions.
5574 *) Add support for CRL distribution points extension. Add Certificate
5575 Policies and CRL distribution points documentation.
5578 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5581 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5582 8 of keying material. Merlin has also confirmed interop with this fix
5583 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5584 [Merlin Hughes <merlin@baltimore.ie>]
5586 *) Fix lots of warnings.
5587 [Richard Levitte <levitte@stacken.kth.se>]
5589 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5590 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5591 [Richard Levitte <levitte@stacken.kth.se>]
5593 *) Fix problems with sizeof(long) == 8.
5594 [Andy Polyakov <appro@fy.chalmers.se>]
5596 *) Change functions to ANSI C.
5599 *) Fix typos in error codes.
5600 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5602 *) Remove defunct assembler files from Configure.
5605 *) SPARC v8 assembler BIGNUM implementation.
5606 [Andy Polyakov <appro@fy.chalmers.se>]
5608 *) Support for Certificate Policies extension: both print and set.
5609 Various additions to support the r2i method this uses.
5612 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5613 return a const string when you are expecting an allocated buffer.
5616 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5617 types DirectoryString and DisplayText.
5620 *) Add code to allow r2i extensions to access the configuration database,
5621 add an LHASH database driver and add several ctx helper functions.
5624 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5625 fail when they extended the size of a BIGNUM.
5628 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5629 support typesafe stack.
5632 *) Fix typo in SSL_[gs]et_options().
5633 [Nils Frostberg <nils@medcom.se>]
5635 *) Delete various functions and files that belonged to the (now obsolete)
5636 old X509V3 handling code.
5639 *) New Configure option "rsaref".
5642 *) Don't auto-generate pem.h.
5645 *) Introduce type-safe ASN.1 SETs.
5648 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5649 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5651 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5652 that links with OpenSSL (well at least cause lots of warnings), but fear
5653 not: the conversion is trivial, and it eliminates loads of evil casts. A
5654 few STACKed things have been converted already. Feel free to convert more.
5655 In the fullness of time, I'll do away with the STACK type altogether.
5658 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5659 specified in <certfile> by updating the entry in the index.txt file.
5660 This way one no longer has to edit the index.txt file manually for
5661 revoking a certificate. The -revoke option does the gory details now.
5662 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5664 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5665 `-text' option at all and this way the `-noout -text' combination was
5666 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5667 [Ralf S. Engelschall]
5669 *) Make sure a corresponding plain text error message exists for the
5670 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5671 verify callback function determined that a certificate was revoked.
5672 [Ralf S. Engelschall]
5674 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5675 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5676 all available cipers including rc5, which was forgotten until now.
5677 In order to let the testing shell script know which algorithms
5678 are available, a new (up to now undocumented) command
5679 "openssl list-cipher-commands" is used.
5682 *) Bugfix: s_client occasionally would sleep in select() when
5683 it should have checked SSL_pending() first.
5686 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5687 the raw DSA values prior to ASN.1 encoding.
5690 *) Tweaks to Configure
5691 [Niels Poppe <niels@netbox.org>]
5693 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5697 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5700 *) New config option to avoid instructions that are illegal on the 80386.
5701 The default code is faster, but requires at least a 486.
5704 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5705 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5706 same as SSL2_VERSION anyway.
5709 *) New "-showcerts" option for s_client.
5712 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5713 application. Various cleanups and fixes.
5716 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5717 modify error routines to work internally. Add error codes and PBE init
5718 to library startup routines.
5721 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5722 packing functions to asn1 and evp. Changed function names and error
5723 codes along the way.
5726 *) PKCS12 integration: and so it begins... First of several patches to
5727 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5728 objects to objects.h
5731 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5732 and display support for Thawte strong extranet extension.
5735 *) Add LinuxPPC support.
5736 [Jeff Dubrule <igor@pobox.org>]
5738 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5739 bn_div_words in alpha.s.
5740 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5742 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5743 OAEP isn't supported when OpenSSL is built with RSAref.
5744 [Ulf Moeller <ulf@fitug.de>]
5746 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5747 so they no longer are missing under -DNOPROTO.
5748 [Soren S. Jorvang <soren@t.dk>]
5751 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5753 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5754 doesn't work when the session is reused. Coming soon!
5757 *) Fix a security hole, that allows sessions to be reused in the wrong
5758 context thus bypassing client cert protection! All software that uses
5759 client certs and session caches in multiple contexts NEEDS PATCHING to
5760 allow session reuse! A fuller solution is in the works.
5761 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5763 *) Some more source tree cleanups (removed obsolete files
5764 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5765 permission on "config" script to be executable) and a fix for the INSTALL
5767 [Ulf Moeller <ulf@fitug.de>]
5769 *) Remove some legacy and erroneous uses of malloc, free instead of
5771 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5773 *) Make rsa_oaep_test return non-zero on error.
5774 [Ulf Moeller <ulf@fitug.de>]
5776 *) Add support for native Solaris shared libraries. Configure
5777 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5778 if someone would make that last step automatic.
5779 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5781 *) ctx_size was not built with the right compiler during "make links". Fixed.
5784 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5785 except NULL ciphers". This means the default cipher list will no longer
5786 enable NULL ciphers. They need to be specifically enabled e.g. with
5787 the string "DEFAULT:eNULL".
5790 *) Fix to RSA private encryption routines: if p < q then it would
5791 occasionally produce an invalid result. This will only happen with
5792 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5795 *) Be less restrictive and allow also `perl util/perlpath.pl
5796 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5797 because this way one can also use an interpreter named `perl5' (which is
5798 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5799 installed as `perl').
5800 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5802 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5803 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5805 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5806 advapi32.lib to Win32 build and change the pem test comparision
5807 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5808 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5809 and crypto/des/ede_cbcm_enc.c.
5812 *) DES quad checksum was broken on big-endian architectures. Fixed.
5815 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5816 Win32 test batch file so it (might) work again. The Win32 test batch file
5817 is horrible: I feel ill....
5820 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5821 in e_os.h. Audit of header files to check ANSI and non ANSI
5822 sections: 10 functions were absent from non ANSI section and not exported
5823 from Windows DLLs. Fixed up libeay.num for new functions.
5826 *) Make `openssl version' output lines consistent.
5827 [Ralf S. Engelschall]
5829 *) Fix Win32 symbol export lists for BIO functions: Added
5830 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5831 to ms/libeay{16,32}.def.
5832 [Ralf S. Engelschall]
5834 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5835 fine under Unix and passes some trivial tests I've now added. But the
5836 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5837 added to make sure no one expects that this stuff really works in the
5838 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5839 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5841 [Ralf S. Engelschall]
5843 *) Fix the generation of two part addresses in perl.
5844 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5846 *) Add config entry for Linux on MIPS.
5847 [John Tobey <jtobey@channel1.com>]
5849 *) Make links whenever Configure is run, unless we are on Windoze.
5852 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5853 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5857 *) Add a useful kludge to allow package maintainers to specify compiler and
5858 other platforms details on the command line without having to patch the
5859 Configure script everytime: One now can use ``perl Configure
5860 <id>:<details>'', i.e. platform ids are allowed to have details appended
5861 to them (seperated by colons). This is treated as there would be a static
5862 pre-configured entry in Configure's %table under key <id> with value
5863 <details> and ``perl Configure <id>'' is called. So, when you want to
5864 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5865 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5866 now, which overrides the FreeBSD-elf entry on-the-fly.
5867 [Ralf S. Engelschall]
5869 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5872 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5873 on the `perl Configure ...' command line. This way one can compile
5874 OpenSSL libraries with Position Independent Code (PIC) which is needed
5875 for linking it into DSOs.
5876 [Ralf S. Engelschall]
5878 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5882 *) Cleaned up the LICENSE document: The official contact for any license
5883 questions now is the OpenSSL core team under openssl-core@openssl.org.
5884 And add a paragraph about the dual-license situation to make sure people
5885 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5886 to the OpenSSL toolkit.
5887 [Ralf S. Engelschall]
5889 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5890 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5891 Additonally cleaned up the `make links' target: Remove unnecessary
5892 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5893 to speed processing and no longer clutter the display with confusing
5894 stuff. Instead only the actually done links are displayed.
5895 [Ralf S. Engelschall]
5897 *) Permit null encryption ciphersuites, used for authentication only. It used
5898 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5899 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5903 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5904 signed attributes when verifying signatures (this would break them),
5905 the detached data encoding was wrong and public keys obtained using
5906 X509_get_pubkey() weren't freed.
5909 *) Add text documentation for the BUFFER functions. Also added a work around
5910 to a Win95 console bug. This was triggered by the password read stuff: the
5911 last character typed gets carried over to the next fread(). If you were
5912 generating a new cert request using 'req' for example then the last
5913 character of the passphrase would be CR which would then enter the first
5917 *) Added the new `Includes OpenSSL Cryptography Software' button as
5918 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5919 button and can be used by applications based on OpenSSL to show the
5920 relationship to the OpenSSL project.
5921 [Ralf S. Engelschall]
5923 *) Remove confusing variables in function signatures in files
5924 ssl/ssl_lib.c and ssl/ssl.h.
5925 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5927 *) Don't install bss_file.c under PREFIX/include/
5928 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5930 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5931 functions that return function pointers and has support for NT specific
5932 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5933 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5934 unsigned to signed types: this was killing the Win32 compile.
5937 *) Add new certificate file to stack functions,
5938 SSL_add_dir_cert_subjects_to_stack() and
5939 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5940 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5941 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5942 This means that Apache-SSL and similar packages don't have to mess around
5943 to add as many CAs as they want to the preferred list.
5946 *) Experiment with doxygen documentation. Currently only partially applied to
5948 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5949 openssl.doxy as the configuration file.
5952 *) Get rid of remaining C++-style comments which strict C compilers hate.
5953 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5955 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5956 compiled in by default: it has problems with large keys.
5959 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5960 DH private keys and/or callback functions which directly correspond to
5961 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5962 is needed for applications which have to configure certificates on a
5963 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5965 For the RSA certificate situation is makes no difference, but
5966 for the DSA certificate situation this fixes the "no shared cipher"
5967 problem where the OpenSSL cipher selection procedure failed because the
5968 temporary keys were not overtaken from the context and the API provided
5969 no way to reconfigure them.
5970 The new functions now let applications reconfigure the stuff and they
5971 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5972 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5973 non-public-API function ssl_cert_instantiate() is used as a helper
5974 function and also to reduce code redundancy inside ssl_rsa.c.
5975 [Ralf S. Engelschall]
5977 *) Move s_server -dcert and -dkey options out of the undocumented feature
5978 area because they are useful for the DSA situation and should be
5979 recognized by the users.
5980 [Ralf S. Engelschall]
5982 *) Fix the cipher decision scheme for export ciphers: the export bits are
5983 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5984 SSL_EXP_MASK. So, the original variable has to be used instead of the
5985 already masked variable.
5986 [Richard Levitte <levitte@stacken.kth.se>]
5988 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5989 [Richard Levitte <levitte@stacken.kth.se>]
5991 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5992 from `int' to `unsigned int' because it's a length and initialized by
5993 EVP_DigestFinal() which expects an `unsigned int *'.
5994 [Richard Levitte <levitte@stacken.kth.se>]
5996 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5997 script. Instead use the usual Shell->Perl transition trick.
5998 [Ralf S. Engelschall]
6000 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6001 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6002 -noout -modulus' as it's already the case for `openssl rsa -noout
6003 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6004 currently the public key is printed (a decision which was already done by
6005 `openssl dsa -modulus' in the past) which serves a similar purpose.
6006 Additionally the NO_RSA no longer completely removes the whole -modulus
6007 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6009 [Ralf S. Engelschall]
6011 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6012 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6013 [Arne Ansper <arne@ats.cyber.ee>]
6015 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6016 to be added. Now both 'req' and 'ca' can use new objects defined in the
6020 *) Add cool BIO that does syslog (or event log on NT).
6021 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6023 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6024 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6025 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6026 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6029 *) Add preliminary config info for new extension code.
6032 *) Make RSA_NO_PADDING really use no padding.
6033 [Ulf Moeller <ulf@fitug.de>]
6035 *) Generate errors when private/public key check is done.
6038 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6039 for some CRL extensions and new objects added.
6042 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6043 key usage extension and fuller support for authority key id.
6046 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6047 padding method for RSA, which is recommended for new applications in PKCS
6048 #1 v2.0 (RFC 2437, October 1998).
6049 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6050 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6051 against Bleichbacher's attack on RSA.
6052 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6055 *) Updates to the new SSL compression code
6056 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6058 *) Fix so that the version number in the master secret, when passed
6059 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6060 (because the server will not accept higher), that the version number
6061 is 0x03,0x01, not 0x03,0x00
6062 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6064 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6065 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6066 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6069 *) Support for RAW extensions where an arbitrary extension can be
6070 created by including its DER encoding. See apps/openssl.cnf for
6074 *) Make sure latest Perl versions don't interpret some generated C array
6075 code as Perl array code in the crypto/err/err_genc.pl script.
6076 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6078 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6079 not many people have the assembler. Various Win32 compilation fixes and
6080 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6084 *) Modify configure script 'Configure' to automatically create crypto/date.h
6085 file under Win32 and also build pem.h from pem.org. New script
6086 util/mkfiles.pl to create the MINFO file on environments that can't do a
6087 'make files': perl util/mkfiles.pl >MINFO should work.
6090 *) Major rework of DES function declarations, in the pursuit of correctness
6091 and purity. As a result, many evil casts evaporated, and some weirdness,
6092 too. You may find this causes warnings in your code. Zapping your evil
6093 casts will probably fix them. Mostly.
6096 *) Fix for a typo in asn1.h. Bug fix to object creation script
6097 obj_dat.pl. It considered a zero in an object definition to mean
6098 "end of object": none of the objects in objects.h have any zeros
6099 so it wasn't spotted.
6100 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6102 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6103 Masking (CBCM). In the absence of test vectors, the best I have been able
6104 to do is check that the decrypt undoes the encrypt, so far. Send me test
6105 vectors if you have them.
6108 *) Correct calculation of key length for export ciphers (too much space was
6109 allocated for null ciphers). This has not been tested!
6112 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6113 message is now correct (it understands "crypto" and "ssl" on its
6114 command line). There is also now an "update" option. This will update
6115 the util/ssleay.num and util/libeay.num files with any new functions.
6117 perl util/mkdef.pl crypto ssl update
6118 it will update them.
6121 *) Overhauled the Perl interface (perl/*):
6122 - ported BN stuff to OpenSSL's different BN library
6123 - made the perl/ source tree CVS-aware
6124 - renamed the package from SSLeay to OpenSSL (the files still contain
6125 their history because I've copied them in the repository)
6126 - removed obsolete files (the test scripts will be replaced
6127 by better Test::Harness variants in the future)
6128 [Ralf S. Engelschall]
6130 *) First cut for a very conservative source tree cleanup:
6131 1. merge various obsolete readme texts into doc/ssleay.txt
6132 where we collect the old documents and readme texts.
6133 2. remove the first part of files where I'm already sure that we no
6134 longer need them because of three reasons: either they are just temporary
6135 files which were left by Eric or they are preserved original files where
6136 I've verified that the diff is also available in the CVS via "cvs diff
6137 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6138 the crypto/md/ stuff).
6139 [Ralf S. Engelschall]
6141 *) More extension code. Incomplete support for subject and issuer alt
6142 name, issuer and authority key id. Change the i2v function parameters
6143 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6144 what that's for :-) Fix to ASN1 macro which messed up
6145 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6148 *) Preliminary support for ENUMERATED type. This is largely copied from the
6152 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6153 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6155 *) Make sure `make rehash' target really finds the `openssl' program.
6156 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6158 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6159 like to hear about it if this slows down other processors.
6162 *) Add CygWin32 platform information to Configure script.
6163 [Alan Batie <batie@aahz.jf.intel.com>]
6165 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6166 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6168 *) New program nseq to manipulate netscape certificate sequences
6171 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6175 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6176 but the BN code had some problems that would cause failures when
6177 doing certificate verification and some other functions.
6178 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6180 *) Add ASN1 and PEM code to support netscape certificate sequences.
6183 *) Add ASN1 and PEM code to support netscape certificate sequences.
6186 *) Add several PKIX and private extended key usage OIDs.
6189 *) Modify the 'ca' program to handle the new extension code. Modify
6190 openssl.cnf for new extension format, add comments.
6193 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6194 and add a sample to openssl.cnf so req -x509 now adds appropriate
6198 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6199 error code, add initial support to X509_print() and x509 application.
6202 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6203 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6204 stuff is currently isolated and isn't even compiled yet.
6207 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6208 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6209 Removed the versions check from X509 routines when loading extensions:
6210 this allows certain broken certificates that don't set the version
6211 properly to be processed.
6214 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6215 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6216 can still be regenerated with "make depend".
6219 *) Spelling mistake in C version of CAST-128.
6220 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6222 *) Changes to the error generation code. The perl script err-code.pl
6223 now reads in the old error codes and retains the old numbers, only
6224 adding new ones if necessary. It also only changes the .err files if new
6225 codes are added. The makefiles have been modified to only insert errors
6226 when needed (to avoid needlessly modifying header files). This is done
6227 by only inserting errors if the .err file is newer than the auto generated
6228 C file. To rebuild all the error codes from scratch (the old behaviour)
6229 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6230 or delete all the .err files.
6233 *) CAST-128 was incorrectly implemented for short keys. The C version has
6234 been fixed, but is untested. The assembler versions are also fixed, but
6235 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6236 to regenerate it if needed.
6237 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6238 Hagino <itojun@kame.net>]
6240 *) File was opened incorrectly in randfile.c.
6241 [Ulf Möller <ulf@fitug.de>]
6243 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6244 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6245 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6246 al: it's just almost always a UTCTime. Note this patch adds new error
6247 codes so do a "make errors" if there are problems.
6250 *) Correct Linux 1 recognition in config.
6251 [Ulf Möller <ulf@fitug.de>]
6253 *) Remove pointless MD5 hash when using DSA keys in ca.
6254 [Anonymous <nobody@replay.com>]
6256 *) Generate an error if given an empty string as a cert directory. Also
6257 generate an error if handed NULL (previously returned 0 to indicate an
6258 error, but didn't set one).
6259 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6261 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6264 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6265 parameters. This was causing a warning which killed off the Win32 compile.
6268 *) Remove C++ style comments from crypto/bn/bn_local.h.
6269 [Neil Costigan <neil.costigan@celocom.com>]
6271 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6272 based on a text string, looking up short and long names and finally
6273 "dot" format. The "dot" format stuff didn't work. Added new function
6274 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6275 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6276 OID is not part of the table.
6279 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6280 X509_LOOKUP_by_alias().
6283 *) Sort openssl functions by name.
6286 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6287 encryption from sample DSA keys (in case anyone is interested the password
6291 *) Make _all_ *_free functions accept a NULL pointer.
6292 [Frans Heymans <fheymans@isaserver.be>]
6294 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6296 [Anonymous <nobody@replay.com>]
6298 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6299 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6301 *) Don't blow it for numeric -newkey arguments to apps/req.
6302 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6304 *) Temp key "for export" tests were wrong in s3_srvr.c.
6305 [Anonymous <nobody@replay.com>]
6307 *) Add prototype for temp key callback functions
6308 SSL_CTX_set_tmp_{rsa,dh}_callback().
6311 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6312 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6315 *) X509_name_add_entry() freed the wrong thing after an error.
6316 [Arne Ansper <arne@ats.cyber.ee>]
6318 *) rsa_eay.c would attempt to free a NULL context.
6319 [Arne Ansper <arne@ats.cyber.ee>]
6321 *) BIO_s_socket() had a broken should_retry() on Windoze.
6322 [Arne Ansper <arne@ats.cyber.ee>]
6324 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6325 [Arne Ansper <arne@ats.cyber.ee>]
6327 *) Make sure the already existing X509_STORE->depth variable is initialized
6328 in X509_STORE_new(), but document the fact that this variable is still
6329 unused in the certificate verification process.
6330 [Ralf S. Engelschall]
6332 *) Fix the various library and apps files to free up pkeys obtained from
6333 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6336 *) Fix reference counting in X509_PUBKEY_get(). This makes
6337 demos/maurice/example2.c work, amongst others, probably.
6338 [Steve Henson and Ben Laurie]
6340 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6341 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6342 are no longer created. This way we have a single and consistent command
6343 line interface `openssl <command>', similar to `cvs <command>'.
6344 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6346 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6347 BIT STRING wrapper always have zero unused bits.
6350 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6353 *) Make the top-level INSTALL documentation easier to understand.
6356 *) Makefiles updated to exit if an error occurs in a sub-directory
6357 make (including if user presses ^C) [Paul Sutton]
6359 *) Make Montgomery context stuff explicit in RSA data structure.
6362 *) Fix build order of pem and err to allow for generated pem.h.
6365 *) Fix renumbering bug in X509_NAME_delete_entry().
6368 *) Enhanced the err-ins.pl script so it makes the error library number
6369 global and can add a library name. This is needed for external ASN1 and
6370 other error libraries.
6373 *) Fixed sk_insert which never worked properly.
6376 *) Fix ASN1 macros so they can handle indefinite length construted
6377 EXPLICIT tags. Some non standard certificates use these: they can now
6381 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6382 into a single doc/ssleay.txt bundle. This way the information is still
6383 preserved but no longer messes up this directory. Now it's new room for
6384 the new set of documenation files.
6385 [Ralf S. Engelschall]
6387 *) SETs were incorrectly DER encoded. This was a major pain, because they
6388 shared code with SEQUENCEs, which aren't coded the same. This means that
6389 almost everything to do with SETs or SEQUENCEs has either changed name or
6390 number of arguments.
6391 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6393 *) Fix test data to work with the above.
6396 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6397 was already fixed by Eric for 0.9.1 it seems.
6398 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6400 *) Autodetect FreeBSD3.
6403 *) Fix various bugs in Configure. This affects the following platforms:
6407 unixware-2.0-pentium
6411 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6412 before they are needed.
6415 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6419 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6421 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6422 changed SSLeay to OpenSSL in version strings.
6423 [Ralf S. Engelschall]
6425 *) Some fixups to the top-level documents.
6428 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6429 because the symlink to include/ was missing.
6430 [Ralf S. Engelschall]
6432 *) Incorporated the popular no-RSA/DSA-only patches
6433 which allow to compile a RSA-free SSLeay.
6434 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6436 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6437 when "ssleay" is still not found.
6438 [Ralf S. Engelschall]
6440 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6441 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6443 *) Updated the README file.
6444 [Ralf S. Engelschall]
6446 *) Added various .cvsignore files in the CVS repository subdirs
6447 to make a "cvs update" really silent.
6448 [Ralf S. Engelschall]
6450 *) Recompiled the error-definition header files and added
6451 missing symbols to the Win32 linker tables.
6452 [Ralf S. Engelschall]
6454 *) Cleaned up the top-level documents;
6455 o new files: CHANGES and LICENSE
6456 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6457 o merged COPYRIGHT into LICENSE
6458 o removed obsolete TODO file
6459 o renamed MICROSOFT to INSTALL.W32
6460 [Ralf S. Engelschall]
6462 *) Removed dummy files from the 0.9.1b source tree:
6463 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6464 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6465 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6466 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6467 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6468 [Ralf S. Engelschall]
6470 *) Added various platform portability fixes.
6473 *) The Genesis of the OpenSSL rpject:
6474 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6475 Young and Tim J. Hudson created while they were working for C2Net until
6477 [The OpenSSL Project]
6480 Changes between 0.9.0b and 0.9.1b [not released]
6482 *) Updated a few CA certificates under certs/
6485 *) Changed some BIGNUM api stuff.
6488 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6489 DGUX x86, Linux Alpha, etc.
6492 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6493 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6497 *) Add -strparse option to asn1pars program which parses nested
6499 [Dr Stephen Henson <shenson@bigfoot.com>]
6501 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6504 *) DSA fix for "ca" program.
6507 *) Added "-genkey" option to "dsaparam" program.
6510 *) Added RIPE MD160 (rmd160) message digest.
6513 *) Added -a (all) option to "ssleay version" command.
6516 *) Added PLATFORM define which is the id given to Configure.
6519 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6522 *) Extended the ASN.1 parser routines.
6525 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6528 *) Added a BN_CTX to the BN library.
6531 *) Fixed the weak key values in DES library
6534 *) Changed API in EVP library for cipher aliases.
6537 *) Added support for RC2/64bit cipher.
6540 *) Converted the lhash library to the crypto/mem.c functions.
6543 *) Added more recognized ASN.1 object ids.
6546 *) Added more RSA padding checks for SSL/TLS.
6549 *) Added BIO proxy/filter functionality.
6552 *) Added extra_certs to SSL_CTX which can be used
6553 send extra CA certificates to the client in the CA cert chain sending
6554 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6557 *) Now Fortezza is denied in the authentication phase because
6558 this is key exchange mechanism is not supported by SSLeay at all.
6561 *) Additional PKCS1 checks.
6564 *) Support the string "TLSv1" for all TLS v1 ciphers.
6567 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6568 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6571 *) Fixed a few memory leaks.
6574 *) Fixed various code and comment typos.
6577 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6578 bytes sent in the client random.
6579 [Edward Bishop <ebishop@spyglass.com>]