5 Changes between 0.9.7d and 0.9.7f [XX xxx XXXX]
9 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11 *) Avoid a race condition when CRLs are checked in a multi threaded
12 environment. This would happen due to the reordering of the revoked
13 entries during signature checking and serial number lookup. Now the
14 encoding is cached and the serial number sort performed under a lock.
15 Add new STACK function sk_is_sorted().
18 *) Add Delta CRL to the extension code.
21 *) Various fixes to s3_pkt.c so alerts are sent properly.
22 [David Holmes <d.holmes@f5.com>]
24 *) Reduce the chances of duplicate issuer name and serial numbers (in
25 violation of RFC3280) using the OpenSSL certificate creation utilities.
26 This is done by creating a random 64 bit value for the initial serial
27 number when a serial number file is created or when a self signed
28 certificate is created using 'openssl req -x509'. The initial serial
29 number file is created using 'openssl x509 -next_serial' in CA.pl
30 rather than being initialized to 1.
33 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
35 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
36 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
37 [Joe Orton, Steve Henson]
39 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
41 [Joe Orton, Steve Henson]
43 *) Make it possible to have multiple active certificates with the same
44 subject in the CA index file. This is done only if the keyword
45 'unique_subject' is set to 'no' in the main CA section (default
46 if 'CA_default') of the configuration file. The value is saved
47 with the database itself in a separate index attribute file,
48 named like the index file with '.attr' appended to the name.
51 *) X509 verify fixes. Disable broken certificate workarounds when
52 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
53 keyUsage extension present. Don't accept CRLs with unhandled critical
54 extensions: since verify currently doesn't process CRL extensions this
55 rejects a CRL with *any* critical extensions. Add new verify error codes
59 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
60 A clarification of RFC2560 will require the use of OCTET STRINGs and
61 some implementations cannot handle the current raw format. Since OpenSSL
62 copies and compares OCSP nonces as opaque blobs without any attempt at
63 parsing them this should not create any compatibility issues.
66 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
67 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
68 this HMAC (and other) operations are several times slower than OpenSSL
72 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
73 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
75 *) Use the correct content when signing type "other".
78 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
80 *) Fix various bugs revealed by running the NISCC test suite:
82 Stop out of bounds reads in the ASN1 code when presented with
83 invalid tags (CAN-2003-0543 and CAN-2003-0544).
85 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
87 If verify callback ignores invalid public key errors don't try to check
88 certificate signature with the NULL public key.
92 *) New -ignore_err option in ocsp application to stop the server
93 exiting on the first error in a request.
96 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
97 if the server requested one: as stated in TLS 1.0 and SSL 3.0
101 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
102 extra data after the compression methods not only for TLS 1.0
103 but also for SSL 3.0 (as required by the specification).
104 [Bodo Moeller; problem pointed out by Matthias Loepfe]
106 *) Change X509_certificate_type() to mark the key as exported/exportable
107 when it's 512 *bits* long, not 512 bytes.
110 *) Change AES_cbc_encrypt() so it outputs exact multiple of
111 blocks during encryption.
114 *) Various fixes to base64 BIO and non blocking I/O. On write
115 flushes were not handled properly if the BIO retried. On read
116 data was not being buffered properly and had various logic bugs.
117 This also affects blocking I/O when the data being decoded is a
121 *) Various S/MIME bugfixes and compatibility changes:
122 output correct application/pkcs7 MIME type if
123 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
124 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
125 of files as .eml work). Correctly handle very long lines in MIME
129 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
131 *) Countermeasure against the Klima-Pokorny-Rosa extension of
132 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
133 a protocol version number mismatch like a decryption error
134 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
137 *) Turn on RSA blinding by default in the default implementation
138 to avoid a timing attack. Applications that don't want it can call
139 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
140 They would be ill-advised to do so in most cases.
141 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
143 *) Change RSA blinding code so that it works when the PRNG is not
144 seeded (in this case, the secret RSA exponent is abused as
145 an unpredictable seed -- if it is not unpredictable, there
146 is no point in blinding anyway). Make RSA blinding thread-safe
147 by remembering the creator's thread ID in rsa->blinding and
148 having all other threads use local one-time blinding factors
149 (this requires more computation than sharing rsa->blinding, but
150 avoids excessive locking; and if an RSA object is not shared
151 between threads, blinding will still be very fast).
154 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
155 ENGINE as defaults for all supported algorithms irrespective of
156 the 'flags' parameter. 'flags' is now honoured, so applications
157 should make sure they are passing it correctly.
160 *) Target "mingw" now allows native Windows code to be generated in
161 the Cygwin environment as well as with the MinGW compiler.
164 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
166 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
167 via timing by performing a MAC computation even if incorrrect
168 block cipher padding has been found. This is a countermeasure
169 against active attacks where the attacker has to distinguish
170 between bad padding and a MAC verification error. (CAN-2003-0078)
172 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
173 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
174 Martin Vuagnoux (EPFL, Ilion)]
176 *) Make the no-err option work as intended. The intention with no-err
177 is not to have the whole error stack handling routines removed from
178 libcrypto, it's only intended to remove all the function name and
179 reason texts, thereby removing some of the footprint that may not
180 be interesting if those errors aren't displayed anyway.
182 NOTE: it's still possible for any application or module to have it's
183 own set of error texts inserted. The routines are there, just not
184 used by default when no-err is given.
187 *) Add support for FreeBSD on IA64.
188 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
190 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
191 Kerberos function mit_des_cbc_cksum(). Before this change,
192 the value returned by DES_cbc_cksum() was like the one from
193 mit_des_cbc_cksum(), except the bytes were swapped.
194 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
196 *) Allow an application to disable the automatic SSL chain building.
197 Before this a rather primitive chain build was always performed in
198 ssl3_output_cert_chain(): an application had no way to send the
199 correct chain if the automatic operation produced an incorrect result.
201 Now the chain builder is disabled if either:
203 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
205 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
207 The reasoning behind this is that an application would not want the
208 auto chain building to take place if extra chain certificates are
209 present and it might also want a means of sending no additional
210 certificates (for example the chain has two certificates and the
214 *) Add the possibility to build without the ENGINE framework.
215 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
217 *) Under Win32 gmtime() can return NULL: check return value in
218 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
221 *) DSA routines: under certain error conditions uninitialized BN objects
222 could be freed. Solution: make sure initialization is performed early
223 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
224 Nils Larsch <nla@trustcenter.de> via PR#459)
227 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
228 checked on reconnect on the client side, therefore session resumption
229 could still fail with a "ssl session id is different" error. This
230 behaviour is masked when SSL_OP_ALL is used due to
231 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
232 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
236 *) IA-32 assembler support enhancements: unified ELF targets, support
237 for SCO/Caldera platforms, fix for Cygwin shared build.
240 *) Add support for FreeBSD on sparc64. As a consequence, support for
241 FreeBSD on non-x86 processors is separate from x86 processors on
242 the config script, much like the NetBSD support.
243 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
245 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
247 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
250 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
251 code (06) was taken as the first octet of the session ID and the last
252 octet was ignored consequently. As a result SSLv2 client side session
253 caching could not have worked due to the session ID mismatch between
255 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
259 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
260 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
264 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
265 seems that in spite of existing for more than a year, many application
266 author have done nothing to provide the necessary callbacks, which
267 means that this particular engine will not work properly anywhere.
268 This is a very unfortunate situation which forces us, in the name
269 of usability, to give the hw_ncipher.c a static lock, which is part
271 NOTE: This is for the 0.9.7 series ONLY. This hack will never
272 appear in 0.9.8 or later. We EXPECT application authors to have
273 dealt properly with this when 0.9.8 is released (unless we actually
274 make such changes in the libcrypto locking code that changes will
275 have to be made anyway).
278 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
279 octets have been read, EOF or an error occurs. Without this change
280 some truncated ASN1 structures will not produce an error.
283 *) Disable Heimdal support, since it hasn't been fully implemented.
284 Still give the possibility to force the use of Heimdal, but with
285 warnings and a request that patches get sent to openssl-dev.
288 *) Add the VC-CE target, introduce the WINCE sysname, and add
289 INSTALL.WCE and appropriate conditionals to make it build.
290 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
292 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
293 cygssl-x.y.z.dll, where x, y and z are the major, minor and
294 edit numbers of the version.
295 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
297 *) Introduce safe string copy and catenation functions
298 (BUF_strlcpy() and BUF_strlcat()).
299 [Ben Laurie (CHATS) and Richard Levitte]
301 *) Avoid using fixed-size buffers for one-line DNs.
304 *) Add BUF_MEM_grow_clean() to avoid information leakage when
305 resizing buffers containing secrets, and use where appropriate.
308 *) Avoid using fixed size buffers for configuration file location.
311 *) Avoid filename truncation for various CA files.
314 *) Use sizeof in preference to magic numbers.
317 *) Avoid filename truncation in cert requests.
320 *) Add assertions to check for (supposedly impossible) buffer
324 *) Don't cache truncated DNS entries in the local cache (this could
325 potentially lead to a spoofing attack).
328 *) Fix various buffers to be large enough for hex/decimal
329 representations in a platform independent manner.
332 *) Add CRYPTO_realloc_clean() to avoid information leakage when
333 resizing buffers containing secrets, and use where appropriate.
336 *) Add BIO_indent() to avoid much slightly worrying code to do
340 *) Convert sprintf()/BIO_puts() to BIO_printf().
343 *) buffer_gets() could terminate with the buffer only half
347 *) Add assertions to prevent user-supplied crypto functions from
348 overflowing internal buffers by having large block sizes, etc.
351 *) New OPENSSL_assert() macro (similar to assert(), but enabled
355 *) Eliminate unused copy of key in RC4.
358 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
361 *) Fix off-by-one error in EGD path.
364 *) If RANDFILE path is too long, ignore instead of truncating.
367 *) Eliminate unused and incorrectly sized X.509 structure
371 *) Eliminate unused and dangerous function knumber().
374 *) Eliminate unused and dangerous structure, KSSL_ERR.
377 *) Protect against overlong session ID context length in an encoded
378 session object. Since these are local, this does not appear to be
382 *) Change from security patch (see 0.9.6e below) that did not affect
383 the 0.9.6 release series:
385 Remote buffer overflow in SSL3 protocol - an attacker could
386 supply an oversized master key in Kerberos-enabled versions.
390 *) Change the SSL kerb5 codes to match RFC 2712.
393 *) Make -nameopt work fully for req and add -reqopt switch.
394 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
396 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
397 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
399 *) Make sure tests can be performed even if the corresponding algorithms
400 have been removed entirely. This was also the last step to make
401 OpenSSL compilable with DJGPP under all reasonable conditions.
402 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
404 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
405 to allow version independent disabling of normally unselected ciphers,
406 which may be activated as a side-effect of selecting a single cipher.
408 (E.g., cipher list string "RSA" enables ciphersuites that are left
409 out of "ALL" because they do not provide symmetric encryption.
410 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
411 [Lutz Jaenicke, Bodo Moeller]
413 *) Add appropriate support for separate platform-dependent build
414 directories. The recommended way to make a platform-dependent
415 build directory is the following (tested on Linux), maybe with
418 # Place yourself outside of the OpenSSL source tree. In
419 # this example, the environment variable OPENSSL_SOURCE
420 # is assumed to contain the absolute OpenSSL source directory.
421 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
422 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
423 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
424 mkdir -p `dirname $F`
425 ln -s $OPENSSL_SOURCE/$F $F
428 To be absolutely sure not to disturb the source tree, a "make clean"
429 is a good thing. If it isn't successfull, don't worry about it,
430 it probably means the source directory is very clean.
433 *) Make sure any ENGINE control commands make local copies of string
434 pointers passed to them whenever necessary. Otherwise it is possible
435 the caller may have overwritten (or deallocated) the original string
436 data when a later ENGINE operation tries to use the stored values.
437 [Götz Babin-Ebell <babinebell@trustcenter.de>]
439 *) Improve diagnostics in file reading and command-line digests.
440 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
442 *) Add AES modes CFB and OFB to the object database. Correct an
443 error in AES-CFB decryption.
446 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
447 allows existing EVP_CIPHER_CTX structures to be reused after
448 calling EVP_*Final(). This behaviour is used by encryption
449 BIOs and some applications. This has the side effect that
450 applications must explicitly clean up cipher contexts with
451 EVP_CIPHER_CTX_cleanup() or they will leak memory.
454 *) Check the values of dna and dnb in bn_mul_recursive before calling
455 bn_mul_comba (a non zero value means the a or b arrays do not contain
456 n2 elements) and fallback to bn_mul_normal if either is not zero.
459 *) Fix escaping of non-ASCII characters when using the -subj option
460 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
463 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
464 form for "surname", serialNumber has no short form.
465 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
466 therefore remove "mail" short name for "internet 7".
467 The OID for unique identifiers in X509 certificates is
468 x500UniqueIdentifier, not uniqueIdentifier.
469 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
472 *) Add an "init" command to the ENGINE config module and auto initialize
473 ENGINEs. Without any "init" command the ENGINE will be initialized
474 after all ctrl commands have been executed on it. If init=1 the
475 ENGINE is initailized at that point (ctrls before that point are run
476 on the uninitialized ENGINE and after on the initialized one). If
477 init=0 then the ENGINE will not be iniatialized at all.
480 *) Fix the 'app_verify_callback' interface so that the user-defined
481 argument is actually passed to the callback: In the
482 SSL_CTX_set_cert_verify_callback() prototype, the callback
483 declaration has been changed from
486 int (*cb)(X509_STORE_CTX *,void *);
487 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
488 i=s->ctx->app_verify_callback(&ctx)
489 has been changed into
490 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
492 To update applications using SSL_CTX_set_cert_verify_callback(),
493 a dummy argument can be added to their callback functions.
494 [D. K. Smetters <smetters@parc.xerox.com>]
496 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
497 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
499 *) Add and OPENSSL_LOAD_CONF define which will cause
500 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
501 This allows older applications to transparently support certain
502 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
503 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
504 load the config file and OPENSSL_add_all_algorithms_conf() which will
505 always load it have also been added.
508 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
509 Adjust NIDs and EVP layer.
510 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
512 *) Config modules support in openssl utility.
514 Most commands now load modules from the config file,
515 though in a few (such as version) this isn't done
516 because it couldn't be used for anything.
518 In the case of ca and req the config file used is
519 the same as the utility itself: that is the -config
520 command line option can be used to specify an
524 *) Move default behaviour from OPENSSL_config(). If appname is NULL
525 use "openssl_conf" if filename is NULL use default openssl config file.
528 *) Add an argument to OPENSSL_config() to allow the use of an alternative
529 config section name. Add a new flag to tolerate a missing config file
530 and move code to CONF_modules_load_file().
533 *) Support for crypto accelerator cards from Accelerated Encryption
534 Processing, www.aep.ie. (Use engine 'aep')
535 The support was copied from 0.9.6c [engine] and adapted/corrected
536 to work with the new engine framework.
537 [AEP Inc. and Richard Levitte]
539 *) Support for SureWare crypto accelerator cards from Baltimore
540 Technologies. (Use engine 'sureware')
541 The support was copied from 0.9.6c [engine] and adapted
542 to work with the new engine framework.
545 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
546 make the newer ENGINE framework commands for the CHIL engine work.
547 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
549 *) Make it possible to produce shared libraries on ReliantUNIX.
550 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
552 *) Add the configuration target debug-linux-ppro.
553 Make 'openssl rsa' use the general key loading routines
554 implemented in apps.c, and make those routines able to
555 handle the key format FORMAT_NETSCAPE and the variant
557 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
559 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
560 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
562 *) Add -keyform to rsautl, and document -engine.
563 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
565 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
566 BIO_R_NO_SUCH_FILE error code rather than the generic
567 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
572 ERR_peek_last_error_line
573 ERR_peek_last_error_line_data.
577 ERR_peek_error_line_data,
578 but report on the latest error recorded rather than the first one
579 still in the error queue.
580 [Ben Laurie, Bodo Moeller]
582 *) default_algorithms option in ENGINE config module. This allows things
584 default_algorithms = ALL
585 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
588 *) Prelminary ENGINE config module.
591 *) New experimental application configuration code.
594 *) Change the AES code to follow the same name structure as all other
595 symmetric ciphers, and behave the same way. Move everything to
596 the directory crypto/aes, thereby obsoleting crypto/rijndael.
597 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
599 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
600 [Ben Laurie and Theo de Raadt]
602 *) Add option to output public keys in req command.
603 [Massimiliano Pala madwolf@openca.org]
605 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
606 (up to about 10% better than before for P-192 and P-224).
609 *) New functions/macros
611 SSL_CTX_set_msg_callback(ctx, cb)
612 SSL_CTX_set_msg_callback_arg(ctx, arg)
613 SSL_set_msg_callback(ssl, cb)
614 SSL_set_msg_callback_arg(ssl, arg)
616 to request calling a callback function
618 void cb(int write_p, int version, int content_type,
619 const void *buf, size_t len, SSL *ssl, void *arg)
621 whenever a protocol message has been completely received
622 (write_p == 0) or sent (write_p == 1). Here 'version' is the
623 protocol version according to which the SSL library interprets
624 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
625 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
626 the content type as defined in the SSL 3.0/TLS 1.0 protocol
627 specification (change_cipher_spec(20), alert(21), handshake(22)).
628 'buf' and 'len' point to the actual message, 'ssl' to the
629 SSL object, and 'arg' is the application-defined value set by
630 SSL[_CTX]_set_msg_callback_arg().
632 'openssl s_client' and 'openssl s_server' have new '-msg' options
633 to enable a callback that displays all protocol messages.
636 *) Change the shared library support so shared libraries are built as
637 soon as the corresponding static library is finished, and thereby get
638 openssl and the test programs linked against the shared library.
639 This still only happens when the keyword "shard" has been given to
640 the configuration scripts.
642 NOTE: shared library support is still an experimental thing, and
643 backward binary compatibility is still not guaranteed.
644 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
646 *) Add support for Subject Information Access extension.
647 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
649 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
650 additional bytes when new memory had to be allocated, not just
651 when reusing an existing buffer.
654 *) New command line and configuration option 'utf8' for the req command.
655 This allows field values to be specified as UTF8 strings.
658 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
659 runs for the former and machine-readable output for the latter.
662 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
663 of the e-mail address in the DN (i.e., it will go into a certificate
664 extension only). The new configuration file option 'email_in_dn = no'
666 [Massimiliano Pala madwolf@openca.org]
668 *) Change all functions with names starting with des_ to be starting
669 with DES_ instead. Add wrappers that are compatible with libdes,
670 but are named _ossl_old_des_*. Finally, add macros that map the
671 des_* symbols to the corresponding _ossl_old_des_* if libdes
672 compatibility is desired. If OpenSSL 0.9.6c compatibility is
673 desired, the des_* symbols will be mapped to DES_*, with one
676 Since we provide two compatibility mappings, the user needs to
677 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
678 compatibility is desired. The default (i.e., when that macro
679 isn't defined) is OpenSSL 0.9.6c compatibility.
681 There are also macros that enable and disable the support of old
682 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
683 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
684 are defined, the default will apply: to support the old des routines.
686 In either case, one must include openssl/des.h to get the correct
687 definitions. Do not try to just include openssl/des_old.h, that
690 NOTE: This is a major break of an old API into a new one. Software
691 authors are encouraged to switch to the DES_ style functions. Some
692 time in the future, des_old.h and the libdes compatibility functions
693 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
694 default), and then completely removed.
697 *) Test for certificates which contain unsupported critical extensions.
698 If such a certificate is found during a verify operation it is
699 rejected by default: this behaviour can be overridden by either
700 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
701 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
702 X509_supported_extension() has also been added which returns 1 if a
703 particular extension is supported.
706 *) Modify the behaviour of EVP cipher functions in similar way to digests
707 to retain compatibility with existing code.
710 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
711 compatibility with existing code. In particular the 'ctx' parameter does
712 not have to be to be initialized before the call to EVP_DigestInit() and
713 it is tidied up after a call to EVP_DigestFinal(). New function
714 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
715 EVP_MD_CTX_copy() changed to not require the destination to be
716 initialized valid and new function EVP_MD_CTX_copy_ex() added which
717 requires the destination to be valid.
719 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
720 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
723 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
724 so that complete 'Handshake' protocol structures are kept in memory
725 instead of overwriting 'msg_type' and 'length' with 'body' data.
728 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
729 [Massimo Santin via Richard Levitte]
731 *) Major restructuring to the underlying ENGINE code. This includes
732 reduction of linker bloat, separation of pure "ENGINE" manipulation
733 (initialisation, etc) from functionality dealing with implementations
734 of specific crypto iterfaces. This change also introduces integrated
735 support for symmetric ciphers and digest implementations - so ENGINEs
736 can now accelerate these by providing EVP_CIPHER and EVP_MD
737 implementations of their own. This is detailed in crypto/engine/README
738 as it couldn't be adequately described here. However, there are a few
739 API changes worth noting - some RSA, DSA, DH, and RAND functions that
740 were changed in the original introduction of ENGINE code have now
741 reverted back - the hooking from this code to ENGINE is now a good
742 deal more passive and at run-time, operations deal directly with
743 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
744 dereferencing through an ENGINE pointer any more. Also, the ENGINE
745 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
746 they were not being used by the framework as there is no concept of a
747 BIGNUM_METHOD and they could not be generalised to the new
748 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
749 ENGINE_cpy() has been removed as it cannot be consistently defined in
753 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
756 *) Change mkdef.pl to sort symbols that get the same entry number,
757 and make sure the automatically generated functions ERR_load_*
758 become part of libeay.num as well.
761 *) New function SSL_renegotiate_pending(). This returns true once
762 renegotiation has been requested (either SSL_renegotiate() call
763 or HelloRequest/ClientHello receveived from the peer) and becomes
764 false once a handshake has been completed.
765 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
766 sends a HelloRequest, but does not ensure that a handshake takes
767 place. SSL_renegotiate_pending() is useful for checking if the
768 client has followed the request.)
771 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
772 By default, clients may request session resumption even during
773 renegotiation (if session ID contexts permit); with this option,
774 session resumption is possible only in the first handshake.
776 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
777 more bits available for options that should not be part of
778 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
781 *) Add some demos for certificate and certificate request creation.
784 *) Make maximum certificate chain size accepted from the peer application
785 settable (SSL*_get/set_max_cert_list()), as proposed by
786 "Douglas E. Engert" <deengert@anl.gov>.
789 *) Add support for shared libraries for Unixware-7
790 (Boyd Lynn Gerber <gerberb@zenez.com>).
793 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
794 be done prior to destruction. Use this to unload error strings from
795 ENGINEs that load their own error strings. NB: This adds two new API
796 functions to "get" and "set" this destroy handler in an ENGINE.
799 *) Alter all existing ENGINE implementations (except "openssl" and
800 "openbsd") to dynamically instantiate their own error strings. This
801 makes them more flexible to be built both as statically-linked ENGINEs
802 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
803 Also, add stub code to each that makes building them as self-contained
804 shared-libraries easier (see README.ENGINE).
807 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
808 implementations into applications that are completely implemented in
809 self-contained shared-libraries. The "dynamic" ENGINE exposes control
810 commands that can be used to configure what shared-library to load and
811 to control aspects of the way it is handled. Also, made an update to
812 the README.ENGINE file that brings its information up-to-date and
813 provides some information and instructions on the "dynamic" ENGINE
814 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
817 *) Make it possible to unload ranges of ERR strings with a new
818 "ERR_unload_strings" function.
821 *) Add a copy() function to EVP_MD.
824 *) Make EVP_MD routines take a context pointer instead of just the
825 md_data void pointer.
828 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
829 that the digest can only process a single chunk of data
830 (typically because it is provided by a piece of
831 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
832 is only going to provide a single chunk of data, and hence the
833 framework needn't accumulate the data for oneshot drivers.
836 *) As with "ERR", make it possible to replace the underlying "ex_data"
837 functions. This change also alters the storage and management of global
838 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
839 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
840 index counters. The API functions that use this state have been changed
841 to take a "class_index" rather than pointers to the class's local STACK
842 and counter, and there is now an API function to dynamically create new
843 classes. This centralisation allows us to (a) plug a lot of the
844 thread-safety problems that existed, and (b) makes it possible to clean
845 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
846 such data would previously have always leaked in application code and
847 workarounds were in place to make the memory debugging turn a blind eye
848 to it. Application code that doesn't use this new function will still
849 leak as before, but their memory debugging output will announce it now
850 rather than letting it slide.
852 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
853 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
854 has a return value to indicate success or failure.
857 *) Make it possible to replace the underlying "ERR" functions such that the
858 global state (2 LHASH tables and 2 locks) is only used by the "default"
859 implementation. This change also adds two functions to "get" and "set"
860 the implementation prior to it being automatically set the first time
861 any other ERR function takes place. Ie. an application can call "get",
862 pass the return value to a module it has just loaded, and that module
863 can call its own "set" function using that value. This means the
864 module's "ERR" operations will use (and modify) the error state in the
865 application and not in its own statically linked copy of OpenSSL code.
868 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
869 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
870 the operation, and provides a more encapsulated way for external code
871 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
872 to use these functions rather than manually incrementing the counts.
874 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
877 *) Add EVP test program.
880 *) Add symmetric cipher support to ENGINE. Expect the API to change!
883 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
884 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
885 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
886 These allow a CRL to be built without having to access X509_CRL fields
887 directly. Modify 'ca' application to use new functions.
890 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
891 bug workarounds. Rollback attack detection is a security feature.
892 The problem will only arise on OpenSSL servers when TLSv1 is not
893 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
894 Software authors not wanting to support TLSv1 will have special reasons
895 for their choice and can explicitly enable this option.
896 [Bodo Moeller, Lutz Jaenicke]
898 *) Rationalise EVP so it can be extended: don't include a union of
899 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
900 (similar to those existing for EVP_CIPHER_CTX).
905 EVP_MD_CTX_init(&md); /* new function call */
906 EVP_DigestInit(&md, EVP_sha1());
907 EVP_DigestUpdate(&md, in, len);
908 EVP_DigestFinal(&md, out, NULL);
909 EVP_MD_CTX_cleanup(&md); /* new function call */
913 *) Make DES key schedule conform to the usual scheme, as well as
914 correcting its structure. This means that calls to DES functions
915 now have to pass a pointer to a des_key_schedule instead of a
916 plain des_key_schedule (which was actually always a pointer
921 des_set_key_checked(..., &ks);
922 des_ncbc_encrypt(..., &ks, ...);
924 (Note that a later change renames 'des_...' into 'DES_...'.)
927 *) Initial reduction of linker bloat: the use of some functions, such as
928 PEM causes large amounts of unused functions to be linked in due to
929 poor organisation. For example pem_all.c contains every PEM function
930 which has a knock on effect of linking in large amounts of (unused)
931 ASN1 code. Grouping together similar functions and splitting unrelated
932 functions prevents this.
935 *) Cleanup of EVP macros.
938 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
942 *) Add initial OCSP responder support to ocsp application. The
943 revocation information is handled using the text based index
944 use by the ca application. The responder can either handle
945 requests generated internally, supplied in files (for example
946 via a CGI script) or using an internal minimal server.
949 *) Add configuration choices to get zlib compression for TLS.
952 *) Changes to Kerberos SSL for RFC 2712 compliance:
953 1. Implemented real KerberosWrapper, instead of just using
954 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
955 2. Implemented optional authenticator field of KerberosWrapper.
957 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
958 and authenticator structs; see crypto/krb5/.
960 Generalized Kerberos calls to support multiple Kerberos libraries.
961 [Vern Staats <staatsvr@asc.hpc.mil>,
962 Jeffrey Altman <jaltman@columbia.edu>
965 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
966 already does with RSA. testdsa.h now has 'priv_key/pub_key'
967 values for each of the key sizes rather than having just
968 parameters (and 'speed' generating keys each time).
971 *) Speed up EVP routines.
974 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
975 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
976 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
977 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
979 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
980 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
981 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
984 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
986 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
989 *) Added the OS2-EMX target.
990 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
992 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
993 to support NCONF routines in extension code. New function CONF_set_nconf()
994 to allow functions which take an NCONF to also handle the old LHASH
995 structure: this means that the old CONF compatible routines can be
996 retained (in particular wrt extensions) without having to duplicate the
997 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1000 *) Enhance the general user interface with mechanisms for inner control
1001 and with possibilities to have yes/no kind of prompts.
1004 *) Change all calls to low level digest routines in the library and
1005 applications to use EVP. Add missing calls to HMAC_cleanup() and
1006 don't assume HMAC_CTX can be copied using memcpy().
1007 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1009 *) Add the possibility to control engines through control names but with
1010 arbitrary arguments instead of just a string.
1011 Change the key loaders to take a UI_METHOD instead of a callback
1012 function pointer. NOTE: this breaks binary compatibility with earlier
1013 versions of OpenSSL [engine].
1014 Adapt the nCipher code for these new conditions and add a card insertion
1018 *) Enhance the general user interface with mechanisms to better support
1019 dialog box interfaces, application-defined prompts, the possibility
1020 to use defaults (for example default passwords from somewhere else)
1021 and interrupts/cancellations.
1024 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1025 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1028 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1029 tidy up some unnecessarily weird code in 'sk_new()').
1030 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1032 *) Change the key loading routines for ENGINEs to use the same kind
1033 callback (pem_password_cb) as all other routines that need this
1037 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1038 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1039 than this minimum value is recommended.
1042 *) New random seeder for OpenVMS, using the system process statistics
1043 that are easily reachable.
1046 *) Windows apparently can't transparently handle global
1047 variables defined in DLLs. Initialisations such as:
1049 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1051 wont compile. This is used by the any applications that need to
1052 declare their own ASN1 modules. This was fixed by adding the option
1053 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1054 needed for static libraries under Win32.
1057 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1058 setting of purpose and trust fields. New X509_STORE trust and
1059 purpose functions and tidy up setting in other SSL functions.
1062 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1063 structure. These are inherited by X509_STORE_CTX when it is
1064 initialised. This allows various defaults to be set in the
1065 X509_STORE structure (such as flags for CRL checking and custom
1066 purpose or trust settings) for functions which only use X509_STORE_CTX
1067 internally such as S/MIME.
1069 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1070 trust settings if they are not set in X509_STORE. This allows X509_STORE
1071 purposes and trust (in S/MIME for example) to override any set by default.
1073 Add command line options for CRL checking to smime, s_client and s_server
1077 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1078 are set then the CRL is looked up in the X509_STORE structure and
1079 its validity and signature checked, then if the certificate is found
1080 in the CRL the verify fails with a revoked error.
1082 Various new CRL related callbacks added to X509_STORE_CTX structure.
1084 Command line options added to 'verify' application to support this.
1086 This needs some additional work, such as being able to handle multiple
1087 CRLs with different times, extension based lookup (rather than just
1088 by subject name) and ultimately more complete V2 CRL extension
1092 *) Add a general user interface API (crypto/ui/). This is designed
1093 to replace things like des_read_password and friends (backward
1094 compatibility functions using this new API are provided).
1095 The purpose is to remove prompting functions from the DES code
1096 section as well as provide for prompting through dialog boxes in
1097 a window system and the like.
1100 *) Add "ex_data" support to ENGINE so implementations can add state at a
1101 per-structure level rather than having to store it globally.
1104 *) Make it possible for ENGINE structures to be copied when retrieved by
1105 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1106 This causes the "original" ENGINE structure to act like a template,
1107 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1108 operational state can be localised to each ENGINE structure, despite the
1109 fact they all share the same "methods". New ENGINE structures returned in
1110 this case have no functional references and the return value is the single
1111 structural reference. This matches the single structural reference returned
1112 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1116 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1117 needs to match any other type at all we need to manually clear the
1121 *) Changes to the "openssl engine" utility to include;
1122 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1123 about an ENGINE's available control commands.
1124 - executing control commands from command line arguments using the
1125 '-pre' and '-post' switches. '-post' is only used if '-t' is
1126 specified and the ENGINE is successfully initialised. The syntax for
1127 the individual commands are colon-separated, for example;
1128 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1131 *) New dynamic control command support for ENGINEs. ENGINEs can now
1132 declare their own commands (numbers), names (strings), descriptions,
1133 and input types for run-time discovery by calling applications. A
1134 subset of these commands are implicitly classed as "executable"
1135 depending on their input type, and only these can be invoked through
1136 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1137 can be based on user input, config files, etc). The distinction is
1138 that "executable" commands cannot return anything other than a boolean
1139 result and can only support numeric or string input, whereas some
1140 discoverable commands may only be for direct use through
1141 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1142 pointers, or other custom uses. The "executable" commands are to
1143 support parameterisations of ENGINE behaviour that can be
1144 unambiguously defined by ENGINEs and used consistently across any
1145 OpenSSL-based application. Commands have been added to all the
1146 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1147 control over shared-library paths without source code alterations.
1150 *) Changed all ENGINE implementations to dynamically allocate their
1151 ENGINEs rather than declaring them statically. Apart from this being
1152 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1153 this also allows the implementations to compile without using the
1154 internal engine_int.h header.
1157 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1158 'const' value. Any code that should be able to modify a RAND_METHOD
1159 should already have non-const pointers to it (ie. they should only
1160 modify their own ones).
1163 *) Made a variety of little tweaks to the ENGINE code.
1164 - "atalla" and "ubsec" string definitions were moved from header files
1165 to C code. "nuron" string definitions were placed in variables
1166 rather than hard-coded - allowing parameterisation of these values
1167 later on via ctrl() commands.
1168 - Removed unused "#if 0"'d code.
1169 - Fixed engine list iteration code so it uses ENGINE_free() to release
1170 structural references.
1171 - Constified the RAND_METHOD element of ENGINE structures.
1172 - Constified various get/set functions as appropriate and added
1173 missing functions (including a catch-all ENGINE_cpy that duplicates
1174 all ENGINE values onto a new ENGINE except reference counts/state).
1175 - Removed NULL parameter checks in get/set functions. Setting a method
1176 or function to NULL is a way of cancelling out a previously set
1177 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1178 and doesn't justify the extra error symbols and code.
1179 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1180 flags from engine_int.h to engine.h.
1181 - Changed prototypes for ENGINE handler functions (init(), finish(),
1182 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1185 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1186 to the algorithm using long division. The binary algorithm can be
1187 used only if the modulus is odd. On 32-bit systems, it is faster
1188 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1189 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1190 up to 450 bits. In 64-bit environments, the binary algorithm
1191 appears to be advantageous for much longer moduli; here we use it
1192 for moduli up to 2048 bits.
1195 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1196 could not support the combine flag in choice fields.
1199 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1200 extensions from a certificate request to the certificate.
1203 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1204 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1205 file: this allows the display of the certificate about to be
1206 signed to be customised, to allow certain fields to be included
1207 or excluded and extension details. The old system didn't display
1208 multicharacter strings properly, omitted fields not in the policy
1209 and couldn't display additional details such as extensions.
1212 *) Function EC_POINTs_mul for multiple scalar multiplication
1213 of an arbitrary number of elliptic curve points
1214 \sum scalars[i]*points[i],
1215 optionally including the generator defined for the EC_GROUP:
1216 scalar*generator + \sum scalars[i]*points[i].
1218 EC_POINT_mul is a simple wrapper function for the typical case
1219 that the point list has just one item (besides the optional
1223 *) First EC_METHODs for curves over GF(p):
1225 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1226 operations and provides various method functions that can also
1227 operate with faster implementations of modular arithmetic.
1229 EC_GFp_mont_method() reuses most functions that are part of
1230 EC_GFp_simple_method, but uses Montgomery arithmetic.
1232 [Bodo Moeller; point addition and point doubling
1233 implementation directly derived from source code provided by
1234 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1236 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1237 crypto/ec/ec_lib.c):
1239 Curves are EC_GROUP objects (with an optional group generator)
1240 based on EC_METHODs that are built into the library.
1242 Points are EC_POINT objects based on EC_GROUP objects.
1244 Most of the framework would be able to handle curves over arbitrary
1245 finite fields, but as there are no obvious types for fields other
1246 than GF(p), some functions are limited to that for now.
1249 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1250 that the file contains a complete HTTP response.
1253 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1254 change the def and num file printf format specifier from "%-40sXXX"
1255 to "%-39s XXX". The latter will always guarantee a space after the
1256 field while the former will cause them to run together if the field
1257 is 40 of more characters long.
1260 *) Constify the cipher and digest 'method' functions and structures
1261 and modify related functions to take constant EVP_MD and EVP_CIPHER
1265 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1266 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1269 *) Modify EVP_Digest*() routines so they now return values. Although the
1270 internal software routines can never fail additional hardware versions
1274 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1276 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1277 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1280 ERR_R_NESTED_ASN1_ERROR
1282 ERR_R_MISSING_ASN1_EOS
1283 were 4 .. 9, conflicting with
1284 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1286 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1287 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1289 Add new error code 'ERR_R_INTERNAL_ERROR'.
1292 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1296 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1297 sets the subject name for a new request or supersedes the
1298 subject name in a given request. Formats that can be parsed are
1299 'CN=Some Name, OU=myOU, C=IT'
1301 'CN=Some Name/OU=myOU/C=IT'.
1303 Add options '-batch' and '-verbose' to 'openssl req'.
1304 [Massimiliano Pala <madwolf@hackmasters.net>]
1306 *) Introduce the possibility to access global variables through
1307 functions on platform were that's the best way to handle exporting
1308 global variables in shared libraries. To enable this functionality,
1309 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1310 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1311 is normally done by Configure or something similar).
1313 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1314 in the source file (foo.c) like this:
1316 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1317 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1319 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1320 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1322 OPENSSL_DECLARE_GLOBAL(int,foo);
1323 #define foo OPENSSL_GLOBAL_REF(foo)
1324 OPENSSL_DECLARE_GLOBAL(double,bar);
1325 #define bar OPENSSL_GLOBAL_REF(bar)
1327 The #defines are very important, and therefore so is including the
1328 header file everywhere where the defined globals are used.
1330 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1331 of ASN.1 items, but that structure is a bit different.
1333 The largest change is in util/mkdef.pl which has been enhanced with
1334 better and easier to understand logic to choose which symbols should
1335 go into the Windows .def files as well as a number of fixes and code
1336 cleanup (among others, algorithm keywords are now sorted
1337 lexicographically to avoid constant rewrites).
1340 *) In BN_div() keep a copy of the sign of 'num' before writing the
1341 result to 'rm' because if rm==num the value will be overwritten
1342 and produce the wrong result if 'num' is negative: this caused
1343 problems with BN_mod() and BN_nnmod().
1346 *) Function OCSP_request_verify(). This checks the signature on an
1347 OCSP request and verifies the signer certificate. The signer
1348 certificate is just checked for a generic purpose and OCSP request
1352 *) Add OCSP_check_validity() function to check the validity of OCSP
1353 responses. OCSP responses are prepared in real time and may only
1354 be a few seconds old. Simply checking that the current time lies
1355 between thisUpdate and nextUpdate max reject otherwise valid responses
1356 caused by either OCSP responder or client clock inaccuracy. Instead
1357 we allow thisUpdate and nextUpdate to fall within a certain period of
1358 the current time. The age of the response can also optionally be
1359 checked. Two new options -validity_period and -status_age added to
1363 *) If signature or public key algorithm is unrecognized print out its
1364 OID rather that just UNKNOWN.
1367 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1368 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1369 ID to be generated from the issuer certificate alone which can then be
1370 passed to OCSP_id_issuer_cmp().
1373 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1374 ASN1 modules to export functions returning ASN1_ITEM pointers
1375 instead of the ASN1_ITEM structures themselves. This adds several
1376 new macros which allow the underlying ASN1 function/structure to
1377 be accessed transparently. As a result code should not use ASN1_ITEM
1378 references directly (such as &X509_it) but instead use the relevant
1379 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1380 use of the new ASN1 code on platforms where exporting structures
1381 is problematical (for example in shared libraries) but exporting
1382 functions returning pointers to structures is not.
1385 *) Add support for overriding the generation of SSL/TLS session IDs.
1386 These callbacks can be registered either in an SSL_CTX or per SSL.
1387 The purpose of this is to allow applications to control, if they wish,
1388 the arbitrary values chosen for use as session IDs, particularly as it
1389 can be useful for session caching in multiple-server environments. A
1390 command-line switch for testing this (and any client code that wishes
1391 to use such a feature) has been added to "s_server".
1392 [Geoff Thorpe, Lutz Jaenicke]
1394 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1395 of the form '#if defined(...) || defined(...) || ...' and
1396 '#if !defined(...) && !defined(...) && ...'. This also avoids
1397 the growing number of special cases it was previously handling.
1400 *) Make all configuration macros available for application by making
1401 sure they are available in opensslconf.h, by giving them names starting
1402 with "OPENSSL_" to avoid conflicts with other packages and by making
1403 sure e_os2.h will cover all platform-specific cases together with
1405 Additionally, it is now possible to define configuration/platform-
1406 specific names (called "system identities"). In the C code, these
1407 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1408 macro with the name beginning with "OPENSSL_SYS_", which is determined
1409 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1413 *) New option -set_serial to 'req' and 'x509' this allows the serial
1414 number to use to be specified on the command line. Previously self
1415 signed certificates were hard coded with serial number 0 and the
1416 CA options of 'x509' had to use a serial number in a file which was
1420 *) New options to 'ca' utility to support V2 CRL entry extensions.
1421 Currently CRL reason, invalidity date and hold instruction are
1422 supported. Add new CRL extensions to V3 code and some new objects.
1425 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1426 disable standard block padding (aka PKCS#5 padding) in the EVP
1427 API, which was previously mandatory. This means that the data is
1428 not padded in any way and so the total length much be a multiple
1429 of the block size, otherwise an error occurs.
1432 *) Initial (incomplete) OCSP SSL support.
1435 *) New function OCSP_parse_url(). This splits up a URL into its host,
1436 port and path components: primarily to parse OCSP URLs. New -url
1437 option to ocsp utility.
1440 *) New nonce behavior. The return value of OCSP_check_nonce() now
1441 reflects the various checks performed. Applications can decide
1442 whether to tolerate certain situations such as an absent nonce
1443 in a response when one was present in a request: the ocsp application
1444 just prints out a warning. New function OCSP_add1_basic_nonce()
1445 this is to allow responders to include a nonce in a response even if
1446 the request is nonce-less.
1449 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1450 skipped when using openssl x509 multiple times on a single input file,
1451 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1454 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1455 set string type: to handle setting ASN1_TIME structures. Fix ca
1456 utility to correctly initialize revocation date of CRLs.
1459 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1460 the clients preferred ciphersuites and rather use its own preferences.
1461 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1462 Internet Explorer by ensuring unchanged hash method during stepup.
1463 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1466 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1467 to aes and add a new 'exist' option to print out symbols that don't
1471 *) Additional options to ocsp utility to allow flags to be set and
1472 additional certificates supplied.
1475 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1476 OCSP client a number of certificate to only verify the response
1480 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1481 handle the new API. Currently only ECB, CBC modes supported. Add new
1484 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1485 Encryption Standard (AES) Ciphersuites for Transport Layer
1486 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1487 not enabled by default and were not part of the "ALL" ciphersuite
1488 alias because they were not yet official; they could be
1489 explicitly requested by specifying the "AESdraft" ciphersuite
1490 group alias. In the final release of OpenSSL 0.9.7, the group
1491 alias is called "AES" and is part of "ALL".)
1492 [Ben Laurie, Steve Henson, Bodo Moeller]
1494 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1495 request to response.
1498 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1499 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1500 extract information from a certificate request. OCSP_response_create()
1501 creates a response and optionally adds a basic response structure.
1502 OCSP_basic_add1_status() adds a complete single response to a basic
1503 response and returns the OCSP_SINGLERESP structure just added (to allow
1504 extensions to be included for example). OCSP_basic_add1_cert() adds a
1505 certificate to a basic response and OCSP_basic_sign() signs a basic
1506 response with various flags. New helper functions ASN1_TIME_check()
1507 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1508 (converts ASN1_TIME to GeneralizedTime).
1511 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1512 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1513 structure from a certificate. X509_pubkey_digest() digests the public_key
1514 contents: this is used in various key identifiers.
1517 *) Make sk_sort() tolerate a NULL argument.
1518 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1520 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1521 passed by the function are trusted implicitly. If any of them signed the
1522 response then it is assumed to be valid and is not verified.
1525 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1526 to data. This was previously part of the PKCS7 ASN1 code. This
1527 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1528 [Steve Henson, reported by Kenneth R. Robinette
1529 <support@securenetterm.com>]
1531 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1532 routines: without these tracing memory leaks is very painful.
1533 Fix leaks in PKCS12 and PKCS7 routines.
1536 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1537 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1538 effectively meant GeneralizedTime would never be used. Now it
1539 is initialised to -1 but X509_time_adj() now has to check the value
1540 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1541 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1542 [Steve Henson, reported by Kenneth R. Robinette
1543 <support@securenetterm.com>]
1545 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1546 result in a zero length in the ASN1_INTEGER structure which was
1547 not consistent with the structure when d2i_ASN1_INTEGER() was used
1548 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1549 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1550 where it did not print out a minus for negative ASN1_INTEGER.
1553 *) Add summary printout to ocsp utility. The various functions which
1554 convert status values to strings have been renamed to:
1555 OCSP_response_status_str(), OCSP_cert_status_str() and
1556 OCSP_crl_reason_str() and are no longer static. New options
1557 to verify nonce values and to disable verification. OCSP response
1558 printout format cleaned up.
1561 *) Add additional OCSP certificate checks. These are those specified
1562 in RFC2560. This consists of two separate checks: the CA of the
1563 certificate being checked must either be the OCSP signer certificate
1564 or the issuer of the OCSP signer certificate. In the latter case the
1565 OCSP signer certificate must contain the OCSP signing extended key
1566 usage. This check is performed by attempting to match the OCSP
1567 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1568 in the OCSP_CERTID structures of the response.
1571 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1572 and related routines. This uses the standard OpenSSL certificate
1573 verify routines to perform initial checks (just CA validity) and
1574 to obtain the certificate chain. Then additional checks will be
1575 performed on the chain. Currently the root CA is checked to see
1576 if it is explicitly trusted for OCSP signing. This is used to set
1577 a root CA as a global signing root: that is any certificate that
1578 chains to that CA is an acceptable OCSP signing certificate.
1581 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1582 extensions from a separate configuration file.
1583 As when reading extensions from the main configuration file,
1584 the '-extensions ...' option may be used for specifying the
1586 [Massimiliano Pala <madwolf@comune.modena.it>]
1588 *) New OCSP utility. Allows OCSP requests to be generated or
1589 read. The request can be sent to a responder and the output
1590 parsed, outputed or printed in text form. Not complete yet:
1591 still needs to check the OCSP response validity.
1594 *) New subcommands for 'openssl ca':
1595 'openssl ca -status <serial>' prints the status of the cert with
1596 the given serial number (according to the index file).
1597 'openssl ca -updatedb' updates the expiry status of certificates
1599 [Massimiliano Pala <madwolf@comune.modena.it>]
1601 *) New '-newreq-nodes' command option to CA.pl. This is like
1602 '-newreq', but calls 'openssl req' with the '-nodes' option
1603 so that the resulting key is not encrypted.
1604 [Damien Miller <djm@mindrot.org>]
1606 *) New configuration for the GNU Hurd.
1607 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1609 *) Initial code to implement OCSP basic response verify. This
1610 is currently incomplete. Currently just finds the signer's
1611 certificate and verifies the signature on the response.
1614 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1615 value of OPENSSLDIR. This is available via the new '-d' option
1616 to 'openssl version', and is also included in 'openssl version -a'.
1619 *) Allowing defining memory allocation callbacks that will be given
1620 file name and line number information in additional arguments
1621 (a const char* and an int). The basic functionality remains, as
1622 well as the original possibility to just replace malloc(),
1623 realloc() and free() by functions that do not know about these
1624 additional arguments. To register and find out the current
1625 settings for extended allocation functions, the following
1626 functions are provided:
1628 CRYPTO_set_mem_ex_functions
1629 CRYPTO_set_locked_mem_ex_functions
1630 CRYPTO_get_mem_ex_functions
1631 CRYPTO_get_locked_mem_ex_functions
1633 These work the same way as CRYPTO_set_mem_functions and friends.
1634 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1635 extended allocation function is enabled.
1636 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1637 a conventional allocation function is enabled.
1638 [Richard Levitte, Bodo Moeller]
1640 *) Finish off removing the remaining LHASH function pointer casts.
1641 There should no longer be any prototype-casting required when using
1642 the LHASH abstraction, and any casts that remain are "bugs". See
1643 the callback types and macros at the head of lhash.h for details
1644 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1647 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1648 If /dev/[u]random devices are not available or do not return enough
1649 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1651 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1652 /etc/entropy will be queried once each in this sequence, quering stops
1653 when enough entropy was collected without querying more sockets.
1656 *) Change the Unix RAND_poll() variant to be able to poll several
1657 random devices, as specified by DEVRANDOM, until a sufficient amount
1658 of data has been collected. We spend at most 10 ms on each file
1659 (select timeout) and read in non-blocking mode. DEVRANDOM now
1660 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1661 (previously it was just the string "/dev/urandom"), so on typical
1662 platforms the 10 ms delay will never occur.
1663 Also separate out the Unix variant to its own file, rand_unix.c.
1664 For VMS, there's a currently-empty rand_vms.c.
1667 *) Move OCSP client related routines to ocsp_cl.c. These
1668 provide utility functions which an application needing
1669 to issue a request to an OCSP responder and analyse the
1670 response will typically need: as opposed to those which an
1671 OCSP responder itself would need which will be added later.
1673 OCSP_request_sign() signs an OCSP request with an API similar
1674 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1675 response. OCSP_response_get1_basic() extracts basic response
1676 from response. OCSP_resp_find_status(): finds and extracts status
1677 information from an OCSP_CERTID structure (which will be created
1678 when the request structure is built). These are built from lower
1679 level functions which work on OCSP_SINGLERESP structures but
1680 wont normally be used unless the application wishes to examine
1681 extensions in the OCSP response for example.
1683 Replace nonce routines with a pair of functions.
1684 OCSP_request_add1_nonce() adds a nonce value and optionally
1685 generates a random value. OCSP_check_nonce() checks the
1686 validity of the nonce in an OCSP response.
1689 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1690 This doesn't copy the supplied OCSP_CERTID and avoids the
1691 need to free up the newly created id. Change return type
1692 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1693 This can then be used to add extensions to the request.
1694 Deleted OCSP_request_new(), since most of its functionality
1695 is now in OCSP_REQUEST_new() (and the case insensitive name
1696 clash) apart from the ability to set the request name which
1697 will be added elsewhere.
1700 *) Update OCSP API. Remove obsolete extensions argument from
1701 various functions. Extensions are now handled using the new
1702 OCSP extension code. New simple OCSP HTTP function which
1703 can be used to send requests and parse the response.
1706 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1707 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1708 uses the special reorder version of SET OF to sort the attributes
1709 and reorder them to match the encoded order. This resolves a long
1710 standing problem: a verify on a PKCS7 structure just after signing
1711 it used to fail because the attribute order did not match the
1712 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1713 it uses the received order. This is necessary to tolerate some broken
1714 software that does not order SET OF. This is handled by encoding
1715 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1716 to produce the required SET OF.
1719 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1720 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1721 files to get correct declarations of the ASN.1 item variables.
1724 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1725 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1726 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1727 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1728 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1729 ASN1_ITEM and no wrapper functions.
1732 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1733 replace the old function pointer based I/O routines. Change most of
1734 the *_d2i_bio() and *_d2i_fp() functions to use these.
1737 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1738 lines, recognice more "algorithms" that can be deselected, and make
1739 it complain about algorithm deselection that isn't recognised.
1742 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1743 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1744 to use new functions. Add NO_ASN1_OLD which can be set to remove
1745 some old style ASN1 functions: this can be used to determine if old
1746 code will still work when these eventually go away.
1749 *) New extension functions for OCSP structures, these follow the
1750 same conventions as certificates and CRLs.
1753 *) New function X509V3_add1_i2d(). This automatically encodes and
1754 adds an extension. Its behaviour can be customised with various
1755 flags to append, replace or delete. Various wrappers added for
1756 certifcates and CRLs.
1759 *) Fix to avoid calling the underlying ASN1 print routine when
1760 an extension cannot be parsed. Correct a typo in the
1761 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1764 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1765 entries for variables.
1768 *) Add functionality to apps/openssl.c for detecting locking
1769 problems: As the program is single-threaded, all we have
1770 to do is register a locking callback using an array for
1771 storing which locks are currently held by the program.
1774 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1775 SSL_get_ex_data_X509_STORE_idx(), which is used in
1776 ssl_verify_cert_chain() and thus can be called at any time
1777 during TLS/SSL handshakes so that thread-safety is essential.
1778 Unfortunately, the ex_data design is not at all suited
1779 for multi-threaded use, so it probably should be abolished.
1782 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1783 [Broadcom, tweaked and integrated by Geoff Thorpe]
1785 *) Move common extension printing code to new function
1786 X509V3_print_extensions(). Reorganise OCSP print routines and
1787 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1790 *) New function X509_signature_print() to remove duplication in some
1794 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1795 set (this was treated exactly the same as SET OF previously). This
1796 is used to reorder the STACK representing the structure to match the
1797 encoding. This will be used to get round a problem where a PKCS7
1798 structure which was signed could not be verified because the STACK
1799 order did not reflect the encoded order.
1802 *) Reimplement the OCSP ASN1 module using the new code.
1805 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1806 for its ASN1 operations. The old style function pointers still exist
1807 for now but they will eventually go away.
1810 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1811 completely replaces the old ASN1 functionality with a table driven
1812 encoder and decoder which interprets an ASN1_ITEM structure describing
1813 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1814 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1815 has also been converted to the new form.
1818 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1819 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1820 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1821 for negative moduli.
1824 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1825 of not touching the result's sign bit.
1828 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1832 *) Changed the LHASH code to use prototypes for callbacks, and created
1833 macros to declare and implement thin (optionally static) functions
1834 that provide type-safety and avoid function pointer casting for the
1835 type-specific callbacks.
1838 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1840 [Veers Staats <staatsvr@asc.hpc.mil>,
1841 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1843 *) Reformat the FAQ so the different questions and answers can be divided
1844 in sections depending on the subject.
1847 *) Have the zlib compression code load ZLIB.DLL dynamically under
1851 *) New function BN_mod_sqrt for computing square roots modulo a prime
1852 (using the probabilistic Tonelli-Shanks algorithm unless
1853 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1854 be handled deterministically).
1855 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1857 *) Make BN_mod_inverse faster by explicitly handling small quotients
1858 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1859 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1862 *) New function BN_kronecker.
1865 *) Fix BN_gcd so that it works on negative inputs; the result is
1866 positive unless both parameters are zero.
1867 Previously something reasonably close to an infinite loop was
1868 possible because numbers could be growing instead of shrinking
1869 in the implementation of Euclid's algorithm.
1872 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1873 sign of the number in question.
1875 Fix BN_is_word(a,w) to work correctly for w == 0.
1877 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1878 because its test if the absolute value of 'a' equals 'w'.
1879 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1880 it exists mostly for use in the implementations of BN_is_zero(),
1881 BN_is_one(), and BN_is_word().
1884 *) New function BN_swap.
1887 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1888 the exponentiation functions are more likely to produce reasonable
1889 results on negative inputs.
1892 *) Change BN_mod_mul so that the result is always non-negative.
1893 Previously, it could be negative if one of the factors was negative;
1894 I don't think anyone really wanted that behaviour.
1897 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1898 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1899 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1900 and add new functions:
1909 BN_mod_lshift1_quick
1913 These functions always generate non-negative results.
1915 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1916 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1918 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1919 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1920 be reduced modulo m.
1921 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1924 The following entry accidentily appeared in the CHANGES file
1925 distributed with OpenSSL 0.9.7. The modifications described in
1926 it do *not* apply to OpenSSL 0.9.7.
1928 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1929 was actually never needed) and in BN_mul(). The removal in BN_mul()
1930 required a small change in bn_mul_part_recursive() and the addition
1931 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1932 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1933 bn_sub_words() and bn_add_words() except they take arrays with
1938 *) In 'openssl passwd', verify passwords read from the terminal
1939 unless the '-salt' option is used (which usually means that
1940 verification would just waste user's time since the resulting
1941 hash is going to be compared with some given password hash)
1942 or the new '-noverify' option is used.
1944 This is an incompatible change, but it does not affect
1945 non-interactive use of 'openssl passwd' (passwords on the command
1946 line, '-stdin' option, '-in ...' option) and thus should not
1950 *) Remove all references to RSAref, since there's no more need for it.
1953 *) Make DSO load along a path given through an environment variable
1954 (SHLIB_PATH) with shl_load().
1957 *) Constify the ENGINE code as a result of BIGNUM constification.
1958 Also constify the RSA code and most things related to it. In a
1959 few places, most notable in the depth of the ASN.1 code, ugly
1960 casts back to non-const were required (to be solved at a later
1964 *) Make it so the openssl application has all engines loaded by default.
1967 *) Constify the BIGNUM routines a little more.
1970 *) Add the following functions:
1972 ENGINE_load_cswift()
1974 ENGINE_load_atalla()
1976 ENGINE_load_builtin_engines()
1978 That way, an application can itself choose if external engines that
1979 are built-in in OpenSSL shall ever be used or not. The benefit is
1980 that applications won't have to be linked with libdl or other dso
1981 libraries unless it's really needed.
1983 Changed 'openssl engine' to load all engines on demand.
1984 Changed the engine header files to avoid the duplication of some
1985 declarations (they differed!).
1988 *) 'openssl engine' can now list capabilities.
1991 *) Better error reporting in 'openssl engine'.
1994 *) Never call load_dh_param(NULL) in s_server.
1997 *) Add engine application. It can currently list engines by name and
1998 identity, and test if they are actually available.
2001 *) Improve RPM specification file by forcing symbolic linking and making
2002 sure the installed documentation is also owned by root.root.
2003 [Damien Miller <djm@mindrot.org>]
2005 *) Give the OpenSSL applications more possibilities to make use of
2006 keys (public as well as private) handled by engines.
2009 *) Add OCSP code that comes from CertCo.
2012 *) Add VMS support for the Rijndael code.
2015 *) Added untested support for Nuron crypto accelerator.
2018 *) Add support for external cryptographic devices. This code was
2019 previously distributed separately as the "engine" branch.
2020 [Geoff Thorpe, Richard Levitte]
2022 *) Rework the filename-translation in the DSO code. It is now possible to
2023 have far greater control over how a "name" is turned into a filename
2024 depending on the operating environment and any oddities about the
2025 different shared library filenames on each system.
2028 *) Support threads on FreeBSD-elf in Configure.
2031 *) Fix for SHA1 assembly problem with MASM: it produces
2032 warnings about corrupt line number information when assembling
2033 with debugging information. This is caused by the overlapping
2035 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2038 NCONF_get_number() has no error checking at all. As a replacement,
2039 NCONF_get_number_e() is defined (_e for "error checking") and is
2040 promoted strongly. The old NCONF_get_number is kept around for
2041 binary backward compatibility.
2042 Make it possible for methods to load from something other than a BIO,
2043 by providing a function pointer that is given a name instead of a BIO.
2044 For example, this could be used to load configuration data from an
2048 *) Fix for non blocking accept BIOs. Added new I/O special reason
2049 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2050 with non blocking I/O was not possible because no retry code was
2051 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2055 *) Added the beginnings of Rijndael support.
2058 *) Fix for bug in DirectoryString mask setting. Add support for
2059 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2060 to allow certificate printing to more controllable, additional
2061 'certopt' option to 'x509' to allow new printing options to be
2065 *) Clean old EAY MD5 hack from e_os.h.
2068 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2070 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2071 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2072 [Joe Orton, Steve Henson]
2074 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2076 *) Fix additional bug revealed by the NISCC test suite:
2078 Stop bug triggering large recursion when presented with
2079 certain ASN.1 tags (CAN-2003-0851)
2082 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2084 *) Fix various bugs revealed by running the NISCC test suite:
2086 Stop out of bounds reads in the ASN1 code when presented with
2087 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2089 If verify callback ignores invalid public key errors don't try to check
2090 certificate signature with the NULL public key.
2094 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2095 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2099 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2100 extra data after the compression methods not only for TLS 1.0
2101 but also for SSL 3.0 (as required by the specification).
2102 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2104 *) Change X509_certificate_type() to mark the key as exported/exportable
2105 when it's 512 *bits* long, not 512 bytes.
2108 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2110 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2111 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2112 a protocol version number mismatch like a decryption error
2113 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2116 *) Turn on RSA blinding by default in the default implementation
2117 to avoid a timing attack. Applications that don't want it can call
2118 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2119 They would be ill-advised to do so in most cases.
2120 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2122 *) Change RSA blinding code so that it works when the PRNG is not
2123 seeded (in this case, the secret RSA exponent is abused as
2124 an unpredictable seed -- if it is not unpredictable, there
2125 is no point in blinding anyway). Make RSA blinding thread-safe
2126 by remembering the creator's thread ID in rsa->blinding and
2127 having all other threads use local one-time blinding factors
2128 (this requires more computation than sharing rsa->blinding, but
2129 avoids excessive locking; and if an RSA object is not shared
2130 between threads, blinding will still be very fast).
2133 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2135 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2136 via timing by performing a MAC computation even if incorrrect
2137 block cipher padding has been found. This is a countermeasure
2138 against active attacks where the attacker has to distinguish
2139 between bad padding and a MAC verification error. (CAN-2003-0078)
2141 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2142 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2143 Martin Vuagnoux (EPFL, Ilion)]
2145 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2147 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2148 memory from it's contents. This is done with a counter that will
2149 place alternating values in each byte. This can be used to solve
2150 two issues: 1) the removal of calls to memset() by highly optimizing
2151 compilers, and 2) cleansing with other values than 0, since those can
2152 be read through on certain media, for example a swap space on disk.
2155 *) Bugfix: client side session caching did not work with external caching,
2156 because the session->cipher setting was not restored when reloading
2157 from the external cache. This problem was masked, when
2158 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2159 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2162 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2163 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2164 [Zeev Lieber <zeev-l@yahoo.com>]
2166 *) Undo an undocumented change introduced in 0.9.6e which caused
2167 repeated calls to OpenSSL_add_all_ciphers() and
2168 OpenSSL_add_all_digests() to be ignored, even after calling
2172 *) Change the default configuration reader to deal with last line not
2173 being properly terminated.
2176 *) Change X509_NAME_cmp() so it applies the special rules on handling
2177 DN values that are of type PrintableString, as well as RDNs of type
2178 emailAddress where the value has the type ia5String.
2179 [stefank@valicert.com via Richard Levitte]
2181 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2182 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2183 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2184 the bitwise-OR of the two for use by the majority of applications
2185 wanting this behaviour, and update the docs. The documented
2186 behaviour and actual behaviour were inconsistent and had been
2187 changing anyway, so this is more a bug-fix than a behavioural
2189 [Geoff Thorpe, diagnosed by Nadav Har'El]
2191 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2192 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2195 *) Fix initialization code race conditions in
2196 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2197 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2198 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2199 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2200 ssl2_get_cipher_by_char(),
2201 ssl3_get_cipher_by_char().
2202 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2204 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2205 the cached sessions are flushed, as the remove_cb() might use ex_data
2206 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2207 (see [openssl.org #212]).
2208 [Geoff Thorpe, Lutz Jaenicke]
2210 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2211 length, instead of the encoding length to d2i_ASN1_OBJECT.
2214 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2216 *) [In 0.9.6g-engine release:]
2217 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2218 [Lynn Gazis <lgazis@rainbow.com>]
2220 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2222 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2223 and get fix the header length calculation.
2224 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2225 Alon Kantor <alonk@checkpoint.com> (and others),
2228 *) Use proper error handling instead of 'assertions' in buffer
2229 overflow checks added in 0.9.6e. This prevents DoS (the
2230 assertions could call abort()).
2231 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2233 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2235 *) Add various sanity checks to asn1_get_length() to reject
2236 the ASN1 length bytes if they exceed sizeof(long), will appear
2237 negative or the content length exceeds the length of the
2239 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2241 *) Fix cipher selection routines: ciphers without encryption had no flags
2242 for the cipher strength set and where therefore not handled correctly
2243 by the selection routines (PR #130).
2246 *) Fix EVP_dsa_sha macro.
2250 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2251 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2252 that was added in OpenSSL 0.9.6d.
2254 As the countermeasure turned out to be incompatible with some
2255 broken SSL implementations, the new option is part of SSL_OP_ALL.
2256 SSL_OP_ALL is usually employed when compatibility with weird SSL
2257 implementations is desired (e.g. '-bugs' option to 's_client' and
2258 's_server'), so the new option is automatically set in many
2262 *) Changes in security patch:
2264 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2265 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2266 Air Force Materiel Command, USAF, under agreement number
2269 *) Add various sanity checks to asn1_get_length() to reject
2270 the ASN1 length bytes if they exceed sizeof(long), will appear
2271 negative or the content length exceeds the length of the
2272 supplied buffer. (CAN-2002-0659)
2273 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2275 *) Assertions for various potential buffer overflows, not known to
2277 [Ben Laurie (CHATS)]
2279 *) Various temporary buffers to hold ASCII versions of integers were
2280 too small for 64 bit platforms. (CAN-2002-0655)
2281 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2283 *) Remote buffer overflow in SSL3 protocol - an attacker could
2284 supply an oversized session ID to a client. (CAN-2002-0656)
2285 [Ben Laurie (CHATS)]
2287 *) Remote buffer overflow in SSL2 protocol - an attacker could
2288 supply an oversized client master key. (CAN-2002-0656)
2289 [Ben Laurie (CHATS)]
2291 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2293 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2294 encoded as NULL) with id-dsa-with-sha1.
2295 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2297 *) Check various X509_...() return values in apps/req.c.
2298 [Nils Larsch <nla@trustcenter.de>]
2300 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2301 an end-of-file condition would erronously be flagged, when the CRLF
2302 was just at the end of a processed block. The bug was discovered when
2303 processing data through a buffering memory BIO handing the data to a
2304 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2305 <ptsekov@syntrex.com> and Nedelcho Stanev.
2308 *) Implement a countermeasure against a vulnerability recently found
2309 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2310 before application data chunks to avoid the use of known IVs
2311 with data potentially chosen by the attacker.
2314 *) Fix length checks in ssl3_get_client_hello().
2317 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2318 to prevent ssl3_read_internal() from incorrectly assuming that
2319 ssl3_read_bytes() found application data while handshake
2320 processing was enabled when in fact s->s3->in_read_app_data was
2321 merely automatically cleared during the initial handshake.
2322 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2324 *) Fix object definitions for Private and Enterprise: they were not
2325 recognized in their shortname (=lowercase) representation. Extend
2326 obj_dat.pl to issue an error when using undefined keywords instead
2327 of silently ignoring the problem (Svenning Sorensen
2328 <sss@sss.dnsalias.net>).
2331 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2332 generators, i.e. generators other than 2 and 5. (Previously, the
2333 code did not properly initialise the 'add' and 'rem' values to
2334 BN_generate_prime().)
2336 In the new general case, we do not insist that 'generator' is
2337 actually a primitive root: This requirement is rather pointless;
2338 a generator of the order-q subgroup is just as good, if not
2342 *) Map new X509 verification errors to alerts. Discovered and submitted by
2343 Tom Wu <tom@arcot.com>.
2346 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2347 returning non-zero before the data has been completely received
2348 when using non-blocking I/O.
2349 [Bodo Moeller; problem pointed out by John Hughes]
2351 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2352 [Ben Laurie, Lutz Jaenicke]
2354 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2355 Yoram Zahavi <YoramZ@gilian.com>).
2358 *) Add information about CygWin 1.3 and on, and preserve proper
2359 configuration for the versions before that.
2360 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2362 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2363 check whether we deal with a copy of a session and do not delete from
2364 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2365 <izhar@checkpoint.com>.
2368 *) Do not store session data into the internal session cache, if it
2369 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2370 flag is set). Proposed by Aslam <aslam@funk.com>.
2373 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2377 *) [In 0.9.6d-engine release:]
2378 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2379 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2381 *) Add the configuration target linux-s390x.
2382 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2384 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2385 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2386 variable as an indication that a ClientHello message has been
2387 received. As the flag value will be lost between multiple
2388 invocations of ssl3_accept when using non-blocking I/O, the
2389 function may not be aware that a handshake has actually taken
2390 place, thus preventing a new session from being added to the
2393 To avoid this problem, we now set s->new_session to 2 instead of
2394 using a local variable.
2395 [Lutz Jaenicke, Bodo Moeller]
2397 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2398 if the SSL_R_LENGTH_MISMATCH error is detected.
2399 [Geoff Thorpe, Bodo Moeller]
2401 *) New 'shared_ldflag' column in Configure platform table.
2404 *) Fix EVP_CIPHER_mode macro.
2405 ["Dan S. Camper" <dan@bti.net>]
2407 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2408 type, we must throw them away by setting rr->length to 0.
2409 [D P Chang <dpc@qualys.com>]
2411 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2413 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2414 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2415 worked incorrectly for those cases where range = 10..._2 and
2416 3*range is two bits longer than range.)
2419 *) Only add signing time to PKCS7 structures if it is not already
2423 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2424 OBJ_ld_ce should be OBJ_id_ce.
2425 Also some ip-pda OIDs in crypto/objects/objects.txt were
2426 incorrect (cf. RFC 3039).
2427 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2429 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2430 returns early because it has nothing to do.
2431 [Andy Schneider <andy.schneider@bjss.co.uk>]
2433 *) [In 0.9.6c-engine release:]
2434 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2435 [Andy Schneider <andy.schneider@bjss.co.uk>]
2437 *) [In 0.9.6c-engine release:]
2438 Add support for Cryptographic Appliance's keyserver technology.
2439 (Use engine 'keyclient')
2440 [Cryptographic Appliances and Geoff Thorpe]
2442 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2443 is called via tools/c89.sh because arguments have to be
2444 rearranged (all '-L' options must appear before the first object
2446 [Richard Shapiro <rshapiro@abinitio.com>]
2448 *) [In 0.9.6c-engine release:]
2449 Add support for Broadcom crypto accelerator cards, backported
2451 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2453 *) [In 0.9.6c-engine release:]
2454 Add support for SureWare crypto accelerator cards from
2455 Baltimore Technologies. (Use engine 'sureware')
2456 [Baltimore Technologies and Mark Cox]
2458 *) [In 0.9.6c-engine release:]
2459 Add support for crypto accelerator cards from Accelerated
2460 Encryption Processing, www.aep.ie. (Use engine 'aep')
2461 [AEP Inc. and Mark Cox]
2463 *) Add a configuration entry for gcc on UnixWare.
2464 [Gary Benson <gbenson@redhat.com>]
2466 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2467 messages are stored in a single piece (fixed-length part and
2468 variable-length part combined) and fix various bugs found on the way.
2471 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2472 instead. BIO_gethostbyname() does not know what timeouts are
2473 appropriate, so entries would stay in cache even when they have
2475 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2477 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2478 faced with a pathologically small ClientHello fragment that does
2479 not contain client_version: Instead of aborting with an error,
2480 simply choose the highest available protocol version (i.e.,
2481 TLS 1.0 unless it is disabled). In practice, ClientHello
2482 messages are never sent like this, but this change gives us
2483 strictly correct behaviour at least for TLS.
2486 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2487 never resets s->method to s->ctx->method when called from within
2488 one of the SSL handshake functions.
2489 [Bodo Moeller; problem pointed out by Niko Baric]
2491 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2492 (sent using the client's version number) if client_version is
2493 smaller than the protocol version in use. Also change
2494 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2495 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2496 the client will at least see that alert.
2499 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2503 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2504 client receives HelloRequest while in a handshake.
2505 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2507 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2508 should end in 'break', not 'goto end' which circuments various
2509 cleanups done in state SSL_ST_OK. But session related stuff
2510 must be disabled for SSL_ST_OK in the case that we just sent a
2513 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2514 before just sending a HelloRequest.
2515 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2517 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2518 reveal whether illegal block cipher padding was found or a MAC
2519 verification error occured. (Neither SSLerr() codes nor alerts
2520 are directly visible to potential attackers, but the information
2521 may leak via logfiles.)
2523 Similar changes are not required for the SSL 2.0 implementation
2524 because the number of padding bytes is sent in clear for SSL 2.0,
2525 and the extra bytes are just ignored. However ssl/s2_pkt.c
2526 failed to verify that the purported number of padding bytes is in
2530 *) Add OpenUNIX-8 support including shared libraries
2531 (Boyd Lynn Gerber <gerberb@zenez.com>).
2534 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2535 'wristwatch attack' using huge encoding parameters (cf.
2536 James H. Manger's CRYPTO 2001 paper). Note that the
2537 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2538 encoding parameters and hence was not vulnerable.
2541 *) BN_sqr() bug fix.
2542 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2544 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2545 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2546 followed by modular reduction.
2547 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2549 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2550 equivalent based on BN_pseudo_rand() instead of BN_rand().
2553 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2554 This function was broken, as the check for a new client hello message
2555 to handle SGC did not allow these large messages.
2556 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2559 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2562 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2563 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2566 *) Rework the configuration and shared library support for Tru64 Unix.
2567 The configuration part makes use of modern compiler features and
2568 still retains old compiler behavior for those that run older versions
2569 of the OS. The shared library support part includes a variant that
2570 uses the RPATH feature, and is available through the special
2571 configuration target "alpha-cc-rpath", which will never be selected
2573 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2575 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2576 with the same message size as in ssl3_get_certificate_request().
2577 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2578 messages might inadvertently be reject as too long.
2579 [Petr Lampa <lampa@fee.vutbr.cz>]
2581 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2584 *) Modified SSL library such that the verify_callback that has been set
2585 specificly for an SSL object with SSL_set_verify() is actually being
2586 used. Before the change, a verify_callback set with this function was
2587 ignored and the verify_callback() set in the SSL_CTX at the time of
2588 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2589 to allow the necessary settings.
2592 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2593 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2594 done automatically (in contradiction to the requirements of the C
2595 standard). This made problems when used from OpenSSH.
2598 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2599 dh->length and always used
2601 BN_rand_range(priv_key, dh->p).
2603 BN_rand_range() is not necessary for Diffie-Hellman, and this
2604 specific range makes Diffie-Hellman unnecessarily inefficient if
2605 dh->length (recommended exponent length) is much smaller than the
2606 length of dh->p. We could use BN_rand_range() if the order of
2607 the subgroup was stored in the DH structure, but we only have
2612 BN_rand(priv_key, l, ...)
2614 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2620 RSA_eay_public_encrypt
2621 RSA_eay_private_decrypt
2622 RSA_eay_private_encrypt (signing)
2623 RSA_eay_public_decrypt (signature verification)
2625 (default implementations for RSA_public_encrypt,
2626 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2627 always reject numbers >= n.
2630 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2631 to synchronize access to 'locking_thread'. This is necessary on
2632 systems where access to 'locking_thread' (an 'unsigned long'
2633 variable) is not atomic.
2636 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2637 *before* setting the 'crypto_lock_rand' flag. The previous code had
2638 a race condition if 0 is a valid thread ID.
2639 [Travis Vitek <vitek@roguewave.com>]
2641 *) Add support for shared libraries under Irix.
2642 [Albert Chin-A-Young <china@thewrittenword.com>]
2644 *) Add configuration option to build on Linux on both big-endian and
2646 [Ralf Baechle <ralf@uni-koblenz.de>]
2648 *) Add the possibility to create shared libraries on HP-UX.
2651 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2653 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2654 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2655 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2656 PRNG state recovery was possible based on the output of
2657 one PRNG request appropriately sized to gain knowledge on
2658 'md' followed by enough consecutive 1-byte PRNG requests
2659 to traverse all of 'state'.
2661 1. When updating 'md_local' (the current thread's copy of 'md')
2662 during PRNG output generation, hash all of the previous
2663 'md_local' value, not just the half used for PRNG output.
2665 2. Make the number of bytes from 'state' included into the hash
2666 independent from the number of PRNG bytes requested.
2668 The first measure alone would be sufficient to avoid
2669 Markku-Juhani's attack. (Actually it had never occurred
2670 to me that the half of 'md_local' used for chaining was the
2671 half from which PRNG output bytes were taken -- I had always
2672 assumed that the secret half would be used.) The second
2673 measure makes sure that additional data from 'state' is never
2674 mixed into 'md_local' in small portions; this heuristically
2675 further strengthens the PRNG.
2678 *) Fix crypto/bn/asm/mips3.s.
2681 *) When only the key is given to "enc", the IV is undefined. Print out
2682 an error message in this case.
2685 *) Handle special case when X509_NAME is empty in X509 printing routines.
2688 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2689 positive and less than q.
2692 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2693 used: it isn't thread safe and the add_lock_callback should handle
2695 [Paul Rose <Paul.Rose@bridge.com>]
2697 *) Verify that incoming data obeys the block size in
2698 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2702 [Ulf Möller, Bodo Möller]
2704 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2705 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2706 when fixing the server behaviour for backwards-compatible 'client
2707 hello' messages. (Note that the attack is impractical against
2708 SSL 3.0 and TLS 1.0 anyway because length and version checking
2709 means that the probability of guessing a valid ciphertext is
2710 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2713 Before 0.9.5, the countermeasure (hide the error by generating a
2714 random 'decryption result') did not work properly because
2715 ERR_clear_error() was missing, meaning that SSL_get_error() would
2716 detect the supposedly ignored error.
2718 Both problems are now fixed.
2721 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2722 (previously it was 1024).
2725 *) Fix for compatibility mode trust settings: ignore trust settings
2726 unless some valid trust or reject settings are present.
2729 *) Fix for blowfish EVP: its a variable length cipher.
2732 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2733 parameters in DSA public key structures and return an error in the
2734 DSA routines if parameters are absent.
2737 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2738 in the current directory if neither $RANDFILE nor $HOME was set.
2739 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2740 caused some confusion to Windows users who haven't defined $HOME.
2741 Thus RAND_file_name() is changed again: e_os.h can define a
2742 DEFAULT_HOME, which will be used if $HOME is not set.
2743 For Windows, we use "C:"; on other platforms, we still require
2744 environment variables.
2746 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2747 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2748 having multiple threads call RAND_poll() concurrently.
2751 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2752 combination of a flag and a thread ID variable.
2753 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2754 flag), *other* threads can enter ssleay_add_bytes without obeying
2755 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2756 that they do not hold after the first thread unsets add_do_not_lock).
2759 *) Change bctest again: '-x' expressions are not available in all
2763 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2765 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2766 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2768 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2769 the default extension for executables, if any. Also, make the perl
2770 scripts that use symlink() to test if it really exists and use "cp"
2771 if it doesn't. All this made OpenSSL compilable and installable in
2775 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2776 If SEQUENCE is length is indefinite just set c->slen to the total
2777 amount of data available.
2778 [Steve Henson, reported by shige@FreeBSD.org]
2779 [This change does not apply to 0.9.7.]
2781 *) Change bctest to avoid here-documents inside command substitution
2782 (workaround for FreeBSD /bin/sh bug).
2783 For compatibility with Ultrix, avoid shell functions (introduced
2784 in the bctest version that searches along $PATH).
2787 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2788 with des_encrypt() defined on some operating systems, like Solaris
2792 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2793 On the Importance of Eliminating Errors in Cryptographic
2794 Computations, J. Cryptology 14 (2001) 2, 101-119,
2795 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2798 *) MIPS assembler BIGNUM division bug fix.
2801 *) Disabled incorrect Alpha assembler code.
2804 *) Fix PKCS#7 decode routines so they correctly update the length
2805 after reading an EOC for the EXPLICIT tag.
2807 [This change does not apply to 0.9.7.]
2809 *) Fix bug in PKCS#12 key generation routines. This was triggered
2810 if a 3DES key was generated with a 0 initial byte. Include
2811 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2812 (but broken) behaviour.
2815 *) Enhance bctest to search for a working bc along $PATH and print
2817 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2819 *) Fix memory leaks in err.c: free err_data string if necessary;
2820 don't write to the wrong index in ERR_set_error_data.
2823 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2827 *) Replace rdtsc with _emit statements for VC++ version 5.
2828 [Jeremy Cooper <jeremy@baymoo.org>]
2830 *) Make it possible to reuse SSLv2 sessions.
2833 *) In copy_email() check for >= 0 as a return value for
2834 X509_NAME_get_index_by_NID() since 0 is a valid index.
2835 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2837 *) Avoid coredump with unsupported or invalid public keys by checking if
2838 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2839 PKCS7_verify() fails with non detached data.
2842 *) Don't use getenv in library functions when run as setuid/setgid.
2843 New function OPENSSL_issetugid().
2846 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2847 due to incorrect handling of multi-threading:
2849 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2851 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2853 3. Count how many times MemCheck_off() has been called so that
2854 nested use can be treated correctly. This also avoids
2855 inband-signalling in the previous code (which relied on the
2856 assumption that thread ID 0 is impossible).
2859 *) Add "-rand" option also to s_client and s_server.
2862 *) Fix CPU detection on Irix 6.x.
2863 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2864 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2866 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2869 [This change does not apply to 0.9.7.]
2871 *) Use the cached encoding of an X509_NAME structure rather than
2872 copying it. This is apparently the reason for the libsafe "errors"
2873 but the code is actually correct.
2876 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2877 Bleichenbacher's DSA attack.
2878 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2879 to be set and top=0 forces the highest bit to be set; top=-1 is new
2880 and leaves the highest bit random.
2881 [Ulf Moeller, Bodo Moeller]
2883 *) In the NCONF_...-based implementations for CONF_... queries
2884 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2885 a temporary CONF structure with the data component set to NULL
2886 (which gives segmentation faults in lh_retrieve).
2887 Instead, use NULL for the CONF pointer in CONF_get_string and
2888 CONF_get_number (which may use environment variables) and directly
2889 return NULL from CONF_get_section.
2892 *) Fix potential buffer overrun for EBCDIC.
2895 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2896 keyUsage if basicConstraints absent for a CA.
2899 *) Make SMIME_write_PKCS7() write mail header values with a format that
2900 is more generally accepted (no spaces before the semicolon), since
2901 some programs can't parse those values properly otherwise. Also make
2902 sure BIO's that break lines after each write do not create invalid
2906 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2907 macros previously used would not encode an empty SEQUENCE OF
2908 and break the signature.
2910 [This change does not apply to 0.9.7.]
2912 *) Zero the premaster secret after deriving the master secret in
2916 *) Add some EVP_add_digest_alias registrations (as found in
2917 OpenSSL_add_all_digests()) to SSL_library_init()
2918 aka OpenSSL_add_ssl_algorithms(). This provides improved
2919 compatibility with peers using X.509 certificates
2920 with unconventional AlgorithmIdentifier OIDs.
2923 *) Fix for Irix with NO_ASM.
2924 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2926 *) ./config script fixes.
2927 [Ulf Moeller, Richard Levitte]
2929 *) Fix 'openssl passwd -1'.
2932 *) Change PKCS12_key_gen_asc() so it can cope with non null
2933 terminated strings whose length is passed in the passlen
2934 parameter, for example from PEM callbacks. This was done
2935 by adding an extra length parameter to asc2uni().
2936 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2938 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2939 call failed, free the DSA structure.
2942 *) Fix to uni2asc() to cope with zero length Unicode strings.
2943 These are present in some PKCS#12 files.
2946 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2947 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2948 when writing a 32767 byte record.
2949 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2951 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2952 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2954 (RSA objects have a reference count access to which is protected
2955 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2956 so they are meant to be shared between threads.)
2957 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2958 "Reddie, Steven" <Steven.Reddie@ca.com>]
2960 *) Fix a deadlock in CRYPTO_mem_leaks().
2963 *) Use better test patterns in bntest.
2966 *) rand_win.c fix for Borland C.
2969 *) BN_rshift bugfix for n == 0.
2972 *) Add a 'bctest' script that checks for some known 'bc' bugs
2973 so that 'make test' does not abort just because 'bc' is broken.
2976 *) Store verify_result within SSL_SESSION also for client side to
2977 avoid potential security hole. (Re-used sessions on the client side
2978 always resulted in verify_result==X509_V_OK, not using the original
2979 result of the server certificate verification.)
2982 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2983 SSL3_RT_APPLICATION_DATA, return 0.
2984 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2988 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2989 releases, have been re-implemented by renaming the previous
2990 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2991 and ssl3_read_internal, respectively, and adding 'peek' parameters
2992 to them. The new ssl[23]_{read,peek} functions are calls to
2993 ssl[23]_read_internal with the 'peek' flag set appropriately.
2994 A 'peek' parameter has also been added to ssl3_read_bytes, which
2995 does the actual work for ssl3_read_internal.
2998 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2999 the method-specific "init()" handler. Also clean up ex_data after
3000 calling the method-specific "finish()" handler. Previously, this was
3001 happening the other way round.
3004 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3005 The previous value, 12, was not always sufficient for BN_mod_exp().
3008 *) Make sure that shared libraries get the internal name engine with
3009 the full version number and not just 0. This should mark the
3010 shared libraries as not backward compatible. Of course, this should
3011 be changed again when we can guarantee backward binary compatibility.
3014 *) Fix typo in get_cert_by_subject() in by_dir.c
3015 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3017 *) Rework the system to generate shared libraries:
3019 - Make note of the expected extension for the shared libraries and
3020 if there is a need for symbolic links from for example libcrypto.so.0
3021 to libcrypto.so.0.9.7. There is extended info in Configure for
3024 - Make as few rebuilds of the shared libraries as possible.
3026 - Still avoid linking the OpenSSL programs with the shared libraries.
3028 - When installing, install the shared libraries separately from the
3032 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3034 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3035 and not in SSL_clear because the latter is also used by the
3036 accept/connect functions; previously, the settings made by
3037 SSL_set_read_ahead would be lost during the handshake.
3038 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3040 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3041 Previously, it would create entries for disableed algorithms no
3045 *) Added several new manual pages for SSL_* function.
3048 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3050 *) In ssl23_get_client_hello, generate an error message when faced
3051 with an initial SSL 3.0/TLS record that is too small to contain the
3052 first two bytes of the ClientHello message, i.e. client_version.
3053 (Note that this is a pathologic case that probably has never happened
3054 in real life.) The previous approach was to use the version number
3055 from the record header as a substitute; but our protocol choice
3056 should not depend on that one because it is not authenticated
3057 by the Finished messages.
3060 *) More robust randomness gathering functions for Windows.
3061 [Jeffrey Altman <jaltman@columbia.edu>]
3063 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3064 not set then we don't setup the error code for issuer check errors
3065 to avoid possibly overwriting other errors which the callback does
3066 handle. If an application does set the flag then we assume it knows
3067 what it is doing and can handle the new informational codes
3071 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3072 a general "ANY" type, as such it should be able to decode anything
3073 including tagged types. However it didn't check the class so it would
3074 wrongly interpret tagged types in the same way as their universal
3075 counterpart and unknown types were just rejected. Changed so that the
3076 tagged and unknown types are handled in the same way as a SEQUENCE:
3077 that is the encoding is stored intact. There is also a new type
3078 "V_ASN1_OTHER" which is used when the class is not universal, in this
3079 case we have no idea what the actual type is so we just lump them all
3083 *) On VMS, stdout may very well lead to a file that is written to
3084 in a record-oriented fashion. That means that every write() will
3085 write a separate record, which will be read separately by the
3086 programs trying to read from it. This can be very confusing.
3088 The solution is to put a BIO filter in the way that will buffer
3089 text until a linefeed is reached, and then write everything a
3090 line at a time, so every record written will be an actual line,
3091 not chunks of lines and not (usually doesn't happen, but I've
3092 seen it once) several lines in one record. BIO_f_linebuffer() is
3095 Currently, it's a VMS-only method, because that's where it has
3096 been tested well enough.
3099 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3100 it can return incorrect results.
3101 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3102 but it was in 0.9.6-beta[12].)
3105 *) Disable the check for content being present when verifying detached
3106 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3107 include zero length content when signing messages.
3110 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3111 BIO_ctrl (for BIO pairs).
3114 *) Add DSO method for VMS.
3117 *) Bug fix: Montgomery multiplication could produce results with the
3121 *) Add RPM specification openssl.spec and modify it to build three
3122 packages. The default package contains applications, application
3123 documentation and run-time libraries. The devel package contains
3124 include files, static libraries and function documentation. The
3125 doc package contains the contents of the doc directory. The original
3126 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3129 *) Add a large number of documentation files for many SSL routines.
3130 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3132 *) Add a configuration entry for Sony News 4.
3133 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3135 *) Don't set the two most significant bits to one when generating a
3136 random number < q in the DSA library.
3139 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3140 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3141 the underlying transport is blocking) if a handshake took place.
3142 (The default behaviour is needed by applications such as s_client
3143 and s_server that use select() to determine when to use SSL_read;
3144 but for applications that know in advance when to expect data, it
3145 just makes things more complicated.)
3148 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3152 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3153 work better on such systems.
3154 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3156 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3157 Update PKCS12_parse() so it copies the friendlyName and the
3158 keyid to the certificates aux info.
3161 *) Fix bug in PKCS7_verify() which caused an infinite loop
3162 if there was more than one signature.
3163 [Sven Uszpelkat <su@celocom.de>]
3165 *) Major change in util/mkdef.pl to include extra information
3166 about each symbol, as well as presentig variables as well
3167 as functions. This change means that there's n more need
3168 to rebuild the .num files when some algorithms are excluded.
3171 *) Allow the verify time to be set by an application,
3172 rather than always using the current time.
3175 *) Phase 2 verify code reorganisation. The certificate
3176 verify code now looks up an issuer certificate by a
3177 number of criteria: subject name, authority key id
3178 and key usage. It also verifies self signed certificates
3179 by the same criteria. The main comparison function is
3180 X509_check_issued() which performs these checks.
3182 Lot of changes were necessary in order to support this
3183 without completely rewriting the lookup code.
3185 Authority and subject key identifier are now cached.
3187 The LHASH 'certs' is X509_STORE has now been replaced
3188 by a STACK_OF(X509_OBJECT). This is mainly because an
3189 LHASH can't store or retrieve multiple objects with
3190 the same hash value.
3192 As a result various functions (which were all internal
3193 use only) have changed to handle the new X509_STORE
3194 structure. This will break anything that messed round
3195 with X509_STORE internally.
3197 The functions X509_STORE_add_cert() now checks for an
3198 exact match, rather than just subject name.
3200 The X509_STORE API doesn't directly support the retrieval
3201 of multiple certificates matching a given criteria, however
3202 this can be worked round by performing a lookup first
3203 (which will fill the cache with candidate certificates)
3204 and then examining the cache for matches. This is probably
3205 the best we can do without throwing out X509_LOOKUP
3206 entirely (maybe later...).
3208 The X509_VERIFY_CTX structure has been enhanced considerably.
3210 All certificate lookup operations now go via a get_issuer()
3211 callback. Although this currently uses an X509_STORE it
3212 can be replaced by custom lookups. This is a simple way
3213 to bypass the X509_STORE hackery necessary to make this
3214 work and makes it possible to use more efficient techniques
3215 in future. A very simple version which uses a simple
3216 STACK for its trusted certificate store is also provided
3217 using X509_STORE_CTX_trusted_stack().
3219 The verify_cb() and verify() callbacks now have equivalents
3220 in the X509_STORE_CTX structure.
3222 X509_STORE_CTX also has a 'flags' field which can be used
3223 to customise the verify behaviour.
3226 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3227 excludes S/MIME capabilities.
3230 *) When a certificate request is read in keep a copy of the
3231 original encoding of the signed data and use it when outputing
3232 again. Signatures then use the original encoding rather than
3233 a decoded, encoded version which may cause problems if the
3234 request is improperly encoded.
3237 *) For consistency with other BIO_puts implementations, call
3238 buffer_write(b, ...) directly in buffer_puts instead of calling
3241 In BIO_puts, increment b->num_write as in BIO_write.
3242 [Peter.Sylvester@EdelWeb.fr]
3244 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3245 BN_zero, we may not return a BIGNUM with an array consisting of
3249 *) Avoid calling abort() from within the library when problems are
3250 detected, except if preprocessor symbols have been defined
3251 (such as REF_CHECK, BN_DEBUG etc.).
3254 *) New openssl application 'rsautl'. This utility can be
3255 used for low level RSA operations. DER public key
3256 BIO/fp routines also added.
3259 *) New Configure entry and patches for compiling on QNX 4.
3260 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3262 *) A demo state-machine implementation was sponsored by
3263 Nuron (http://www.nuron.com/) and is now available in
3264 demos/state_machine.
3267 *) New options added to the 'dgst' utility for signature
3268 generation and verification.
3271 *) Unrecognized PKCS#7 content types are now handled via a
3272 catch all ASN1_TYPE structure. This allows unsupported
3273 types to be stored as a "blob" and an application can
3274 encode and decode it manually.
3277 *) Fix various signed/unsigned issues to make a_strex.c
3279 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3281 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3282 length if passed a buffer. ASN1_INTEGER_to_BN failed
3283 if passed a NULL BN and its argument was negative.
3284 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3286 *) Modification to PKCS#7 encoding routines to output definite
3287 length encoding. Since currently the whole structures are in
3288 memory there's not real point in using indefinite length
3289 constructed encoding. However if OpenSSL is compiled with
3290 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3293 *) Added BIO_vprintf() and BIO_vsnprintf().
3296 *) Added more prefixes to parse for in the the strings written
3297 through a logging bio, to cover all the levels that are available
3298 through syslog. The prefixes are now:
3300 PANIC, EMERG, EMR => LOG_EMERG
3301 ALERT, ALR => LOG_ALERT
3302 CRIT, CRI => LOG_CRIT
3303 ERROR, ERR => LOG_ERR
3304 WARNING, WARN, WAR => LOG_WARNING
3305 NOTICE, NOTE, NOT => LOG_NOTICE
3306 INFO, INF => LOG_INFO
3307 DEBUG, DBG => LOG_DEBUG
3309 and as before, if none of those prefixes are present at the
3310 beginning of the string, LOG_ERR is chosen.
3312 On Win32, the LOG_* levels are mapped according to this:
3314 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3315 LOG_WARNING => EVENTLOG_WARNING_TYPE
3316 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3320 *) Made it possible to reconfigure with just the configuration
3321 argument "reconf" or "reconfigure". The command line arguments
3322 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3323 and are retrieved from there when reconfiguring.
3327 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3329 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3332 *) The obj_dat.pl script was messing up the sorting of object
3333 names. The reason was that it compared the quoted version
3334 of strings as a result "OCSP" > "OCSP Signing" because
3335 " > SPACE. Changed script to store unquoted versions of
3336 names and add quotes on output. It was also omitting some
3337 names from the lookup table if they were given a default
3338 value (that is if SN is missing it is given the same
3339 value as LN and vice versa), these are now added on the
3340 grounds that if an object has a name we should be able to
3341 look it up. Finally added warning output when duplicate
3342 short or long names are found.
3345 *) Changes needed for Tandem NSK.
3346 [Scott Uroff <scott@xypro.com>]
3348 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3349 RSA_padding_check_SSLv23(), special padding was never detected
3350 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3351 version rollback attacks was not effective.
3353 In s23_clnt.c, don't use special rollback-attack detection padding
3354 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3355 client; similarly, in s23_srvr.c, don't do the rollback check if
3356 SSL 2.0 is the only protocol enabled in the server.
3359 *) Make it possible to get hexdumps of unprintable data with 'openssl
3360 asn1parse'. By implication, the functions ASN1_parse_dump() and
3361 BIO_dump_indent() are added.
3364 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3365 these print out strings and name structures based on various
3366 flags including RFC2253 support and proper handling of
3367 multibyte characters. Added options to the 'x509' utility
3368 to allow the various flags to be set.
3371 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3372 Also change the functions X509_cmp_current_time() and
3373 X509_gmtime_adj() work with an ASN1_TIME structure,
3374 this will enable certificates using GeneralizedTime in validity
3375 dates to be checked.
3378 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3379 negative public key encodings) on by default,
3380 NO_NEG_PUBKEY_BUG can be set to disable it.
3383 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3384 content octets. An i2c_ASN1_OBJECT is unnecessary because
3385 the encoding can be trivially obtained from the structure.
3388 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3389 not read locks (CRYPTO_r_[un]lock).
3392 *) A first attempt at creating official support for shared
3393 libraries through configuration. I've kept it so the
3394 default is static libraries only, and the OpenSSL programs
3395 are always statically linked for now, but there are
3396 preparations for dynamic linking in place.
3397 This has been tested on Linux and Tru64.
3400 *) Randomness polling function for Win9x, as described in:
3401 Peter Gutmann, Software Generation of Practically Strong
3405 *) Fix so PRNG is seeded in req if using an already existing
3409 *) New options to smime application. -inform and -outform
3410 allow alternative formats for the S/MIME message including
3411 PEM and DER. The -content option allows the content to be
3412 specified separately. This should allow things like Netscape
3413 form signing output easier to verify.
3416 *) Fix the ASN1 encoding of tags using the 'long form'.
3419 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3420 STRING types. These convert content octets to and from the
3421 underlying type. The actual tag and length octets are
3422 already assumed to have been read in and checked. These
3423 are needed because all other string types have virtually
3424 identical handling apart from the tag. By having versions
3425 of the ASN1 functions that just operate on content octets
3426 IMPLICIT tagging can be handled properly. It also allows
3427 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3428 and ASN1_INTEGER are identical apart from the tag.
3431 *) Change the handling of OID objects as follows:
3433 - New object identifiers are inserted in objects.txt, following
3434 the syntax given in objects.README.
3435 - objects.pl is used to process obj_mac.num and create a new
3437 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3440 This is currently kind of a hack, and the perl code in objects.pl
3441 isn't very elegant, but it works as I intended. The simplest way
3442 to check that it worked correctly is to look in obj_dat.h and
3443 check the array nid_objs and make sure the objects haven't moved
3444 around (this is important!). Additions are OK, as well as
3445 consistent name changes.
3448 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3451 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3452 The given file adds to whatever has already been seeded into the
3453 random pool through the RANDFILE configuration file option or
3454 environment variable, or the default random state file.
3457 *) mkstack.pl now sorts each macro group into lexical order.
3458 Previously the output order depended on the order the files
3459 appeared in the directory, resulting in needless rewriting
3463 *) Patches to make OpenSSL compile under Win32 again. Mostly
3464 work arounds for the VC++ problem that it treats func() as
3465 func(void). Also stripped out the parts of mkdef.pl that
3466 added extra typesafe functions: these no longer exist.
3469 *) Reorganisation of the stack code. The macros are now all
3470 collected in safestack.h . Each macro is defined in terms of
3471 a "stack macro" of the form SKM_<name>(type, a, b). The
3472 DEBUG_SAFESTACK is now handled in terms of function casts,
3473 this has the advantage of retaining type safety without the
3474 use of additional functions. If DEBUG_SAFESTACK is not defined
3475 then the non typesafe macros are used instead. Also modified the
3476 mkstack.pl script to handle the new form. Needs testing to see
3477 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3478 the default if no major problems. Similar behaviour for ASN1_SET_OF
3479 and PKCS12_STACK_OF.
3482 *) When some versions of IIS use the 'NET' form of private key the
3483 key derivation algorithm is different. Normally MD5(password) is
3484 used as a 128 bit RC4 key. In the modified case
3485 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3486 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3487 as the old Netscape_RSA functions except they have an additional
3488 'sgckey' parameter which uses the modified algorithm. Also added
3489 an -sgckey command line option to the rsa utility. Thanks to
3490 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3491 algorithm to openssl-dev.
3494 *) The evp_local.h macros were using 'c.##kname' which resulted in
3495 invalid expansion on some systems (SCO 5.0.5 for example).
3496 Corrected to 'c.kname'.
3497 [Phillip Porch <root@theporch.com>]
3499 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3500 a STACK of email addresses from a certificate or request, these look
3501 in the subject name and the subject alternative name extensions and
3502 omit any duplicate addresses.
3505 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3506 This makes DSA verification about 2 % faster.
3509 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3510 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3511 plus overhead for 1024 bit moduli).
3512 This makes exponentiations about 0.5 % faster for 1024 bit
3513 exponents (as measured by "openssl speed rsa2048").
3516 *) Rename memory handling macros to avoid conflicts with other
3518 Malloc => OPENSSL_malloc
3519 Malloc_locked => OPENSSL_malloc_locked
3520 Realloc => OPENSSL_realloc
3521 Free => OPENSSL_free
3524 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3525 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3528 *) CygWin32 support.
3529 [John Jarvie <jjarvie@newsguy.com>]
3531 *) The type-safe stack code has been rejigged. It is now only compiled
3532 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3533 by default all type-specific stack functions are "#define"d back to
3534 standard stack functions. This results in more streamlined output
3535 but retains the type-safety checking possibilities of the original
3539 *) The STACK code has been cleaned up, and certain type declarations
3540 that didn't make a lot of sense have been brought in line. This has
3541 also involved a cleanup of sorts in safestack.h to more correctly
3542 map type-safe stack functions onto their plain stack counterparts.
3543 This work has also resulted in a variety of "const"ifications of
3544 lots of the code, especially "_cmp" operations which should normally
3545 be prototyped with "const" parameters anyway.
3548 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3549 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3550 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3551 where all of 'md' is used each time the PRNG is used, but 'state'
3552 is used only indexed by a cyclic counter. As entropy may not be
3553 well distributed from the beginning, 'md' is important as a
3554 chaining variable. However, the output function chains only half
3555 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3556 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3557 in all of 'state' being rewritten, with the new values depending
3558 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3561 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3562 the handshake is continued after ssl_verify_cert_chain();
3563 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3564 can lead to 'unexplainable' connection aborts later.
3565 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3567 *) Major EVP API cipher revision.
3568 Add hooks for extra EVP features. This allows various cipher
3569 parameters to be set in the EVP interface. Support added for variable
3570 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3571 setting of RC2 and RC5 parameters.
3573 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3576 Remove lots of duplicated code from the EVP library. For example *every*
3577 cipher init() function handles the 'iv' in the same way according to the
3578 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3579 for CFB and OFB modes they zero ctx->num.
3581 New functionality allows removal of S/MIME code RC2 hack.
3583 Most of the routines have the same form and so can be declared in terms
3586 By shifting this to the top level EVP_CipherInit() it can be removed from
3587 all individual ciphers. If the cipher wants to handle IVs or keys
3588 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3591 Change lots of functions like EVP_EncryptUpdate() to now return a
3592 value: although software versions of the algorithms cannot fail
3593 any installed hardware versions can.
3596 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3597 this option is set, tolerate broken clients that send the negotiated
3598 protocol version number instead of the requested protocol version
3602 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3603 i.e. non-zero for export ciphersuites, zero otherwise.
3604 Previous versions had this flag inverted, inconsistent with
3605 rsa_tmp_cb (..._TMP_RSA_CB).
3606 [Bodo Moeller; problem reported by Amit Chopra]
3608 *) Add missing DSA library text string. Work around for some IIS
3609 key files with invalid SEQUENCE encoding.
3612 *) Add a document (doc/standards.txt) that list all kinds of standards
3613 and so on that are implemented in OpenSSL.
3616 *) Enhance c_rehash script. Old version would mishandle certificates
3617 with the same subject name hash and wouldn't handle CRLs at all.
3618 Added -fingerprint option to crl utility, to support new c_rehash
3622 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3625 *) Fix for SSL server purpose checking. Server checking was
3626 rejecting certificates which had extended key usage present
3627 but no ssl client purpose.
3628 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3630 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3631 is a little unclear about how a blank password is handled.
3632 Since the password in encoded as a BMPString with terminating
3633 double NULL a zero length password would end up as just the
3634 double NULL. However no password at all is different and is
3635 handled differently in the PKCS#12 key generation code. NS
3636 treats a blank password as zero length. MSIE treats it as no
3637 password on export: but it will try both on import. We now do
3638 the same: PKCS12_parse() tries zero length and no password if
3639 the password is set to "" or NULL (NULL is now a valid password:
3640 it wasn't before) as does the pkcs12 application.
3643 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3644 perror when PEM_read_bio_X509_REQ fails, the error message must
3645 be obtained from the error queue.
3648 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3649 it in ERR_remove_state if appropriate, and change ERR_get_state
3650 accordingly to avoid race conditions (this is necessary because
3651 thread_hash is no longer constant once set).
3654 *) Bugfix for linux-elf makefile.one.
3657 *) RSA_get_default_method() will now cause a default
3658 RSA_METHOD to be chosen if one doesn't exist already.
3659 Previously this was only set during a call to RSA_new()
3660 or RSA_new_method(NULL) meaning it was possible for
3661 RSA_get_default_method() to return NULL.
3664 *) Added native name translation to the existing DSO code
3665 that will convert (if the flag to do so is set) filenames
3666 that are sufficiently small and have no path information
3667 into a canonical native form. Eg. "blah" converted to
3668 "libblah.so" or "blah.dll" etc.
3671 *) New function ERR_error_string_n(e, buf, len) which is like
3672 ERR_error_string(e, buf), but writes at most 'len' bytes
3673 including the 0 terminator. For ERR_error_string_n, 'buf'
3675 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3677 *) CONF library reworked to become more general. A new CONF
3678 configuration file reader "class" is implemented as well as a
3679 new functions (NCONF_*, for "New CONF") to handle it. The now
3680 old CONF_* functions are still there, but are reimplemented to
3681 work in terms of the new functions. Also, a set of functions
3682 to handle the internal storage of the configuration data is
3683 provided to make it easier to write new configuration file
3684 reader "classes" (I can definitely see something reading a
3685 configuration file in XML format, for example), called _CONF_*,
3686 or "the configuration storage API"...
3688 The new configuration file reading functions are:
3690 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3691 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3693 NCONF_default, NCONF_WIN32
3695 NCONF_dump_fp, NCONF_dump_bio
3697 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3698 NCONF_new creates a new CONF object. This works in the same way
3699 as other interfaces in OpenSSL, like the BIO interface.
3700 NCONF_dump_* dump the internal storage of the configuration file,
3701 which is useful for debugging. All other functions take the same
3702 arguments as the old CONF_* functions wth the exception of the
3703 first that must be a `CONF *' instead of a `LHASH *'.
3705 To make it easer to use the new classes with the old CONF_* functions,
3706 the function CONF_set_default_method is provided.
3709 *) Add '-tls1' option to 'openssl ciphers', which was already
3710 mentioned in the documentation but had not been implemented.
3711 (This option is not yet really useful because even the additional
3712 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3715 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3716 OpenSSL-based applications) load shared libraries and bind to
3717 them in a portable way.
3718 [Geoff Thorpe, with contributions from Richard Levitte]
3720 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3722 *) Make sure _lrotl and _lrotr are only used with MSVC.
3724 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3725 (the default implementation of RAND_status).
3727 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3728 to '-clrext' (= clear extensions), as intended and documented.
3729 [Bodo Moeller; inconsistency pointed out by Michael Attili
3732 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3733 was larger than the MD block size.
3734 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3736 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3737 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3738 using the passed key: if the passed key was a private key the result
3739 of X509_print(), for example, would be to print out all the private key
3743 *) des_quad_cksum() byte order bug fix.
3744 [Ulf Möller, using the problem description in krb4-0.9.7, where
3745 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3747 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3749 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3751 *) For easily testing in shell scripts whether some command
3752 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3753 returns with exit code 0 iff no command of the given name is available.
3754 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3755 the output goes to stdout and nothing is printed to stderr.
3756 Additional arguments are always ignored.
3758 Since for each cipher there is a command of the same name,
3759 the 'no-cipher' compilation switches can be tested this way.
3761 ('openssl no-XXX' is not able to detect pseudo-commands such
3762 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3765 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3768 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3769 is set; it will be thrown away anyway because each handshake creates
3771 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3772 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3773 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3774 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3777 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3778 'Q' and 'R' lose their special meanings (quit/renegotiate).
3779 This is part of what -quiet does; unlike -quiet, -ign_eof
3780 does not suppress any output.
3783 *) Add compatibility options to the purpose and trust code. The
3784 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3785 accepts a certificate or CA, this was the previous behaviour,
3786 with all the associated security issues.
3788 X509_TRUST_COMPAT is the old trust behaviour: only and
3789 automatically trust self signed roots in certificate store. A
3790 new trust setting X509_TRUST_DEFAULT is used to specify that
3791 a purpose has no associated trust setting and it should instead
3792 use the value in the default purpose.
3795 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3796 and fix a memory leak.
3799 *) In util/mkerr.pl (which implements 'make errors'), preserve
3800 reason strings from the previous version of the .c file, as
3801 the default to have only downcase letters (and digits) in
3802 automatically generated reasons codes is not always appropriate.
3805 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3806 using strerror. Previously, ERR_reason_error_string() returned
3807 library names as reason strings for SYSerr; but SYSerr is a special
3808 case where small numbers are errno values, not library numbers.
3811 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3812 converts DSA parameters into DH parameters. (When creating parameters,
3813 DSA_generate_parameters is used.)
3816 *) Include 'length' (recommended exponent length) in C code generated
3817 by 'openssl dhparam -C'.
3820 *) The second argument to set_label in perlasm was already being used
3821 so couldn't be used as a "file scope" flag. Moved to third argument
3825 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3826 instead of RAND_bytes for encryption IVs and salts.
3829 *) Include RAND_status() into RAND_METHOD instead of implementing
3830 it only for md_rand.c Otherwise replacing the PRNG by calling
3831 RAND_set_rand_method would be impossible.
3834 *) Don't let DSA_generate_key() enter an infinite loop if the random
3835 number generation fails.
3838 *) New 'rand' application for creating pseudo-random output.
3841 *) Added configuration support for Linux/IA64
3842 [Rolf Haberrecker <rolf@suse.de>]
3844 *) Assembler module support for Mingw32.
3847 *) Shared library support for HPUX (in shlib/).
3848 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3850 *) Shared library support for Solaris gcc.
3851 [Lutz Behnke <behnke@trustcenter.de>]
3853 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3855 *) PKCS7_encrypt() was adding text MIME headers twice because they
3856 were added manually and by SMIME_crlf_copy().
3859 *) In bntest.c don't call BN_rand with zero bits argument.
3860 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3862 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3863 case was implemented. This caused BN_div_recp() to fail occasionally.
3866 *) Add an optional second argument to the set_label() in the perl
3867 assembly language builder. If this argument exists and is set
3868 to 1 it signals that the assembler should use a symbol whose
3869 scope is the entire file, not just the current function. This
3870 is needed with MASM which uses the format label:: for this scope.
3871 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3873 *) Change the ASN1 types so they are typedefs by default. Before
3874 almost all types were #define'd to ASN1_STRING which was causing
3875 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3879 *) Change names of new functions to the new get1/get0 naming
3880 convention: After 'get1', the caller owns a reference count
3881 and has to call ..._free; 'get0' returns a pointer to some
3882 data structure without incrementing reference counters.
3883 (Some of the existing 'get' functions increment a reference
3884 counter, some don't.)
3885 Similarly, 'set1' and 'add1' functions increase reference
3886 counters or duplicate objects.
3889 *) Allow for the possibility of temp RSA key generation failure:
3890 the code used to assume it always worked and crashed on failure.
3893 *) Fix potential buffer overrun problem in BIO_printf().
3894 [Ulf Möller, using public domain code by Patrick Powell; problem
3895 pointed out by David Sacerdote <das33@cornell.edu>]
3897 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3898 RAND_egd() and RAND_status(). In the command line application,
3899 the EGD socket can be specified like a seed file using RANDFILE
3903 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3904 Some CAs (e.g. Verisign) distribute certificates in this form.
3907 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3908 list to exclude them. This means that no special compilation option
3909 is needed to use anonymous DH: it just needs to be included in the
3913 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3914 EVP_MD_type. The old functionality is available in a new macro called
3915 EVP_MD_md(). Change code that uses it and update docs.
3918 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3919 where the 'void *' argument is replaced by a function pointer argument.
3920 Previously 'void *' was abused to point to functions, which works on
3921 many platforms, but is not correct. As these functions are usually
3922 called by macros defined in OpenSSL header files, most source code
3923 should work without changes.
3926 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3927 sections with information on -D... compiler switches used for
3928 compiling the library so that applications can see them. To enable
3929 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3930 must be defined. E.g.,
3931 #define OPENSSL_ALGORITHM_DEFINES
3932 #include <openssl/opensslconf.h>
3933 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3934 [Richard Levitte, Ulf and Bodo Möller]
3936 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3940 *) Change the 'other' type in certificate aux info to a STACK_OF
3941 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3942 the required ASN1 format: arbitrary types determined by an OID.
3945 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3946 argument to 'req'. This is not because the function is newer or
3947 better than others it just uses the work 'NEW' in the certificate
3948 request header lines. Some software needs this.
3951 *) Reorganise password command line arguments: now passwords can be
3952 obtained from various sources. Delete the PEM_cb function and make
3953 it the default behaviour: i.e. if the callback is NULL and the
3954 usrdata argument is not NULL interpret it as a null terminated pass
3955 phrase. If usrdata and the callback are NULL then the pass phrase
3956 is prompted for as usual.
3959 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3960 the support is automatically enabled. The resulting binaries will
3961 autodetect the card and use it if present.
3962 [Ben Laurie and Compaq Inc.]
3964 *) Work around for Netscape hang bug. This sends certificate request
3965 and server done in one record. Since this is perfectly legal in the
3966 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3967 the bugs/SSLv3 entry for more info.
3970 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3973 *) Add -rand argument to smime and pkcs12 applications and read/write
3977 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3980 *) Add command line password options to the remaining applications.
3983 *) Bug fix for BN_div_recp() for numerators with an even number of
3987 *) More tests in bntest.c, and changed test_bn output.
3990 *) ./config recognizes MacOS X now.
3993 *) Bug fix for BN_div() when the first words of num and divsor are
3994 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3997 *) Add support for various broken PKCS#8 formats, and command line
3998 options to produce them.
4001 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4002 get temporary BIGNUMs from a BN_CTX.
4005 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4009 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4010 include a #define from the old name to the new. The original intent
4011 was that statically linked binaries could for example just call
4012 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4013 link with digests. This never worked becayse SSLeay_add_all_digests()
4014 and SSLeay_add_all_ciphers() were in the same source file so calling
4015 one would link with the other. They are now in separate source files.
4018 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4021 *) Use a less unusual form of the Miller-Rabin primality test (it used
4022 a binary algorithm for exponentiation integrated into the Miller-Rabin
4023 loop, our standard modexp algorithms are faster).
4026 *) Support for the EBCDIC character set completed.
4027 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4029 *) Source code cleanups: use const where appropriate, eliminate casts,
4030 use void * instead of char * in lhash.
4033 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4034 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4035 this the server could overwrite ephemeral keys that the client
4039 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4040 using 50 iterations of the Rabin-Miller test.
4042 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4043 iterations of the Rabin-Miller test as required by the appendix
4044 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4045 As BN_is_prime_fasttest includes trial division, DSA parameter
4046 generation becomes much faster.
4048 This implies a change for the callback functions in DSA_is_prime
4049 and DSA_generate_parameters: The callback function is called once
4050 for each positive witness in the Rabin-Miller test, not just
4051 occasionally in the inner loop; and the parameters to the
4052 callback function now provide an iteration count for the outer
4053 loop rather than for the current invocation of the inner loop.
4054 DSA_generate_parameters additionally can call the callback
4055 function with an 'iteration count' of -1, meaning that a
4056 candidate has passed the trial division test (when q is generated
4057 from an application-provided seed, trial division is skipped).
4060 *) New function BN_is_prime_fasttest that optionally does trial
4061 division before starting the Rabin-Miller test and has
4062 an additional BN_CTX * argument (whereas BN_is_prime always
4063 has to allocate at least one BN_CTX).
4064 'callback(1, -1, cb_arg)' is called when a number has passed the
4065 trial division stage.
4068 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4072 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4075 *) New function BN_pseudo_rand().
4078 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4079 bignum version of BN_from_montgomery() with the working code from
4080 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4084 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4085 made it impossible to use the same SSL_SESSION data structure in
4086 SSL2 clients in multiple threads.
4089 *) The return value of RAND_load_file() no longer counts bytes obtained
4090 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4091 to seed the PRNG (previously an explicit byte count was required).
4092 [Ulf Möller, Bodo Möller]
4094 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4095 used (char *) instead of (void *) and had casts all over the place.
4098 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4101 *) Retain source code compatibility for BN_prime_checks macro:
4102 BN_is_prime(..., BN_prime_checks, ...) now uses
4103 BN_prime_checks_for_size to determine the appropriate number of
4104 Rabin-Miller iterations.
4107 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4108 DH_CHECK_P_NOT_SAFE_PRIME.
4109 (Check if this is true? OpenPGP calls them "strong".)
4112 *) Merge the functionality of "dh" and "gendh" programs into a new program
4113 "dhparam". The old programs are retained for now but will handle DH keys
4114 (instead of parameters) in future.
4117 *) Make the ciphers, s_server and s_client programs check the return values
4118 when a new cipher list is set.
4121 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4122 ciphers. Before when the 56bit ciphers were enabled the sorting was
4125 The syntax for the cipher sorting has been extended to support sorting by
4126 cipher-strength (using the strength_bits hard coded in the tables).
4127 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4129 Fix a bug in the cipher-command parser: when supplying a cipher command
4130 string with an "undefined" symbol (neither command nor alphanumeric
4131 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4132 an error is flagged.
4134 Due to the strength-sorting extension, the code of the
4135 ssl_create_cipher_list() function was completely rearranged. I hope that
4136 the readability was also increased :-)
4137 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4139 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4140 for the first serial number and places 2 in the serial number file. This
4141 avoids problems when the root CA is created with serial number zero and
4142 the first user certificate has the same issuer name and serial number
4146 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4147 the new code. Add documentation for this stuff.
4150 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4151 X509_*() to X509at_*() on the grounds that they don't handle X509
4152 structures and behave in an analagous way to the X509v3 functions:
4153 they shouldn't be called directly but wrapper functions should be used
4156 So we also now have some wrapper functions that call the X509at functions
4157 when passed certificate requests. (TO DO: similar things can be done with
4158 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4159 things. Some of these need some d2i or i2d and print functionality
4160 because they handle more complex structures.)
4163 *) Add missing #ifndefs that caused missing symbols when building libssl
4164 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4165 NO_RSA in ssl/s2*.c.
4166 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4168 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4169 has a return value which indicates the quality of the random data
4170 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4171 error queue. New function RAND_pseudo_bytes() generates output that is
4172 guaranteed to be unique but not unpredictable. RAND_add is like
4173 RAND_seed, but takes an extra argument for an entropy estimate
4174 (RAND_seed always assumes full entropy).
4177 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4178 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4179 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4180 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4181 false-positive rate of at most 2^-80 for random input.
4184 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4187 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4188 in the 0.9.5 release), this returns the chain
4189 from an X509_CTX structure with a dup of the stack and all
4190 the X509 reference counts upped: so the stack will exist
4191 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4194 Also make SSL_SESSION_print() print out the verify return
4198 *) Add manpage for the pkcs12 command. Also change the default
4199 behaviour so MAC iteration counts are used unless the new
4200 -nomaciter option is used. This improves file security and
4201 only older versions of MSIE (4.0 for example) need it.
4204 *) Honor the no-xxx Configure options when creating .DEF files.
4207 *) Add PKCS#10 attributes to field table: challengePassword,
4208 unstructuredName and unstructuredAddress. These are taken from
4209 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4210 international characters are used.
4212 More changes to X509_ATTRIBUTE code: allow the setting of types
4213 based on strings. Remove the 'loc' parameter when adding
4214 attributes because these will be a SET OF encoding which is sorted
4218 *) Initial changes to the 'req' utility to allow request generation
4219 automation. This will allow an application to just generate a template
4220 file containing all the field values and have req construct the
4223 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4224 used all over the place including certificate requests and PKCS#7
4225 structures. They are currently handled manually where necessary with
4226 some primitive wrappers for PKCS#7. The new functions behave in a
4227 manner analogous to the X509 extension functions: they allow
4228 attributes to be looked up by NID and added.
4230 Later something similar to the X509V3 code would be desirable to
4231 automatically handle the encoding, decoding and printing of the
4232 more complex types. The string types like challengePassword can
4233 be handled by the string table functions.
4235 Also modified the multi byte string table handling. Now there is
4236 a 'global mask' which masks out certain types. The table itself
4237 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4238 is useful when for example there is only one permissible type
4239 (as in countryName) and using the mask might result in no valid
4243 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4244 SSL_get_peer_finished to allow applications to obtain the latest
4245 Finished messages sent to the peer or expected from the peer,
4246 respectively. (SSL_get_peer_finished is usually the Finished message
4247 actually received from the peer, otherwise the protocol will be aborted.)
4249 As the Finished message are message digests of the complete handshake
4250 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4251 be used for external authentication procedures when the authentication
4252 provided by SSL/TLS is not desired or is not enough.
4255 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4256 the host supports BWX extension and if Compaq C is present on the
4257 $PATH. Just exploiting of the BWX extension results in 20-30%
4258 performance kick for some algorithms, e.g. DES and RC4 to mention
4259 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4263 *) Add support for MS "fast SGC". This is arguably a violation of the
4264 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4265 weak crypto and after checking the certificate is SGC a second one
4266 with strong crypto. MS SGC stops the first handshake after receiving
4267 the server certificate message and sends a second client hello. Since
4268 a server will typically do all the time consuming operations before
4269 expecting any further messages from the client (server key exchange
4270 is the most expensive) there is little difference between the two.
4272 To get OpenSSL to support MS SGC we have to permit a second client
4273 hello message after we have sent server done. In addition we have to
4274 reset the MAC if we do get this second client hello.
4277 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4278 if a DER encoded private key is RSA or DSA traditional format. Changed
4279 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4280 format DER encoded private key. Newer code should use PKCS#8 format which
4281 has the key type encoded in the ASN1 structure. Added DER private key
4282 support to pkcs8 application.
4285 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4286 ciphersuites has been selected (as required by the SSL 3/TLS 1
4287 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4288 is set, we interpret this as a request to violate the specification
4289 (the worst that can happen is a handshake failure, and 'correct'
4290 behaviour would result in a handshake failure anyway).
4293 *) In SSL_CTX_add_session, take into account that there might be multiple
4294 SSL_SESSION structures with the same session ID (e.g. when two threads
4295 concurrently obtain them from an external cache).
4296 The internal cache can handle only one SSL_SESSION with a given ID,
4297 so if there's a conflict, we now throw out the old one to achieve
4301 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4302 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4303 some routines that use cipher OIDs: some ciphers do not have OIDs
4304 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4308 *) Simplify the trust setting structure and code. Now we just have
4309 two sequences of OIDs for trusted and rejected settings. These will
4310 typically have values the same as the extended key usage extension
4311 and any application specific purposes.
4313 The trust checking code now has a default behaviour: it will just
4314 check for an object with the same NID as the passed id. Functions can
4315 be provided to override either the default behaviour or the behaviour
4316 for a given id. SSL client, server and email already have functions
4317 in place for compatibility: they check the NID and also return "trusted"
4318 if the certificate is self signed.
4321 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4322 traditional format into an EVP_PKEY structure.
4325 *) Add a password callback function PEM_cb() which either prompts for
4326 a password if usr_data is NULL or otherwise assumes it is a null
4327 terminated password. Allow passwords to be passed on command line
4328 environment or config files in a few more utilities.
4331 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4332 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4333 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4334 Update documentation.
4337 *) Support for ASN1 "NULL" type. This could be handled before by using
4338 ASN1_TYPE but there wasn't any function that would try to read a NULL
4339 and produce an error if it couldn't. For compatibility we also have
4340 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4341 don't allocate anything because they don't need to.
4344 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4346 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4348 *) Rebuild of the memory allocation routines used by OpenSSL code and
4349 possibly others as well. The purpose is to make an interface that
4350 provide hooks so anyone can build a separate set of allocation and
4351 deallocation routines to be used by OpenSSL, for example memory
4352 pool implementations, or something else, which was previously hard
4353 since Malloc(), Realloc() and Free() were defined as macros having
4354 the values malloc, realloc and free, respectively (except for Win32
4355 compilations). The same is provided for memory debugging code.
4356 OpenSSL already comes with functionality to find memory leaks, but
4357 this gives people a chance to debug other memory problems.
4359 With these changes, a new set of functions and macros have appeared:
4361 CRYPTO_set_mem_debug_functions() [F]
4362 CRYPTO_get_mem_debug_functions() [F]
4363 CRYPTO_dbg_set_options() [F]
4364 CRYPTO_dbg_get_options() [F]
4365 CRYPTO_malloc_debug_init() [M]
4367 The memory debug functions are NULL by default, unless the library
4368 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4369 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4370 gives the standard debugging functions that come with OpenSSL) or
4371 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4372 provided by the library user) must be used. When the standard
4373 debugging functions are used, CRYPTO_dbg_set_options can be used to
4374 request additional information:
4375 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4376 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4378 Also, things like CRYPTO_set_mem_functions will always give the
4379 expected result (the new set of functions is used for allocation
4380 and deallocation) at all times, regardless of platform and compiler
4383 To finish it up, some functions that were never use in any other
4384 way than through macros have a new API and new semantic:
4387 CRYPTO_dbg_realloc()
4390 All macros of value have retained their old syntax.
4391 [Richard Levitte and Bodo Moeller]
4393 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4394 ordering of SMIMECapabilities wasn't in "strength order" and there
4395 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4399 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4400 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4401 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4403 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4404 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4405 functionality to handle multipart/signed properly) and a utility
4406 called 'smime' to call all this stuff. This is based on code I
4407 originally wrote for Celo who have kindly allowed it to be
4408 included in OpenSSL.
4411 *) Add variants des_set_key_checked and des_set_key_unchecked of
4412 des_set_key (aka des_key_sched). Global variable des_check_key
4413 decides which of these is called by des_set_key; this way
4414 des_check_key behaves as it always did, but applications and
4415 the library itself, which was buggy for des_check_key == 1,
4416 have a cleaner way to pick the version they need.
4419 *) New function PKCS12_newpass() which changes the password of a
4423 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4424 dynamic mix. In both cases the ids can be used as an index into the
4425 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4426 functions so they accept a list of the field values and the
4427 application doesn't need to directly manipulate the X509_TRUST
4431 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4435 *) Modify the way the V3 extension code looks up extensions. This now
4436 works in a similar way to the object code: we have some "standard"
4437 extensions in a static table which is searched with OBJ_bsearch()
4438 and the application can add dynamic ones if needed. The file
4439 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4440 updated whenever a new extension is added to the core code and kept
4441 in ext_nid order. There is a simple program 'tabtest.c' which checks
4442 this. New extensions are not added too often so this file can readily
4443 be maintained manually.
4445 There are two big advantages in doing things this way. The extensions
4446 can be looked up immediately and no longer need to be "added" using
4447 X509V3_add_standard_extensions(): this function now does nothing.
4448 [Side note: I get *lots* of email saying the extension code doesn't
4449 work because people forget to call this function]
4450 Also no dynamic allocation is done unless new extensions are added:
4451 so if we don't add custom extensions there is no need to call
4452 X509V3_EXT_cleanup().
4455 *) Modify enc utility's salting as follows: make salting the default. Add a
4456 magic header, so unsalted files fail gracefully instead of just decrypting
4457 to garbage. This is because not salting is a big security hole, so people
4458 should be discouraged from doing it.
4461 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4462 digest to be passed on the command line but it only used this
4463 parameter when signing a certificate. Modified so all relevant
4464 operations are affected by the digest parameter including the
4465 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4466 DSA key was used because it didn't fix the digest.
4469 *) Initial certificate chain verify code. Currently tests the untrusted
4470 certificates for consistency with the verify purpose (which is set
4471 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4473 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4474 this is because it will reject chains with invalid extensions whereas
4475 every previous version of OpenSSL and SSLeay made no checks at all.
4477 Trust code: checks the root CA for the relevant trust settings. Trust
4478 settings have an initial value consistent with the verify purpose: e.g.
4479 if the verify purpose is for SSL client use it expects the CA to be
4480 trusted for SSL client use. However the default value can be changed to
4481 permit custom trust settings: one example of this would be to only trust
4482 certificates from a specific "secure" set of CAs.
4484 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4485 which should be used for version portability: especially since the
4486 verify structure is likely to change more often now.
4488 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4489 to set them. If not set then assume SSL clients will verify SSL servers
4492 Two new options to the verify program: -untrusted allows a set of
4493 untrusted certificates to be passed in and -purpose which sets the
4494 intended purpose of the certificate. If a purpose is set then the
4495 new chain verify code is used to check extension consistency.
4498 *) Support for the authority information access extension.
4501 *) Modify RSA and DSA PEM read routines to transparently handle
4502 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4503 public keys in a format compatible with certificate
4504 SubjectPublicKeyInfo structures. Unfortunately there were already
4505 functions called *_PublicKey_* which used various odd formats so
4506 these are retained for compatibility: however the DSA variants were
4507 never in a public release so they have been deleted. Changed dsa/rsa
4508 utilities to handle the new format: note no releases ever handled public
4509 keys so we should be OK.
4511 The primary motivation for this change is to avoid the same fiasco
4512 that dogs private keys: there are several incompatible private key
4513 formats some of which are standard and some OpenSSL specific and
4514 require various evil hacks to allow partial transparent handling and
4515 even then it doesn't work with DER formats. Given the option anything
4516 other than PKCS#8 should be dumped: but the other formats have to
4517 stay in the name of compatibility.
4519 With public keys and the benefit of hindsight one standard format
4520 is used which works with EVP_PKEY, RSA or DSA structures: though
4521 it clearly returns an error if you try to read the wrong kind of key.
4523 Added a -pubkey option to the 'x509' utility to output the public key.
4524 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4525 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4526 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4527 that do the same as the EVP_PKEY_assign_*() except they up the
4528 reference count of the added key (they don't "swallow" the
4532 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4533 CRLs would fail if the file contained no certificates or no CRLs:
4534 added a new function to read in both types and return the number
4535 read: this means that if none are read it will be an error. The
4536 DER versions of the certificate and CRL reader would always fail
4537 because it isn't possible to mix certificates and CRLs in DER format
4538 without choking one or the other routine. Changed this to just read
4539 a certificate: this is the best we can do. Also modified the code
4540 in apps/verify.c to take notice of return codes: it was previously
4541 attempting to read in certificates from NULL pointers and ignoring
4542 any errors: this is one reason why the cert and CRL reader seemed
4543 to work. It doesn't check return codes from the default certificate
4544 routines: these may well fail if the certificates aren't installed.
4547 *) Code to support otherName option in GeneralName.
4550 *) First update to verify code. Change the verify utility
4551 so it warns if it is passed a self signed certificate:
4552 for consistency with the normal behaviour. X509_verify
4553 has been modified to it will now verify a self signed
4554 certificate if *exactly* the same certificate appears
4555 in the store: it was previously impossible to trust a
4556 single self signed certificate. This means that:
4557 openssl verify ss.pem
4558 now gives a warning about a self signed certificate but
4559 openssl verify -CAfile ss.pem ss.pem
4563 *) For servers, store verify_result in SSL_SESSION data structure
4564 (and add it to external session representation).
4565 This is needed when client certificate verifications fails,
4566 but an application-provided verification callback (set by
4567 SSL_CTX_set_cert_verify_callback) allows accepting the session
4568 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4569 but returns 1): When the session is reused, we have to set
4570 ssl->verify_result to the appropriate error code to avoid
4572 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4574 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4575 case in PKCS7_dataInit() where the signed PKCS7 structure
4576 didn't contain any existing data because it was being created.
4577 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4579 *) Add a salt to the key derivation routines in enc.c. This
4580 forms the first 8 bytes of the encrypted file. Also add a
4581 -S option to allow a salt to be input on the command line.
4584 *) New function X509_cmp(). Oddly enough there wasn't a function
4585 to compare two certificates. We do this by working out the SHA1
4586 hash and comparing that. X509_cmp() will be needed by the trust
4590 *) SSL_get1_session() is like SSL_get_session(), but increments
4591 the reference count in the SSL_SESSION returned.
4592 [Geoff Thorpe <geoff@eu.c2.net>]
4594 *) Fix for 'req': it was adding a null to request attributes.
4595 Also change the X509_LOOKUP and X509_INFO code to handle
4596 certificate auxiliary information.
4599 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4603 *) Add the possibility to add extra information to the memory leak
4604 detecting output, to form tracebacks, showing from where each
4605 allocation was originated: CRYPTO_push_info("constant string") adds
4606 the string plus current file name and line number to a per-thread
4607 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4608 is like calling CYRPTO_pop_info() until the stack is empty.
4609 Also updated memory leak detection code to be multi-thread-safe.
4612 *) Add options -text and -noout to pkcs7 utility and delete the
4613 encryption options which never did anything. Update docs.
4616 *) Add options to some of the utilities to allow the pass phrase
4617 to be included on either the command line (not recommended on
4618 OSes like Unix) or read from the environment. Update the
4619 manpages and fix a few bugs.
4622 *) Add a few manpages for some of the openssl commands.
4625 *) Fix the -revoke option in ca. It was freeing up memory twice,
4626 leaking and not finding already revoked certificates.
4629 *) Extensive changes to support certificate auxiliary information.
4630 This involves the use of X509_CERT_AUX structure and X509_AUX
4631 functions. An X509_AUX function such as PEM_read_X509_AUX()
4632 can still read in a certificate file in the usual way but it
4633 will also read in any additional "auxiliary information". By
4634 doing things this way a fair degree of compatibility can be
4635 retained: existing certificates can have this information added
4636 using the new 'x509' options.
4638 Current auxiliary information includes an "alias" and some trust
4639 settings. The trust settings will ultimately be used in enhanced
4640 certificate chain verification routines: currently a certificate
4641 can only be trusted if it is self signed and then it is trusted
4645 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4646 The problem was that one of the replacement routines had not been working
4647 since SSLeay releases. For now the offending routine has been replaced
4648 with non-optimised assembler. Even so, this now gives around 95%
4649 performance improvement for 1024 bit RSA signs.
4652 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4653 handling. Most clients have the effective key size in bits equal to
4654 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4655 A few however don't do this and instead use the size of the decrypted key
4656 to determine the RC2 key length and the AlgorithmIdentifier to determine
4657 the effective key length. In this case the effective key length can still
4658 be 40 bits but the key length can be 168 bits for example. This is fixed
4659 by manually forcing an RC2 key into the EVP_PKEY structure because the
4660 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4661 the key length and effective key length are equal.
4664 *) Add a bunch of functions that should simplify the creation of
4665 X509_NAME structures. Now you should be able to do:
4666 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4667 and have it automatically work out the correct field type and fill in
4668 the structures. The more adventurous can try:
4669 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4670 and it will (hopefully) work out the correct multibyte encoding.
4673 *) Change the 'req' utility to use the new field handling and multibyte
4674 copy routines. Before the DN field creation was handled in an ad hoc
4675 way in req, ca, and x509 which was rather broken and didn't support
4676 BMPStrings or UTF8Strings. Since some software doesn't implement
4677 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4678 using the dirstring_type option. See the new comment in the default
4679 openssl.cnf for more info.
4682 *) Make crypto/rand/md_rand.c more robust:
4683 - Assure unique random numbers after fork().
4684 - Make sure that concurrent threads access the global counter and
4685 md serializably so that we never lose entropy in them
4686 or use exactly the same state in multiple threads.
4687 Access to the large state is not always serializable because
4688 the additional locking could be a performance killer, and
4689 md should be large enough anyway.
4692 *) New file apps/app_rand.c with commonly needed functionality
4693 for handling the random seed file.
4695 Use the random seed file in some applications that previously did not:
4697 dsaparam -genkey (which also ignored its '-rand' option),
4700 x509 (when signing).
4701 Except on systems with /dev/urandom, it is crucial to have a random
4702 seed file at least for key creation, DSA signing, and for DH exchanges;
4703 for RSA signatures we could do without one.
4705 gendh and gendsa (unlike genrsa) used to read only the first byte
4706 of each file listed in the '-rand' option. The function as previously
4707 found in genrsa is now in app_rand.c and is used by all programs
4708 that support '-rand'.
4711 *) In RAND_write_file, use mode 0600 for creating files;
4712 don't just chmod when it may be too late.
4715 *) Report an error from X509_STORE_load_locations
4716 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4719 *) New function ASN1_mbstring_copy() this copies a string in either
4720 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4721 into an ASN1_STRING type. A mask of permissible types is passed
4722 and it chooses the "minimal" type to use or an error if not type
4726 *) Add function equivalents to the various macros in asn1.h. The old
4727 macros are retained with an M_ prefix. Code inside the library can
4728 use the M_ macros. External code (including the openssl utility)
4729 should *NOT* in order to be "shared library friendly".
4732 *) Add various functions that can check a certificate's extensions
4733 to see if it usable for various purposes such as SSL client,
4734 server or S/MIME and CAs of these types. This is currently
4735 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4736 verification. Also added a -purpose flag to x509 utility to
4737 print out all the purposes.
4740 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4744 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4745 for, obtain and decode and extension and obtain its critical flag.
4746 This allows all the necessary extension code to be handled in a
4747 single function call.
4750 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4751 platforms. See crypto/rc4/rc4_enc.c for further details.
4754 *) New -noout option to asn1parse. This causes no output to be produced
4755 its main use is when combined with -strparse and -out to extract data
4756 from a file (which may not be in ASN.1 format).
4759 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4760 when producing the local key id.
4761 [Richard Levitte <levitte@stacken.kth.se>]
4763 *) New option -dhparam in s_server. This allows a DH parameter file to be
4764 stated explicitly. If it is not stated then it tries the first server
4765 certificate file. The previous behaviour hard coded the filename
4769 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4770 a public key to be input or output. For example:
4771 openssl rsa -in key.pem -pubout -out pubkey.pem
4772 Also added necessary DSA public key functions to handle this.
4775 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4776 in the message. This was handled by allowing
4777 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4778 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4780 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4781 to the end of the strings whereas this didn't. This would cause problems
4782 if strings read with d2i_ASN1_bytes() were later modified.
4783 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4785 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4786 data and it contains EOF it will end up returning an error. This is
4787 caused by input 46 bytes long. The cause is due to the way base64
4788 BIOs find the start of base64 encoded data. They do this by trying a
4789 trial decode on each line until they find one that works. When they
4790 do a flag is set and it starts again knowing it can pass all the
4791 data directly through the decoder. Unfortunately it doesn't reset
4792 the context it uses. This means that if EOF is reached an attempt
4793 is made to pass two EOFs through the context and this causes the
4794 resulting error. This can also cause other problems as well. As is
4795 usual with these problems it takes *ages* to find and the fix is
4796 trivial: move one line.
4797 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4799 *) Ugly workaround to get s_client and s_server working under Windows. The
4800 old code wouldn't work because it needed to select() on sockets and the
4801 tty (for keypresses and to see if data could be written). Win32 only
4802 supports select() on sockets so we select() with a 1s timeout on the
4803 sockets and then see if any characters are waiting to be read, if none
4804 are present then we retry, we also assume we can always write data to
4805 the tty. This isn't nice because the code then blocks until we've
4806 received a complete line of data and it is effectively polling the
4807 keyboard at 1s intervals: however it's quite a bit better than not
4808 working at all :-) A dedicated Windows application might handle this
4809 with an event loop for example.
4812 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4813 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4814 will be called when RSA_sign() and RSA_verify() are used. This is useful
4815 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4816 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4817 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4818 This necessitated the support of an extra signature type NID_md5_sha1
4819 for SSL signatures and modifications to the SSL library to use it instead
4820 of calling RSA_public_decrypt() and RSA_private_encrypt().
4823 *) Add new -verify -CAfile and -CApath options to the crl program, these
4824 will lookup a CRL issuers certificate and verify the signature in a
4825 similar way to the verify program. Tidy up the crl program so it
4826 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4827 less strict. It will now permit CRL extensions even if it is not
4828 a V2 CRL: this will allow it to tolerate some broken CRLs.
4831 *) Initialize all non-automatic variables each time one of the openssl
4832 sub-programs is started (this is necessary as they may be started
4833 multiple times from the "OpenSSL>" prompt).
4834 [Lennart Bang, Bodo Moeller]
4836 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4837 removing all other RSA functionality (this is what NO_RSA does). This
4838 is so (for example) those in the US can disable those operations covered
4839 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4843 *) Non-copying interface to BIO pairs.
4844 (still largely untested)
4847 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4848 ASCII string. This was handled independently in various places before.
4851 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4852 UTF8 strings a character at a time.
4855 *) Use client_version from client hello to select the protocol
4856 (s23_srvr.c) and for RSA client key exchange verification
4857 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4860 *) Add various utility functions to handle SPKACs, these were previously
4861 handled by poking round in the structure internals. Added new function
4862 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4863 print, verify and generate SPKACs. Based on an original idea from
4864 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4867 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4870 *) Allow the config file extension section to be overwritten on the
4871 command line. Based on an original idea from Massimiliano Pala
4872 <madwolf@comune.modena.it>. The new option is called -extensions
4873 and can be applied to ca, req and x509. Also -reqexts to override
4874 the request extensions in req and -crlexts to override the crl extensions
4878 *) Add new feature to the SPKAC handling in ca. Now you can include
4879 the same field multiple times by preceding it by "XXXX." for example:
4882 this is the same syntax as used in the req config file.
4885 *) Allow certificate extensions to be added to certificate requests. These
4886 are specified in a 'req_extensions' option of the req section of the
4887 config file. They can be printed out with the -text option to req but
4888 are otherwise ignored at present.
4891 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4892 data read consists of only the final block it would not decrypted because
4893 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4894 A misplaced 'break' also meant the decrypted final block might not be
4895 copied until the next read.
4898 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4899 a few extra parameters to the DH structure: these will be useful if
4900 for example we want the value of 'q' or implement X9.42 DH.
4903 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4904 provides hooks that allow the default DSA functions or functions on a
4905 "per key" basis to be replaced. This allows hardware acceleration and
4906 hardware key storage to be handled without major modification to the
4907 library. Also added low level modexp hooks and CRYPTO_EX structure and
4908 associated functions.
4911 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4912 as "read only": it can't be written to and the buffer it points to will
4913 not be freed. Reading from a read only BIO is much more efficient than
4914 a normal memory BIO. This was added because there are several times when
4915 an area of memory needs to be read from a BIO. The previous method was
4916 to create a memory BIO and write the data to it, this results in two
4917 copies of the data and an O(n^2) reading algorithm. There is a new
4918 function BIO_new_mem_buf() which creates a read only memory BIO from
4919 an area of memory. Also modified the PKCS#7 routines to use read only
4923 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4924 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4925 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4926 but a retry condition occured while trying to read the rest.
4929 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4930 NID_pkcs7_encrypted by default: this was wrong since this should almost
4931 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4932 the encrypted data type: this is a more sensible place to put it and it
4933 allows the PKCS#12 code to be tidied up that duplicated this
4937 *) Changed obj_dat.pl script so it takes its input and output files on
4938 the command line. This should avoid shell escape redirection problems
4942 *) Initial support for certificate extension requests, these are included
4943 in things like Xenroll certificate requests. Included functions to allow
4944 extensions to be obtained and added.
4947 *) -crlf option to s_client and s_server for sending newlines as
4948 CRLF (as required by many protocols).
4951 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4953 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4954 [Ralf S. Engelschall]
4956 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4957 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4959 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4963 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4964 DH parameters/keys (q is lost during that conversion, but the resulting
4965 DH parameters contain its length).
4967 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4968 much faster than DH_generate_parameters (which creates parameters
4969 where p = 2*q + 1), and also the smaller q makes DH computations
4970 much more efficient (160-bit exponentiation instead of 1024-bit
4971 exponentiation); so this provides a convenient way to support DHE
4972 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4973 utter importance to use
4974 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4976 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4977 when such DH parameters are used, because otherwise small subgroup
4978 attacks may become possible!
4981 *) Avoid memory leak in i2d_DHparams.
4984 *) Allow the -k option to be used more than once in the enc program:
4985 this allows the same encrypted message to be read by multiple recipients.
4988 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4989 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4990 it will always use the numerical form of the OID, even if it has a short
4994 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4995 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4996 otherwise bn_mod_exp was called. In the case of hardware keys for example
4997 no private key components need be present and it might store extra data
4998 in the RSA structure, which cannot be accessed from bn_mod_exp.
4999 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5000 private key operations.
5003 *) Added support for SPARC Linux.
5006 *) pem_password_cb function type incompatibly changed from
5007 typedef int pem_password_cb(char *buf, int size, int rwflag);
5009 ....(char *buf, int size, int rwflag, void *userdata);
5010 so that applications can pass data to their callbacks:
5011 The PEM[_ASN1]_{read,write}... functions and macros now take an
5012 additional void * argument, which is just handed through whenever
5013 the password callback is called.
5014 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5016 New function SSL_CTX_set_default_passwd_cb_userdata.
5018 Compatibility note: As many C implementations push function arguments
5019 onto the stack in reverse order, the new library version is likely to
5020 interoperate with programs that have been compiled with the old
5021 pem_password_cb definition (PEM_whatever takes some data that
5022 happens to be on the stack as its last argument, and the callback
5023 just ignores this garbage); but there is no guarantee whatsoever that
5026 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5027 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5028 problems not only on Windows, but also on some Unix platforms.
5029 To avoid problematic command lines, these definitions are now in an
5030 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5031 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5034 *) MIPS III/IV assembler module is reimplemented.
5037 *) More DES library cleanups: remove references to srand/rand and
5038 delete an unused file.
5041 *) Add support for the the free Netwide assembler (NASM) under Win32,
5042 since not many people have MASM (ml) and it can be hard to obtain.
5043 This is currently experimental but it seems to work OK and pass all
5044 the tests. Check out INSTALL.W32 for info.
5047 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5048 without temporary keys kept an extra copy of the server key,
5049 and connections with temporary keys did not free everything in case
5053 *) New function RSA_check_key and new openssl rsa option -check
5054 for verifying the consistency of RSA keys.
5055 [Ulf Moeller, Bodo Moeller]
5057 *) Various changes to make Win32 compile work:
5058 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5059 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5060 comparison" warnings.
5061 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5064 *) Add a debugging option to PKCS#5 v2 key generation function: when
5065 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5066 derived keys are printed to stderr.
5069 *) Copy the flags in ASN1_STRING_dup().
5070 [Roman E. Pavlov <pre@mo.msk.ru>]
5072 *) The x509 application mishandled signing requests containing DSA
5073 keys when the signing key was also DSA and the parameters didn't match.
5075 It was supposed to omit the parameters when they matched the signing key:
5076 the verifying software was then supposed to automatically use the CA's
5077 parameters if they were absent from the end user certificate.
5079 Omitting parameters is no longer recommended. The test was also
5080 the wrong way round! This was probably due to unusual behaviour in
5081 EVP_cmp_parameters() which returns 1 if the parameters match.
5082 This meant that parameters were omitted when they *didn't* match and
5083 the certificate was useless. Certificates signed with 'ca' didn't have
5085 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5087 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5088 The interface is as follows:
5089 Applications can use
5090 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5091 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5092 "off" is now the default.
5093 The library internally uses
5094 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5095 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5096 to disable memory-checking temporarily.
5098 Some inconsistent states that previously were possible (and were
5099 even the default) are now avoided.
5101 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5102 with each memory chunk allocated; this is occasionally more helpful
5103 than just having a counter.
5105 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5107 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5111 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5112 which largely parallels "options", but is for changing API behaviour,
5113 whereas "options" are about protocol behaviour.
5114 Initial "mode" flags are:
5116 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5117 a single record has been written.
5118 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5119 retries use the same buffer location.
5120 (But all of the contents must be
5124 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5127 *) Fix problems with no-hmac etc.
5128 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5130 *) New functions RSA_get_default_method(), RSA_set_method() and
5131 RSA_get_method(). These allows replacement of RSA_METHODs without having
5132 to mess around with the internals of an RSA structure.
5135 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5136 Also really enable memory leak checks in openssl.c and in some
5138 [Chad C. Mulligan, Bodo Moeller]
5140 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5141 up the length of negative integers. This has now been simplified to just
5142 store the length when it is first determined and use it later, rather
5143 than trying to keep track of where data is copied and updating it to
5145 [Steve Henson, reported by Brien Wheeler
5146 <bwheeler@authentica-security.com>]
5148 *) Add a new function PKCS7_signatureVerify. This allows the verification
5149 of a PKCS#7 signature but with the signing certificate passed to the
5150 function itself. This contrasts with PKCS7_dataVerify which assumes the
5151 certificate is present in the PKCS#7 structure. This isn't always the
5152 case: certificates can be omitted from a PKCS#7 structure and be
5153 distributed by "out of band" means (such as a certificate database).
5156 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5157 function prototypes in pem.h, also change util/mkdef.pl to add the
5158 necessary function names.
5161 *) mk1mf.pl (used by Windows builds) did not properly read the
5162 options set by Configure in the top level Makefile, and Configure
5163 was not even able to write more than one option correctly.
5164 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5167 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5168 file to be loaded from a BIO or FILE pointer. The BIO version will
5169 for example allow memory BIOs to contain config info.
5172 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5173 Whoever hopes to achieve shared-library compatibility across versions
5174 must use this, not the compile-time macro.
5175 (Exercise 0.9.4: Which is the minimum library version required by
5177 Note: All this applies only to multi-threaded programs, others don't
5181 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5182 through a BIO pair triggered the default case, i.e.
5183 SSLerr(...,SSL_R_UNKNOWN_STATE).
5186 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5187 can use the SSL library even if none of the specific BIOs is
5191 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5192 for the encoded length.
5193 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5195 *) Add initial documentation of the X509V3 functions.
5198 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5199 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5200 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5201 secure PKCS#8 private key format with a high iteration count.
5204 *) Fix determination of Perl interpreter: A perl or perl5
5205 _directory_ in $PATH was also accepted as the interpreter.
5206 [Ralf S. Engelschall]
5208 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5209 wrong with it but it was very old and did things like calling
5210 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5214 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5215 to use the new extension code.
5218 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5219 with macros. This should make it easier to change their form, add extra
5220 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5224 *) Add to configuration table a new entry that can specify an alternative
5225 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5226 according to Mark Crispin <MRC@Panda.COM>.
5230 *) DES CBC did not update the IV. Weird.
5233 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5234 Changing the behaviour of the former might break existing programs --
5235 where IV updating is needed, des_ncbc_encrypt can be used.
5238 *) When bntest is run from "make test" it drives bc to check its
5239 calculations, as well as internally checking them. If an internal check
5240 fails, it needs to cause bc to give a non-zero result or make test carries
5241 on without noticing the failure. Fixed.
5244 *) DES library cleanups.
5247 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5248 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5249 ciphers. NOTE: although the key derivation function has been verified
5250 against some published test vectors it has not been extensively tested
5251 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5255 *) Instead of "mkdir -p", which is not fully portable, use new
5256 Perl script "util/mkdir-p.pl".
5259 *) Rewrite the way password based encryption (PBE) is handled. It used to
5260 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5261 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5262 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5263 the 'parameter' field of the AlgorithmIdentifier is passed to the
5264 underlying key generation function so it must do its own ASN1 parsing.
5265 This has also changed the EVP_PBE_CipherInit() function which now has a
5266 'parameter' argument instead of literal salt and iteration count values
5267 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5270 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5271 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5272 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5273 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5274 value was just used as a "magic string" and not used directly its
5275 value doesn't matter.
5278 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5282 *) "linux-sparc64" configuration (ultrapenguin).
5283 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5284 "linux-sparc" configuration.
5285 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5287 *) config now generates no-xxx options for missing ciphers.
5290 *) Support the EBCDIC character set (work in progress).
5291 File ebcdic.c not yet included because it has a different license.
5292 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5294 *) Support BS2000/OSD-POSIX.
5295 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5297 *) Make callbacks for key generation use void * instead of char *.
5300 *) Make S/MIME samples compile (not yet tested).
5303 *) Additional typesafe stacks.
5306 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5310 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5312 *) New configuration variant "sco5-gcc".
5314 *) Updated some demos.
5315 [Sean O Riordain, Wade Scholine]
5317 *) Add missing BIO_free at exit of pkcs12 application.
5320 *) Fix memory leak in conf.c.
5323 *) Updates for Win32 to assembler version of MD5.
5326 *) Set #! path to perl in apps/der_chop to where we found it
5327 instead of using a fixed path.
5330 *) SHA library changes for irix64-mips4-cc.
5333 *) Improvements for VMS support.
5337 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5339 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5340 This also avoids the problems with SC4.2 and unpatched SC5.
5341 [Andy Polyakov <appro@fy.chalmers.se>]
5343 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5344 These are required because of the typesafe stack would otherwise break
5345 existing code. If old code used a structure member which used to be STACK
5346 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5347 sk_num or sk_value it would produce an error because the num, data members
5348 are not present in STACK_OF. Now it just produces a warning. sk_set
5349 replaces the old method of assigning a value to sk_value
5350 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5351 that does this will no longer work (and should use sk_set instead) but
5352 this could be regarded as a "questionable" behaviour anyway.
5355 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5356 correctly handle encrypted S/MIME data.
5359 *) Change type of various DES function arguments from des_cblock
5360 (which means, in function argument declarations, pointer to char)
5361 to des_cblock * (meaning pointer to array with 8 char elements),
5362 which allows the compiler to do more typechecking; it was like
5363 that back in SSLeay, but with lots of ugly casts.
5365 Introduce new type const_des_cblock.
5368 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5369 problems: find RecipientInfo structure that matches recipient certificate
5370 and initialise the ASN1 structures properly based on passed cipher.
5373 *) Belatedly make the BN tests actually check the results.
5376 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5377 to and from BNs: it was completely broken. New compilation option
5378 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5379 key elements as negative integers.
5382 *) Reorganize and speed up MD5.
5383 [Andy Polyakov <appro@fy.chalmers.se>]
5386 [Richard Levitte <richard@levitte.org>]
5388 *) New option -out to asn1parse to allow the parsed structure to be
5389 output to a file. This is most useful when combined with the -strparse
5390 option to examine the output of things like OCTET STRINGS.
5393 *) Make SSL library a little more fool-proof by not requiring any longer
5394 that SSL_set_{accept,connect}_state be called before
5395 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5396 in many applications because usually everything *appeared* to work as
5397 intended anyway -- now it really works as intended).
5400 *) Move openssl.cnf out of lib/.
5403 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5404 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5405 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5406 [Ralf S. Engelschall]
5408 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5409 handle PKCS#7 enveloped data properly.
5410 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5412 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5413 copying pointers. The cert_st handling is changed by this in
5414 various ways (and thus what used to be known as ctx->default_cert
5415 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5416 any longer when s->cert does not give us what we need).
5417 ssl_cert_instantiate becomes obsolete by this change.
5418 As soon as we've got the new code right (possibly it already is?),
5419 we have solved a couple of bugs of the earlier code where s->cert
5420 was used as if it could not have been shared with other SSL structures.
5422 Note that using the SSL API in certain dirty ways now will result
5423 in different behaviour than observed with earlier library versions:
5424 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5425 does not influence s as it used to.
5427 In order to clean up things more thoroughly, inside SSL_SESSION
5428 we don't use CERT any longer, but a new structure SESS_CERT
5429 that holds per-session data (if available); currently, this is
5430 the peer's certificate chain and, for clients, the server's certificate
5431 and temporary key. CERT holds only those values that can have
5432 meaningful defaults in an SSL_CTX.
5435 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5436 from the internal representation. Various PKCS#7 fixes: remove some
5437 evil casts and set the enc_dig_alg field properly based on the signing
5441 *) Allow PKCS#12 password to be set from the command line or the
5442 environment. Let 'ca' get its config file name from the environment
5443 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5447 *) Allow certificate policies extension to use an IA5STRING for the
5448 organization field. This is contrary to the PKIX definition but
5449 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5453 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5454 without disallowing inline assembler and the like for non-pedantic builds.
5457 *) Support Borland C++ builder.
5458 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5463 *) SHA-1 cleanups and performance enhancements.
5464 [Andy Polyakov <appro@fy.chalmers.se>]
5466 *) Sparc v8plus assembler for the bignum library.
5467 [Andy Polyakov <appro@fy.chalmers.se>]
5469 *) Accept any -xxx and +xxx compiler options in Configure.
5472 *) Update HPUX configuration.
5475 *) Add missing sk_<type>_unshift() function to safestack.h
5476 [Ralf S. Engelschall]
5478 *) New function SSL_CTX_use_certificate_chain_file that sets the
5479 "extra_cert"s in addition to the certificate. (This makes sense
5480 only for "PEM" format files, as chains as a whole are not
5484 *) Support verify_depth from the SSL API.
5485 x509_vfy.c had what can be considered an off-by-one-error:
5486 Its depth (which was not part of the external interface)
5487 was actually counting the number of certificates in a chain;
5488 now it really counts the depth.
5491 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5492 instead of X509err, which often resulted in confusing error
5493 messages since the error codes are not globally unique
5494 (e.g. an alleged error in ssl3_accept when a certificate
5495 didn't match the private key).
5497 *) New function SSL_CTX_set_session_id_context that allows to set a default
5498 value (so that you don't need SSL_set_session_id_context for each
5499 connection using the SSL_CTX).
5502 *) OAEP decoding bug fix.
5505 *) Support INSTALL_PREFIX for package builders, as proposed by
5509 *) New Configure options "threads" and "no-threads". For systems
5510 where the proper compiler options are known (currently Solaris
5511 and Linux), "threads" is the default.
5514 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5517 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5518 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5519 such as /usr/local/bin.
5522 *) "make linux-shared" to build shared libraries.
5523 [Niels Poppe <niels@netbox.org>]
5525 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5528 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5529 extension adding in x509 utility.
5532 *) Remove NOPROTO sections and error code comments.
5535 *) Partial rewrite of the DEF file generator to now parse the ANSI
5539 *) New Configure options --prefix=DIR and --openssldir=DIR.
5542 *) Complete rewrite of the error code script(s). It is all now handled
5543 by one script at the top level which handles error code gathering,
5544 header rewriting and C source file generation. It should be much better
5545 than the old method: it now uses a modified version of Ulf's parser to
5546 read the ANSI prototypes in all header files (thus the old K&R definitions
5547 aren't needed for error creation any more) and do a better job of
5548 translating function codes into names. The old 'ASN1 error code imbedded
5549 in a comment' is no longer necessary and it doesn't use .err files which
5550 have now been deleted. Also the error code call doesn't have to appear all
5551 on one line (which resulted in some large lines...).
5554 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5557 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5558 0 (which usually indicates a closed connection), but continue reading.
5561 *) Fix some race conditions.
5564 *) Add support for CRL distribution points extension. Add Certificate
5565 Policies and CRL distribution points documentation.
5568 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5571 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5572 8 of keying material. Merlin has also confirmed interop with this fix
5573 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5574 [Merlin Hughes <merlin@baltimore.ie>]
5576 *) Fix lots of warnings.
5577 [Richard Levitte <levitte@stacken.kth.se>]
5579 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5580 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5581 [Richard Levitte <levitte@stacken.kth.se>]
5583 *) Fix problems with sizeof(long) == 8.
5584 [Andy Polyakov <appro@fy.chalmers.se>]
5586 *) Change functions to ANSI C.
5589 *) Fix typos in error codes.
5590 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5592 *) Remove defunct assembler files from Configure.
5595 *) SPARC v8 assembler BIGNUM implementation.
5596 [Andy Polyakov <appro@fy.chalmers.se>]
5598 *) Support for Certificate Policies extension: both print and set.
5599 Various additions to support the r2i method this uses.
5602 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5603 return a const string when you are expecting an allocated buffer.
5606 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5607 types DirectoryString and DisplayText.
5610 *) Add code to allow r2i extensions to access the configuration database,
5611 add an LHASH database driver and add several ctx helper functions.
5614 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5615 fail when they extended the size of a BIGNUM.
5618 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5619 support typesafe stack.
5622 *) Fix typo in SSL_[gs]et_options().
5623 [Nils Frostberg <nils@medcom.se>]
5625 *) Delete various functions and files that belonged to the (now obsolete)
5626 old X509V3 handling code.
5629 *) New Configure option "rsaref".
5632 *) Don't auto-generate pem.h.
5635 *) Introduce type-safe ASN.1 SETs.
5638 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5639 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5641 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5642 that links with OpenSSL (well at least cause lots of warnings), but fear
5643 not: the conversion is trivial, and it eliminates loads of evil casts. A
5644 few STACKed things have been converted already. Feel free to convert more.
5645 In the fullness of time, I'll do away with the STACK type altogether.
5648 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5649 specified in <certfile> by updating the entry in the index.txt file.
5650 This way one no longer has to edit the index.txt file manually for
5651 revoking a certificate. The -revoke option does the gory details now.
5652 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5654 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5655 `-text' option at all and this way the `-noout -text' combination was
5656 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5657 [Ralf S. Engelschall]
5659 *) Make sure a corresponding plain text error message exists for the
5660 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5661 verify callback function determined that a certificate was revoked.
5662 [Ralf S. Engelschall]
5664 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5665 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5666 all available cipers including rc5, which was forgotten until now.
5667 In order to let the testing shell script know which algorithms
5668 are available, a new (up to now undocumented) command
5669 "openssl list-cipher-commands" is used.
5672 *) Bugfix: s_client occasionally would sleep in select() when
5673 it should have checked SSL_pending() first.
5676 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5677 the raw DSA values prior to ASN.1 encoding.
5680 *) Tweaks to Configure
5681 [Niels Poppe <niels@netbox.org>]
5683 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5687 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5690 *) New config option to avoid instructions that are illegal on the 80386.
5691 The default code is faster, but requires at least a 486.
5694 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5695 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5696 same as SSL2_VERSION anyway.
5699 *) New "-showcerts" option for s_client.
5702 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5703 application. Various cleanups and fixes.
5706 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5707 modify error routines to work internally. Add error codes and PBE init
5708 to library startup routines.
5711 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5712 packing functions to asn1 and evp. Changed function names and error
5713 codes along the way.
5716 *) PKCS12 integration: and so it begins... First of several patches to
5717 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5718 objects to objects.h
5721 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5722 and display support for Thawte strong extranet extension.
5725 *) Add LinuxPPC support.
5726 [Jeff Dubrule <igor@pobox.org>]
5728 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5729 bn_div_words in alpha.s.
5730 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5732 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5733 OAEP isn't supported when OpenSSL is built with RSAref.
5734 [Ulf Moeller <ulf@fitug.de>]
5736 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5737 so they no longer are missing under -DNOPROTO.
5738 [Soren S. Jorvang <soren@t.dk>]
5741 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5743 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5744 doesn't work when the session is reused. Coming soon!
5747 *) Fix a security hole, that allows sessions to be reused in the wrong
5748 context thus bypassing client cert protection! All software that uses
5749 client certs and session caches in multiple contexts NEEDS PATCHING to
5750 allow session reuse! A fuller solution is in the works.
5751 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5753 *) Some more source tree cleanups (removed obsolete files
5754 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5755 permission on "config" script to be executable) and a fix for the INSTALL
5757 [Ulf Moeller <ulf@fitug.de>]
5759 *) Remove some legacy and erroneous uses of malloc, free instead of
5761 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5763 *) Make rsa_oaep_test return non-zero on error.
5764 [Ulf Moeller <ulf@fitug.de>]
5766 *) Add support for native Solaris shared libraries. Configure
5767 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5768 if someone would make that last step automatic.
5769 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5771 *) ctx_size was not built with the right compiler during "make links". Fixed.
5774 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5775 except NULL ciphers". This means the default cipher list will no longer
5776 enable NULL ciphers. They need to be specifically enabled e.g. with
5777 the string "DEFAULT:eNULL".
5780 *) Fix to RSA private encryption routines: if p < q then it would
5781 occasionally produce an invalid result. This will only happen with
5782 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5785 *) Be less restrictive and allow also `perl util/perlpath.pl
5786 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5787 because this way one can also use an interpreter named `perl5' (which is
5788 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5789 installed as `perl').
5790 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5792 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5793 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5795 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5796 advapi32.lib to Win32 build and change the pem test comparision
5797 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5798 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5799 and crypto/des/ede_cbcm_enc.c.
5802 *) DES quad checksum was broken on big-endian architectures. Fixed.
5805 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5806 Win32 test batch file so it (might) work again. The Win32 test batch file
5807 is horrible: I feel ill....
5810 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5811 in e_os.h. Audit of header files to check ANSI and non ANSI
5812 sections: 10 functions were absent from non ANSI section and not exported
5813 from Windows DLLs. Fixed up libeay.num for new functions.
5816 *) Make `openssl version' output lines consistent.
5817 [Ralf S. Engelschall]
5819 *) Fix Win32 symbol export lists for BIO functions: Added
5820 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5821 to ms/libeay{16,32}.def.
5822 [Ralf S. Engelschall]
5824 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5825 fine under Unix and passes some trivial tests I've now added. But the
5826 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5827 added to make sure no one expects that this stuff really works in the
5828 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5829 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5831 [Ralf S. Engelschall]
5833 *) Fix the generation of two part addresses in perl.
5834 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5836 *) Add config entry for Linux on MIPS.
5837 [John Tobey <jtobey@channel1.com>]
5839 *) Make links whenever Configure is run, unless we are on Windoze.
5842 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5843 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5847 *) Add a useful kludge to allow package maintainers to specify compiler and
5848 other platforms details on the command line without having to patch the
5849 Configure script everytime: One now can use ``perl Configure
5850 <id>:<details>'', i.e. platform ids are allowed to have details appended
5851 to them (seperated by colons). This is treated as there would be a static
5852 pre-configured entry in Configure's %table under key <id> with value
5853 <details> and ``perl Configure <id>'' is called. So, when you want to
5854 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5855 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5856 now, which overrides the FreeBSD-elf entry on-the-fly.
5857 [Ralf S. Engelschall]
5859 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5862 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5863 on the `perl Configure ...' command line. This way one can compile
5864 OpenSSL libraries with Position Independent Code (PIC) which is needed
5865 for linking it into DSOs.
5866 [Ralf S. Engelschall]
5868 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5872 *) Cleaned up the LICENSE document: The official contact for any license
5873 questions now is the OpenSSL core team under openssl-core@openssl.org.
5874 And add a paragraph about the dual-license situation to make sure people
5875 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5876 to the OpenSSL toolkit.
5877 [Ralf S. Engelschall]
5879 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5880 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5881 Additonally cleaned up the `make links' target: Remove unnecessary
5882 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5883 to speed processing and no longer clutter the display with confusing
5884 stuff. Instead only the actually done links are displayed.
5885 [Ralf S. Engelschall]
5887 *) Permit null encryption ciphersuites, used for authentication only. It used
5888 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5889 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5893 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5894 signed attributes when verifying signatures (this would break them),
5895 the detached data encoding was wrong and public keys obtained using
5896 X509_get_pubkey() weren't freed.
5899 *) Add text documentation for the BUFFER functions. Also added a work around
5900 to a Win95 console bug. This was triggered by the password read stuff: the
5901 last character typed gets carried over to the next fread(). If you were
5902 generating a new cert request using 'req' for example then the last
5903 character of the passphrase would be CR which would then enter the first
5907 *) Added the new `Includes OpenSSL Cryptography Software' button as
5908 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5909 button and can be used by applications based on OpenSSL to show the
5910 relationship to the OpenSSL project.
5911 [Ralf S. Engelschall]
5913 *) Remove confusing variables in function signatures in files
5914 ssl/ssl_lib.c and ssl/ssl.h.
5915 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5917 *) Don't install bss_file.c under PREFIX/include/
5918 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5920 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5921 functions that return function pointers and has support for NT specific
5922 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5923 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5924 unsigned to signed types: this was killing the Win32 compile.
5927 *) Add new certificate file to stack functions,
5928 SSL_add_dir_cert_subjects_to_stack() and
5929 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5930 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5931 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5932 This means that Apache-SSL and similar packages don't have to mess around
5933 to add as many CAs as they want to the preferred list.
5936 *) Experiment with doxygen documentation. Currently only partially applied to
5938 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5939 openssl.doxy as the configuration file.
5942 *) Get rid of remaining C++-style comments which strict C compilers hate.
5943 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5945 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5946 compiled in by default: it has problems with large keys.
5949 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5950 DH private keys and/or callback functions which directly correspond to
5951 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5952 is needed for applications which have to configure certificates on a
5953 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5955 For the RSA certificate situation is makes no difference, but
5956 for the DSA certificate situation this fixes the "no shared cipher"
5957 problem where the OpenSSL cipher selection procedure failed because the
5958 temporary keys were not overtaken from the context and the API provided
5959 no way to reconfigure them.
5960 The new functions now let applications reconfigure the stuff and they
5961 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5962 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5963 non-public-API function ssl_cert_instantiate() is used as a helper
5964 function and also to reduce code redundancy inside ssl_rsa.c.
5965 [Ralf S. Engelschall]
5967 *) Move s_server -dcert and -dkey options out of the undocumented feature
5968 area because they are useful for the DSA situation and should be
5969 recognized by the users.
5970 [Ralf S. Engelschall]
5972 *) Fix the cipher decision scheme for export ciphers: the export bits are
5973 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5974 SSL_EXP_MASK. So, the original variable has to be used instead of the
5975 already masked variable.
5976 [Richard Levitte <levitte@stacken.kth.se>]
5978 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5979 [Richard Levitte <levitte@stacken.kth.se>]
5981 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5982 from `int' to `unsigned int' because it's a length and initialized by
5983 EVP_DigestFinal() which expects an `unsigned int *'.
5984 [Richard Levitte <levitte@stacken.kth.se>]
5986 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5987 script. Instead use the usual Shell->Perl transition trick.
5988 [Ralf S. Engelschall]
5990 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5991 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5992 -noout -modulus' as it's already the case for `openssl rsa -noout
5993 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5994 currently the public key is printed (a decision which was already done by
5995 `openssl dsa -modulus' in the past) which serves a similar purpose.
5996 Additionally the NO_RSA no longer completely removes the whole -modulus
5997 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5999 [Ralf S. Engelschall]
6001 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6002 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6003 [Arne Ansper <arne@ats.cyber.ee>]
6005 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6006 to be added. Now both 'req' and 'ca' can use new objects defined in the
6010 *) Add cool BIO that does syslog (or event log on NT).
6011 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6013 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6014 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6015 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6016 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6019 *) Add preliminary config info for new extension code.
6022 *) Make RSA_NO_PADDING really use no padding.
6023 [Ulf Moeller <ulf@fitug.de>]
6025 *) Generate errors when private/public key check is done.
6028 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6029 for some CRL extensions and new objects added.
6032 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6033 key usage extension and fuller support for authority key id.
6036 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6037 padding method for RSA, which is recommended for new applications in PKCS
6038 #1 v2.0 (RFC 2437, October 1998).
6039 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6040 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6041 against Bleichbacher's attack on RSA.
6042 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6045 *) Updates to the new SSL compression code
6046 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6048 *) Fix so that the version number in the master secret, when passed
6049 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6050 (because the server will not accept higher), that the version number
6051 is 0x03,0x01, not 0x03,0x00
6052 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6054 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6055 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6056 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6059 *) Support for RAW extensions where an arbitrary extension can be
6060 created by including its DER encoding. See apps/openssl.cnf for
6064 *) Make sure latest Perl versions don't interpret some generated C array
6065 code as Perl array code in the crypto/err/err_genc.pl script.
6066 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6068 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6069 not many people have the assembler. Various Win32 compilation fixes and
6070 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6074 *) Modify configure script 'Configure' to automatically create crypto/date.h
6075 file under Win32 and also build pem.h from pem.org. New script
6076 util/mkfiles.pl to create the MINFO file on environments that can't do a
6077 'make files': perl util/mkfiles.pl >MINFO should work.
6080 *) Major rework of DES function declarations, in the pursuit of correctness
6081 and purity. As a result, many evil casts evaporated, and some weirdness,
6082 too. You may find this causes warnings in your code. Zapping your evil
6083 casts will probably fix them. Mostly.
6086 *) Fix for a typo in asn1.h. Bug fix to object creation script
6087 obj_dat.pl. It considered a zero in an object definition to mean
6088 "end of object": none of the objects in objects.h have any zeros
6089 so it wasn't spotted.
6090 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6092 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6093 Masking (CBCM). In the absence of test vectors, the best I have been able
6094 to do is check that the decrypt undoes the encrypt, so far. Send me test
6095 vectors if you have them.
6098 *) Correct calculation of key length for export ciphers (too much space was
6099 allocated for null ciphers). This has not been tested!
6102 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6103 message is now correct (it understands "crypto" and "ssl" on its
6104 command line). There is also now an "update" option. This will update
6105 the util/ssleay.num and util/libeay.num files with any new functions.
6107 perl util/mkdef.pl crypto ssl update
6108 it will update them.
6111 *) Overhauled the Perl interface (perl/*):
6112 - ported BN stuff to OpenSSL's different BN library
6113 - made the perl/ source tree CVS-aware
6114 - renamed the package from SSLeay to OpenSSL (the files still contain
6115 their history because I've copied them in the repository)
6116 - removed obsolete files (the test scripts will be replaced
6117 by better Test::Harness variants in the future)
6118 [Ralf S. Engelschall]
6120 *) First cut for a very conservative source tree cleanup:
6121 1. merge various obsolete readme texts into doc/ssleay.txt
6122 where we collect the old documents and readme texts.
6123 2. remove the first part of files where I'm already sure that we no
6124 longer need them because of three reasons: either they are just temporary
6125 files which were left by Eric or they are preserved original files where
6126 I've verified that the diff is also available in the CVS via "cvs diff
6127 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6128 the crypto/md/ stuff).
6129 [Ralf S. Engelschall]
6131 *) More extension code. Incomplete support for subject and issuer alt
6132 name, issuer and authority key id. Change the i2v function parameters
6133 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6134 what that's for :-) Fix to ASN1 macro which messed up
6135 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6138 *) Preliminary support for ENUMERATED type. This is largely copied from the
6142 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6143 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6145 *) Make sure `make rehash' target really finds the `openssl' program.
6146 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6148 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6149 like to hear about it if this slows down other processors.
6152 *) Add CygWin32 platform information to Configure script.
6153 [Alan Batie <batie@aahz.jf.intel.com>]
6155 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6156 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6158 *) New program nseq to manipulate netscape certificate sequences
6161 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6165 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6166 but the BN code had some problems that would cause failures when
6167 doing certificate verification and some other functions.
6168 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6170 *) Add ASN1 and PEM code to support netscape certificate sequences.
6173 *) Add ASN1 and PEM code to support netscape certificate sequences.
6176 *) Add several PKIX and private extended key usage OIDs.
6179 *) Modify the 'ca' program to handle the new extension code. Modify
6180 openssl.cnf for new extension format, add comments.
6183 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6184 and add a sample to openssl.cnf so req -x509 now adds appropriate
6188 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6189 error code, add initial support to X509_print() and x509 application.
6192 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6193 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6194 stuff is currently isolated and isn't even compiled yet.
6197 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6198 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6199 Removed the versions check from X509 routines when loading extensions:
6200 this allows certain broken certificates that don't set the version
6201 properly to be processed.
6204 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6205 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6206 can still be regenerated with "make depend".
6209 *) Spelling mistake in C version of CAST-128.
6210 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6212 *) Changes to the error generation code. The perl script err-code.pl
6213 now reads in the old error codes and retains the old numbers, only
6214 adding new ones if necessary. It also only changes the .err files if new
6215 codes are added. The makefiles have been modified to only insert errors
6216 when needed (to avoid needlessly modifying header files). This is done
6217 by only inserting errors if the .err file is newer than the auto generated
6218 C file. To rebuild all the error codes from scratch (the old behaviour)
6219 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6220 or delete all the .err files.
6223 *) CAST-128 was incorrectly implemented for short keys. The C version has
6224 been fixed, but is untested. The assembler versions are also fixed, but
6225 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6226 to regenerate it if needed.
6227 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6228 Hagino <itojun@kame.net>]
6230 *) File was opened incorrectly in randfile.c.
6231 [Ulf Möller <ulf@fitug.de>]
6233 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6234 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6235 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6236 al: it's just almost always a UTCTime. Note this patch adds new error
6237 codes so do a "make errors" if there are problems.
6240 *) Correct Linux 1 recognition in config.
6241 [Ulf Möller <ulf@fitug.de>]
6243 *) Remove pointless MD5 hash when using DSA keys in ca.
6244 [Anonymous <nobody@replay.com>]
6246 *) Generate an error if given an empty string as a cert directory. Also
6247 generate an error if handed NULL (previously returned 0 to indicate an
6248 error, but didn't set one).
6249 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6251 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6254 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6255 parameters. This was causing a warning which killed off the Win32 compile.
6258 *) Remove C++ style comments from crypto/bn/bn_local.h.
6259 [Neil Costigan <neil.costigan@celocom.com>]
6261 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6262 based on a text string, looking up short and long names and finally
6263 "dot" format. The "dot" format stuff didn't work. Added new function
6264 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6265 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6266 OID is not part of the table.
6269 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6270 X509_LOOKUP_by_alias().
6273 *) Sort openssl functions by name.
6276 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6277 encryption from sample DSA keys (in case anyone is interested the password
6281 *) Make _all_ *_free functions accept a NULL pointer.
6282 [Frans Heymans <fheymans@isaserver.be>]
6284 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6286 [Anonymous <nobody@replay.com>]
6288 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6289 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6291 *) Don't blow it for numeric -newkey arguments to apps/req.
6292 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6294 *) Temp key "for export" tests were wrong in s3_srvr.c.
6295 [Anonymous <nobody@replay.com>]
6297 *) Add prototype for temp key callback functions
6298 SSL_CTX_set_tmp_{rsa,dh}_callback().
6301 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6302 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6305 *) X509_name_add_entry() freed the wrong thing after an error.
6306 [Arne Ansper <arne@ats.cyber.ee>]
6308 *) rsa_eay.c would attempt to free a NULL context.
6309 [Arne Ansper <arne@ats.cyber.ee>]
6311 *) BIO_s_socket() had a broken should_retry() on Windoze.
6312 [Arne Ansper <arne@ats.cyber.ee>]
6314 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6315 [Arne Ansper <arne@ats.cyber.ee>]
6317 *) Make sure the already existing X509_STORE->depth variable is initialized
6318 in X509_STORE_new(), but document the fact that this variable is still
6319 unused in the certificate verification process.
6320 [Ralf S. Engelschall]
6322 *) Fix the various library and apps files to free up pkeys obtained from
6323 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6326 *) Fix reference counting in X509_PUBKEY_get(). This makes
6327 demos/maurice/example2.c work, amongst others, probably.
6328 [Steve Henson and Ben Laurie]
6330 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6331 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6332 are no longer created. This way we have a single and consistent command
6333 line interface `openssl <command>', similar to `cvs <command>'.
6334 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6336 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6337 BIT STRING wrapper always have zero unused bits.
6340 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6343 *) Make the top-level INSTALL documentation easier to understand.
6346 *) Makefiles updated to exit if an error occurs in a sub-directory
6347 make (including if user presses ^C) [Paul Sutton]
6349 *) Make Montgomery context stuff explicit in RSA data structure.
6352 *) Fix build order of pem and err to allow for generated pem.h.
6355 *) Fix renumbering bug in X509_NAME_delete_entry().
6358 *) Enhanced the err-ins.pl script so it makes the error library number
6359 global and can add a library name. This is needed for external ASN1 and
6360 other error libraries.
6363 *) Fixed sk_insert which never worked properly.
6366 *) Fix ASN1 macros so they can handle indefinite length construted
6367 EXPLICIT tags. Some non standard certificates use these: they can now
6371 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6372 into a single doc/ssleay.txt bundle. This way the information is still
6373 preserved but no longer messes up this directory. Now it's new room for
6374 the new set of documenation files.
6375 [Ralf S. Engelschall]
6377 *) SETs were incorrectly DER encoded. This was a major pain, because they
6378 shared code with SEQUENCEs, which aren't coded the same. This means that
6379 almost everything to do with SETs or SEQUENCEs has either changed name or
6380 number of arguments.
6381 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6383 *) Fix test data to work with the above.
6386 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6387 was already fixed by Eric for 0.9.1 it seems.
6388 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6390 *) Autodetect FreeBSD3.
6393 *) Fix various bugs in Configure. This affects the following platforms:
6397 unixware-2.0-pentium
6401 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6402 before they are needed.
6405 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6409 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6411 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6412 changed SSLeay to OpenSSL in version strings.
6413 [Ralf S. Engelschall]
6415 *) Some fixups to the top-level documents.
6418 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6419 because the symlink to include/ was missing.
6420 [Ralf S. Engelschall]
6422 *) Incorporated the popular no-RSA/DSA-only patches
6423 which allow to compile a RSA-free SSLeay.
6424 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6426 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6427 when "ssleay" is still not found.
6428 [Ralf S. Engelschall]
6430 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6431 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6433 *) Updated the README file.
6434 [Ralf S. Engelschall]
6436 *) Added various .cvsignore files in the CVS repository subdirs
6437 to make a "cvs update" really silent.
6438 [Ralf S. Engelschall]
6440 *) Recompiled the error-definition header files and added
6441 missing symbols to the Win32 linker tables.
6442 [Ralf S. Engelschall]
6444 *) Cleaned up the top-level documents;
6445 o new files: CHANGES and LICENSE
6446 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6447 o merged COPYRIGHT into LICENSE
6448 o removed obsolete TODO file
6449 o renamed MICROSOFT to INSTALL.W32
6450 [Ralf S. Engelschall]
6452 *) Removed dummy files from the 0.9.1b source tree:
6453 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6454 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6455 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6456 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6457 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6458 [Ralf S. Engelschall]
6460 *) Added various platform portability fixes.
6463 *) The Genesis of the OpenSSL rpject:
6464 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6465 Young and Tim J. Hudson created while they were working for C2Net until
6467 [The OpenSSL Project]
6470 Changes between 0.9.0b and 0.9.1b [not released]
6472 *) Updated a few CA certificates under certs/
6475 *) Changed some BIGNUM api stuff.
6478 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6479 DGUX x86, Linux Alpha, etc.
6482 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6483 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6487 *) Add -strparse option to asn1pars program which parses nested
6489 [Dr Stephen Henson <shenson@bigfoot.com>]
6491 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6494 *) DSA fix for "ca" program.
6497 *) Added "-genkey" option to "dsaparam" program.
6500 *) Added RIPE MD160 (rmd160) message digest.
6503 *) Added -a (all) option to "ssleay version" command.
6506 *) Added PLATFORM define which is the id given to Configure.
6509 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6512 *) Extended the ASN.1 parser routines.
6515 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6518 *) Added a BN_CTX to the BN library.
6521 *) Fixed the weak key values in DES library
6524 *) Changed API in EVP library for cipher aliases.
6527 *) Added support for RC2/64bit cipher.
6530 *) Converted the lhash library to the crypto/mem.c functions.
6533 *) Added more recognized ASN.1 object ids.
6536 *) Added more RSA padding checks for SSL/TLS.
6539 *) Added BIO proxy/filter functionality.
6542 *) Added extra_certs to SSL_CTX which can be used
6543 send extra CA certificates to the client in the CA cert chain sending
6544 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6547 *) Now Fortezza is denied in the authentication phase because
6548 this is key exchange mechanism is not supported by SSLeay at all.
6551 *) Additional PKCS1 checks.
6554 *) Support the string "TLSv1" for all TLS v1 ciphers.
6557 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6558 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6561 *) Fixed a few memory leaks.
6564 *) Fixed various code and comment typos.
6567 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6568 bytes sent in the client random.
6569 [Edward Bishop <ebishop@spyglass.com>]