5 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
8 (CVE-2006-4339) [Ben Laurie and Google Security Team]
10 *) Change the Unix randomness entropy gathering to use poll() when
11 possible instead of select(), since the latter has some
12 undesirable limitations.
13 [Darryl Miles via Richard Levitte and Bodo Moeller]
15 *) Disable rogue ciphersuites:
17 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
18 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
19 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
21 The latter two were purportedly from
22 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
25 Also deactive the remaining ciphersuites from
26 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
27 unofficial, and the ID has long expired.
30 *) Fix RSA blinding Heisenbug (problems sometimes occured on
31 dual-core machines) and other potential thread-safety issues.
34 Changes between 0.9.7i and 0.9.7j [04 May 2006]
36 *) Adapt fipsld and the build system to link against the validated FIPS
40 *) Fixes for VC++ 2005 build under Windows.
43 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
44 from a Windows bash shell such as MSYS. It is autodetected from the
45 "config" script when run from a VC++ environment. Modify standard VC++
46 build to use fipscanister.o from the GNU make build.
49 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
51 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
52 The value now differs depending on if you build for FIPS or not.
53 BEWARE! A program linked with a shared FIPSed libcrypto can't be
54 safely run with a non-FIPSed libcrypto, as it may crash because of
55 the difference induced by this change.
58 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
60 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
61 (part of SSL_OP_ALL). This option used to disable the
62 countermeasure against man-in-the-middle protocol-version
63 rollback in the SSL 2.0 server implementation, which is a bad
66 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
67 for Information Security, National Institute of Advanced Industrial
68 Science and Technology [AIST], Japan)]
70 *) Minimal support for X9.31 signatures and PSS padding modes. This is
71 mainly for FIPS compliance and not fully integrated at this stage.
74 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
75 the exponentiation using a fixed-length exponent. (Otherwise,
76 the information leaked through timing could expose the secret key
77 after many signatures; cf. Bleichenbacher's attack on DSA with
81 *) Make a new fixed-window mod_exp implementation the default for
82 RSA, DSA, and DH private-key operations so that the sequence of
83 squares and multiplies and the memory access pattern are
84 independent of the particular secret key. This will mitigate
85 cache-timing and potential related attacks.
87 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
88 and this is automatically used by BN_mod_exp_mont() if the new flag
89 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
90 will use this BN flag for private exponents unless the flag
91 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
92 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
94 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
96 *) Change the client implementation for SSLv23_method() and
97 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
98 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
99 (Previously, the SSL 2.0 backwards compatible Client Hello
100 message format would be used even with SSL_OP_NO_SSLv2.)
103 *) Add support for smime-type MIME parameter in S/MIME messages which some
107 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
108 a threadsafe manner. Modify rsa code to use new function and add calls
109 to dsa and dh code (which had race conditions before).
112 *) Include the fixed error library code in the C error file definitions
113 instead of fixing them up at runtime. This keeps the error code
117 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
119 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
122 *) Fixes for newer kerberos headers. NB: the casts are needed because
123 the 'length' field is signed on one version and unsigned on another
124 with no (?) obvious way to tell the difference, without these VC++
125 complains. Also the "definition" of FAR (blank) is no longer included
126 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
127 some needed definitions.
130 *) Undo Cygwin change.
133 *) Added support for proxy certificates according to RFC 3820.
134 Because they may be a security thread to unaware applications,
135 they must be explicitely allowed in run-time. See
136 docs/HOWTO/proxy_certificates.txt for further information.
139 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
141 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
142 server and client random values. Previously
143 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
144 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
146 This change has negligible security impact because:
148 1. Server and client random values still have 24 bytes of pseudo random
151 2. Server and client random values are sent in the clear in the initial
154 3. The master secret is derived using the premaster secret (48 bytes in
155 size for static RSA ciphersuites) as well as client server and random
158 The OpenSSL team would like to thank the UK NISCC for bringing this issue
161 [Stephen Henson, reported by UK NISCC]
163 *) Use Windows randomness collection on Cygwin.
166 *) Fix hang in EGD/PRNGD query when communication socket is closed
167 prematurely by EGD/PRNGD.
168 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
170 *) Prompt for pass phrases when appropriate for PKCS12 input format.
173 *) Back-port of selected performance improvements from development
174 branch, as well as improved support for PowerPC platforms.
177 *) Add lots of checks for memory allocation failure, error codes to indicate
178 failure and freeing up memory if a failure occurs.
179 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
181 *) Add new -passin argument to dgst.
184 *) Perform some character comparisons of different types in X509_NAME_cmp:
185 this is needed for some certificates that reencode DNs into UTF8Strings
186 (in violation of RFC3280) and can't or wont issue name rollover
190 *) Make an explicit check during certificate validation to see that
191 the CA setting in each certificate on the chain is correct. As a
192 side effect always do the following basic checks on extensions,
193 not just when there's an associated purpose to the check:
195 - if there is an unhandled critical extension (unless the user
196 has chosen to ignore this fault)
197 - if the path length has been exceeded (if one is set at all)
198 - that certain extensions fit the associated purpose (if one has
202 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
204 *) Avoid a race condition when CRLs are checked in a multi threaded
205 environment. This would happen due to the reordering of the revoked
206 entries during signature checking and serial number lookup. Now the
207 encoding is cached and the serial number sort performed under a lock.
208 Add new STACK function sk_is_sorted().
211 *) Add Delta CRL to the extension code.
214 *) Various fixes to s3_pkt.c so alerts are sent properly.
215 [David Holmes <d.holmes@f5.com>]
217 *) Reduce the chances of duplicate issuer name and serial numbers (in
218 violation of RFC3280) using the OpenSSL certificate creation utilities.
219 This is done by creating a random 64 bit value for the initial serial
220 number when a serial number file is created or when a self signed
221 certificate is created using 'openssl req -x509'. The initial serial
222 number file is created using 'openssl x509 -next_serial' in CA.pl
223 rather than being initialized to 1.
226 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
228 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
229 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
230 [Joe Orton, Steve Henson]
232 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
234 [Joe Orton, Steve Henson]
236 *) Make it possible to have multiple active certificates with the same
237 subject in the CA index file. This is done only if the keyword
238 'unique_subject' is set to 'no' in the main CA section (default
239 if 'CA_default') of the configuration file. The value is saved
240 with the database itself in a separate index attribute file,
241 named like the index file with '.attr' appended to the name.
244 *) X509 verify fixes. Disable broken certificate workarounds when
245 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
246 keyUsage extension present. Don't accept CRLs with unhandled critical
247 extensions: since verify currently doesn't process CRL extensions this
248 rejects a CRL with *any* critical extensions. Add new verify error codes
252 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
253 A clarification of RFC2560 will require the use of OCTET STRINGs and
254 some implementations cannot handle the current raw format. Since OpenSSL
255 copies and compares OCSP nonces as opaque blobs without any attempt at
256 parsing them this should not create any compatibility issues.
259 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
260 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
261 this HMAC (and other) operations are several times slower than OpenSSL
265 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
266 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
268 *) Use the correct content when signing type "other".
271 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
273 *) Fix various bugs revealed by running the NISCC test suite:
275 Stop out of bounds reads in the ASN1 code when presented with
276 invalid tags (CVE-2003-0543 and CVE-2003-0544).
278 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
280 If verify callback ignores invalid public key errors don't try to check
281 certificate signature with the NULL public key.
285 *) New -ignore_err option in ocsp application to stop the server
286 exiting on the first error in a request.
289 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
290 if the server requested one: as stated in TLS 1.0 and SSL 3.0
294 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
295 extra data after the compression methods not only for TLS 1.0
296 but also for SSL 3.0 (as required by the specification).
297 [Bodo Moeller; problem pointed out by Matthias Loepfe]
299 *) Change X509_certificate_type() to mark the key as exported/exportable
300 when it's 512 *bits* long, not 512 bytes.
303 *) Change AES_cbc_encrypt() so it outputs exact multiple of
304 blocks during encryption.
307 *) Various fixes to base64 BIO and non blocking I/O. On write
308 flushes were not handled properly if the BIO retried. On read
309 data was not being buffered properly and had various logic bugs.
310 This also affects blocking I/O when the data being decoded is a
314 *) Various S/MIME bugfixes and compatibility changes:
315 output correct application/pkcs7 MIME type if
316 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
317 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
318 of files as .eml work). Correctly handle very long lines in MIME
322 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
324 *) Countermeasure against the Klima-Pokorny-Rosa extension of
325 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
326 a protocol version number mismatch like a decryption error
327 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
330 *) Turn on RSA blinding by default in the default implementation
331 to avoid a timing attack. Applications that don't want it can call
332 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
333 They would be ill-advised to do so in most cases.
334 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
336 *) Change RSA blinding code so that it works when the PRNG is not
337 seeded (in this case, the secret RSA exponent is abused as
338 an unpredictable seed -- if it is not unpredictable, there
339 is no point in blinding anyway). Make RSA blinding thread-safe
340 by remembering the creator's thread ID in rsa->blinding and
341 having all other threads use local one-time blinding factors
342 (this requires more computation than sharing rsa->blinding, but
343 avoids excessive locking; and if an RSA object is not shared
344 between threads, blinding will still be very fast).
347 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
348 ENGINE as defaults for all supported algorithms irrespective of
349 the 'flags' parameter. 'flags' is now honoured, so applications
350 should make sure they are passing it correctly.
353 *) Target "mingw" now allows native Windows code to be generated in
354 the Cygwin environment as well as with the MinGW compiler.
357 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
359 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
360 via timing by performing a MAC computation even if incorrrect
361 block cipher padding has been found. This is a countermeasure
362 against active attacks where the attacker has to distinguish
363 between bad padding and a MAC verification error. (CVE-2003-0078)
365 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
366 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
367 Martin Vuagnoux (EPFL, Ilion)]
369 *) Make the no-err option work as intended. The intention with no-err
370 is not to have the whole error stack handling routines removed from
371 libcrypto, it's only intended to remove all the function name and
372 reason texts, thereby removing some of the footprint that may not
373 be interesting if those errors aren't displayed anyway.
375 NOTE: it's still possible for any application or module to have it's
376 own set of error texts inserted. The routines are there, just not
377 used by default when no-err is given.
380 *) Add support for FreeBSD on IA64.
381 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
383 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
384 Kerberos function mit_des_cbc_cksum(). Before this change,
385 the value returned by DES_cbc_cksum() was like the one from
386 mit_des_cbc_cksum(), except the bytes were swapped.
387 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
389 *) Allow an application to disable the automatic SSL chain building.
390 Before this a rather primitive chain build was always performed in
391 ssl3_output_cert_chain(): an application had no way to send the
392 correct chain if the automatic operation produced an incorrect result.
394 Now the chain builder is disabled if either:
396 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
398 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
400 The reasoning behind this is that an application would not want the
401 auto chain building to take place if extra chain certificates are
402 present and it might also want a means of sending no additional
403 certificates (for example the chain has two certificates and the
407 *) Add the possibility to build without the ENGINE framework.
408 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
410 *) Under Win32 gmtime() can return NULL: check return value in
411 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
414 *) DSA routines: under certain error conditions uninitialized BN objects
415 could be freed. Solution: make sure initialization is performed early
416 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
417 Nils Larsch <nla@trustcenter.de> via PR#459)
420 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
421 checked on reconnect on the client side, therefore session resumption
422 could still fail with a "ssl session id is different" error. This
423 behaviour is masked when SSL_OP_ALL is used due to
424 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
425 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
429 *) IA-32 assembler support enhancements: unified ELF targets, support
430 for SCO/Caldera platforms, fix for Cygwin shared build.
433 *) Add support for FreeBSD on sparc64. As a consequence, support for
434 FreeBSD on non-x86 processors is separate from x86 processors on
435 the config script, much like the NetBSD support.
436 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
438 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
440 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
443 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
444 code (06) was taken as the first octet of the session ID and the last
445 octet was ignored consequently. As a result SSLv2 client side session
446 caching could not have worked due to the session ID mismatch between
448 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
452 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
453 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
457 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
458 seems that in spite of existing for more than a year, many application
459 author have done nothing to provide the necessary callbacks, which
460 means that this particular engine will not work properly anywhere.
461 This is a very unfortunate situation which forces us, in the name
462 of usability, to give the hw_ncipher.c a static lock, which is part
464 NOTE: This is for the 0.9.7 series ONLY. This hack will never
465 appear in 0.9.8 or later. We EXPECT application authors to have
466 dealt properly with this when 0.9.8 is released (unless we actually
467 make such changes in the libcrypto locking code that changes will
468 have to be made anyway).
471 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
472 octets have been read, EOF or an error occurs. Without this change
473 some truncated ASN1 structures will not produce an error.
476 *) Disable Heimdal support, since it hasn't been fully implemented.
477 Still give the possibility to force the use of Heimdal, but with
478 warnings and a request that patches get sent to openssl-dev.
481 *) Add the VC-CE target, introduce the WINCE sysname, and add
482 INSTALL.WCE and appropriate conditionals to make it build.
483 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
485 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
486 cygssl-x.y.z.dll, where x, y and z are the major, minor and
487 edit numbers of the version.
488 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
490 *) Introduce safe string copy and catenation functions
491 (BUF_strlcpy() and BUF_strlcat()).
492 [Ben Laurie (CHATS) and Richard Levitte]
494 *) Avoid using fixed-size buffers for one-line DNs.
497 *) Add BUF_MEM_grow_clean() to avoid information leakage when
498 resizing buffers containing secrets, and use where appropriate.
501 *) Avoid using fixed size buffers for configuration file location.
504 *) Avoid filename truncation for various CA files.
507 *) Use sizeof in preference to magic numbers.
510 *) Avoid filename truncation in cert requests.
513 *) Add assertions to check for (supposedly impossible) buffer
517 *) Don't cache truncated DNS entries in the local cache (this could
518 potentially lead to a spoofing attack).
521 *) Fix various buffers to be large enough for hex/decimal
522 representations in a platform independent manner.
525 *) Add CRYPTO_realloc_clean() to avoid information leakage when
526 resizing buffers containing secrets, and use where appropriate.
529 *) Add BIO_indent() to avoid much slightly worrying code to do
533 *) Convert sprintf()/BIO_puts() to BIO_printf().
536 *) buffer_gets() could terminate with the buffer only half
540 *) Add assertions to prevent user-supplied crypto functions from
541 overflowing internal buffers by having large block sizes, etc.
544 *) New OPENSSL_assert() macro (similar to assert(), but enabled
548 *) Eliminate unused copy of key in RC4.
551 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
554 *) Fix off-by-one error in EGD path.
557 *) If RANDFILE path is too long, ignore instead of truncating.
560 *) Eliminate unused and incorrectly sized X.509 structure
564 *) Eliminate unused and dangerous function knumber().
567 *) Eliminate unused and dangerous structure, KSSL_ERR.
570 *) Protect against overlong session ID context length in an encoded
571 session object. Since these are local, this does not appear to be
575 *) Change from security patch (see 0.9.6e below) that did not affect
576 the 0.9.6 release series:
578 Remote buffer overflow in SSL3 protocol - an attacker could
579 supply an oversized master key in Kerberos-enabled versions.
583 *) Change the SSL kerb5 codes to match RFC 2712.
586 *) Make -nameopt work fully for req and add -reqopt switch.
587 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
589 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
590 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
592 *) Make sure tests can be performed even if the corresponding algorithms
593 have been removed entirely. This was also the last step to make
594 OpenSSL compilable with DJGPP under all reasonable conditions.
595 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
597 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
598 to allow version independent disabling of normally unselected ciphers,
599 which may be activated as a side-effect of selecting a single cipher.
601 (E.g., cipher list string "RSA" enables ciphersuites that are left
602 out of "ALL" because they do not provide symmetric encryption.
603 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
604 [Lutz Jaenicke, Bodo Moeller]
606 *) Add appropriate support for separate platform-dependent build
607 directories. The recommended way to make a platform-dependent
608 build directory is the following (tested on Linux), maybe with
611 # Place yourself outside of the OpenSSL source tree. In
612 # this example, the environment variable OPENSSL_SOURCE
613 # is assumed to contain the absolute OpenSSL source directory.
614 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
615 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
616 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
617 mkdir -p `dirname $F`
618 ln -s $OPENSSL_SOURCE/$F $F
621 To be absolutely sure not to disturb the source tree, a "make clean"
622 is a good thing. If it isn't successfull, don't worry about it,
623 it probably means the source directory is very clean.
626 *) Make sure any ENGINE control commands make local copies of string
627 pointers passed to them whenever necessary. Otherwise it is possible
628 the caller may have overwritten (or deallocated) the original string
629 data when a later ENGINE operation tries to use the stored values.
630 [Götz Babin-Ebell <babinebell@trustcenter.de>]
632 *) Improve diagnostics in file reading and command-line digests.
633 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
635 *) Add AES modes CFB and OFB to the object database. Correct an
636 error in AES-CFB decryption.
639 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
640 allows existing EVP_CIPHER_CTX structures to be reused after
641 calling EVP_*Final(). This behaviour is used by encryption
642 BIOs and some applications. This has the side effect that
643 applications must explicitly clean up cipher contexts with
644 EVP_CIPHER_CTX_cleanup() or they will leak memory.
647 *) Check the values of dna and dnb in bn_mul_recursive before calling
648 bn_mul_comba (a non zero value means the a or b arrays do not contain
649 n2 elements) and fallback to bn_mul_normal if either is not zero.
652 *) Fix escaping of non-ASCII characters when using the -subj option
653 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
656 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
657 form for "surname", serialNumber has no short form.
658 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
659 therefore remove "mail" short name for "internet 7".
660 The OID for unique identifiers in X509 certificates is
661 x500UniqueIdentifier, not uniqueIdentifier.
662 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
665 *) Add an "init" command to the ENGINE config module and auto initialize
666 ENGINEs. Without any "init" command the ENGINE will be initialized
667 after all ctrl commands have been executed on it. If init=1 the
668 ENGINE is initailized at that point (ctrls before that point are run
669 on the uninitialized ENGINE and after on the initialized one). If
670 init=0 then the ENGINE will not be iniatialized at all.
673 *) Fix the 'app_verify_callback' interface so that the user-defined
674 argument is actually passed to the callback: In the
675 SSL_CTX_set_cert_verify_callback() prototype, the callback
676 declaration has been changed from
679 int (*cb)(X509_STORE_CTX *,void *);
680 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
681 i=s->ctx->app_verify_callback(&ctx)
682 has been changed into
683 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
685 To update applications using SSL_CTX_set_cert_verify_callback(),
686 a dummy argument can be added to their callback functions.
687 [D. K. Smetters <smetters@parc.xerox.com>]
689 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
690 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
692 *) Add and OPENSSL_LOAD_CONF define which will cause
693 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
694 This allows older applications to transparently support certain
695 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
696 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
697 load the config file and OPENSSL_add_all_algorithms_conf() which will
698 always load it have also been added.
701 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
702 Adjust NIDs and EVP layer.
703 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
705 *) Config modules support in openssl utility.
707 Most commands now load modules from the config file,
708 though in a few (such as version) this isn't done
709 because it couldn't be used for anything.
711 In the case of ca and req the config file used is
712 the same as the utility itself: that is the -config
713 command line option can be used to specify an
717 *) Move default behaviour from OPENSSL_config(). If appname is NULL
718 use "openssl_conf" if filename is NULL use default openssl config file.
721 *) Add an argument to OPENSSL_config() to allow the use of an alternative
722 config section name. Add a new flag to tolerate a missing config file
723 and move code to CONF_modules_load_file().
726 *) Support for crypto accelerator cards from Accelerated Encryption
727 Processing, www.aep.ie. (Use engine 'aep')
728 The support was copied from 0.9.6c [engine] and adapted/corrected
729 to work with the new engine framework.
730 [AEP Inc. and Richard Levitte]
732 *) Support for SureWare crypto accelerator cards from Baltimore
733 Technologies. (Use engine 'sureware')
734 The support was copied from 0.9.6c [engine] and adapted
735 to work with the new engine framework.
738 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
739 make the newer ENGINE framework commands for the CHIL engine work.
740 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
742 *) Make it possible to produce shared libraries on ReliantUNIX.
743 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
745 *) Add the configuration target debug-linux-ppro.
746 Make 'openssl rsa' use the general key loading routines
747 implemented in apps.c, and make those routines able to
748 handle the key format FORMAT_NETSCAPE and the variant
750 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
752 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
753 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
755 *) Add -keyform to rsautl, and document -engine.
756 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
758 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
759 BIO_R_NO_SUCH_FILE error code rather than the generic
760 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
765 ERR_peek_last_error_line
766 ERR_peek_last_error_line_data.
770 ERR_peek_error_line_data,
771 but report on the latest error recorded rather than the first one
772 still in the error queue.
773 [Ben Laurie, Bodo Moeller]
775 *) default_algorithms option in ENGINE config module. This allows things
777 default_algorithms = ALL
778 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
781 *) Prelminary ENGINE config module.
784 *) New experimental application configuration code.
787 *) Change the AES code to follow the same name structure as all other
788 symmetric ciphers, and behave the same way. Move everything to
789 the directory crypto/aes, thereby obsoleting crypto/rijndael.
790 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
792 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
793 [Ben Laurie and Theo de Raadt]
795 *) Add option to output public keys in req command.
796 [Massimiliano Pala madwolf@openca.org]
798 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
799 (up to about 10% better than before for P-192 and P-224).
802 *) New functions/macros
804 SSL_CTX_set_msg_callback(ctx, cb)
805 SSL_CTX_set_msg_callback_arg(ctx, arg)
806 SSL_set_msg_callback(ssl, cb)
807 SSL_set_msg_callback_arg(ssl, arg)
809 to request calling a callback function
811 void cb(int write_p, int version, int content_type,
812 const void *buf, size_t len, SSL *ssl, void *arg)
814 whenever a protocol message has been completely received
815 (write_p == 0) or sent (write_p == 1). Here 'version' is the
816 protocol version according to which the SSL library interprets
817 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
818 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
819 the content type as defined in the SSL 3.0/TLS 1.0 protocol
820 specification (change_cipher_spec(20), alert(21), handshake(22)).
821 'buf' and 'len' point to the actual message, 'ssl' to the
822 SSL object, and 'arg' is the application-defined value set by
823 SSL[_CTX]_set_msg_callback_arg().
825 'openssl s_client' and 'openssl s_server' have new '-msg' options
826 to enable a callback that displays all protocol messages.
829 *) Change the shared library support so shared libraries are built as
830 soon as the corresponding static library is finished, and thereby get
831 openssl and the test programs linked against the shared library.
832 This still only happens when the keyword "shard" has been given to
833 the configuration scripts.
835 NOTE: shared library support is still an experimental thing, and
836 backward binary compatibility is still not guaranteed.
837 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
839 *) Add support for Subject Information Access extension.
840 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
842 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
843 additional bytes when new memory had to be allocated, not just
844 when reusing an existing buffer.
847 *) New command line and configuration option 'utf8' for the req command.
848 This allows field values to be specified as UTF8 strings.
851 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
852 runs for the former and machine-readable output for the latter.
855 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
856 of the e-mail address in the DN (i.e., it will go into a certificate
857 extension only). The new configuration file option 'email_in_dn = no'
859 [Massimiliano Pala madwolf@openca.org]
861 *) Change all functions with names starting with des_ to be starting
862 with DES_ instead. Add wrappers that are compatible with libdes,
863 but are named _ossl_old_des_*. Finally, add macros that map the
864 des_* symbols to the corresponding _ossl_old_des_* if libdes
865 compatibility is desired. If OpenSSL 0.9.6c compatibility is
866 desired, the des_* symbols will be mapped to DES_*, with one
869 Since we provide two compatibility mappings, the user needs to
870 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
871 compatibility is desired. The default (i.e., when that macro
872 isn't defined) is OpenSSL 0.9.6c compatibility.
874 There are also macros that enable and disable the support of old
875 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
876 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
877 are defined, the default will apply: to support the old des routines.
879 In either case, one must include openssl/des.h to get the correct
880 definitions. Do not try to just include openssl/des_old.h, that
883 NOTE: This is a major break of an old API into a new one. Software
884 authors are encouraged to switch to the DES_ style functions. Some
885 time in the future, des_old.h and the libdes compatibility functions
886 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
887 default), and then completely removed.
890 *) Test for certificates which contain unsupported critical extensions.
891 If such a certificate is found during a verify operation it is
892 rejected by default: this behaviour can be overridden by either
893 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
894 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
895 X509_supported_extension() has also been added which returns 1 if a
896 particular extension is supported.
899 *) Modify the behaviour of EVP cipher functions in similar way to digests
900 to retain compatibility with existing code.
903 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
904 compatibility with existing code. In particular the 'ctx' parameter does
905 not have to be to be initialized before the call to EVP_DigestInit() and
906 it is tidied up after a call to EVP_DigestFinal(). New function
907 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
908 EVP_MD_CTX_copy() changed to not require the destination to be
909 initialized valid and new function EVP_MD_CTX_copy_ex() added which
910 requires the destination to be valid.
912 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
913 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
916 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
917 so that complete 'Handshake' protocol structures are kept in memory
918 instead of overwriting 'msg_type' and 'length' with 'body' data.
921 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
922 [Massimo Santin via Richard Levitte]
924 *) Major restructuring to the underlying ENGINE code. This includes
925 reduction of linker bloat, separation of pure "ENGINE" manipulation
926 (initialisation, etc) from functionality dealing with implementations
927 of specific crypto iterfaces. This change also introduces integrated
928 support for symmetric ciphers and digest implementations - so ENGINEs
929 can now accelerate these by providing EVP_CIPHER and EVP_MD
930 implementations of their own. This is detailed in crypto/engine/README
931 as it couldn't be adequately described here. However, there are a few
932 API changes worth noting - some RSA, DSA, DH, and RAND functions that
933 were changed in the original introduction of ENGINE code have now
934 reverted back - the hooking from this code to ENGINE is now a good
935 deal more passive and at run-time, operations deal directly with
936 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
937 dereferencing through an ENGINE pointer any more. Also, the ENGINE
938 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
939 they were not being used by the framework as there is no concept of a
940 BIGNUM_METHOD and they could not be generalised to the new
941 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
942 ENGINE_cpy() has been removed as it cannot be consistently defined in
946 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
949 *) Change mkdef.pl to sort symbols that get the same entry number,
950 and make sure the automatically generated functions ERR_load_*
951 become part of libeay.num as well.
954 *) New function SSL_renegotiate_pending(). This returns true once
955 renegotiation has been requested (either SSL_renegotiate() call
956 or HelloRequest/ClientHello receveived from the peer) and becomes
957 false once a handshake has been completed.
958 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
959 sends a HelloRequest, but does not ensure that a handshake takes
960 place. SSL_renegotiate_pending() is useful for checking if the
961 client has followed the request.)
964 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
965 By default, clients may request session resumption even during
966 renegotiation (if session ID contexts permit); with this option,
967 session resumption is possible only in the first handshake.
969 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
970 more bits available for options that should not be part of
971 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
974 *) Add some demos for certificate and certificate request creation.
977 *) Make maximum certificate chain size accepted from the peer application
978 settable (SSL*_get/set_max_cert_list()), as proposed by
979 "Douglas E. Engert" <deengert@anl.gov>.
982 *) Add support for shared libraries for Unixware-7
983 (Boyd Lynn Gerber <gerberb@zenez.com>).
986 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
987 be done prior to destruction. Use this to unload error strings from
988 ENGINEs that load their own error strings. NB: This adds two new API
989 functions to "get" and "set" this destroy handler in an ENGINE.
992 *) Alter all existing ENGINE implementations (except "openssl" and
993 "openbsd") to dynamically instantiate their own error strings. This
994 makes them more flexible to be built both as statically-linked ENGINEs
995 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
996 Also, add stub code to each that makes building them as self-contained
997 shared-libraries easier (see README.ENGINE).
1000 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1001 implementations into applications that are completely implemented in
1002 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1003 commands that can be used to configure what shared-library to load and
1004 to control aspects of the way it is handled. Also, made an update to
1005 the README.ENGINE file that brings its information up-to-date and
1006 provides some information and instructions on the "dynamic" ENGINE
1007 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1010 *) Make it possible to unload ranges of ERR strings with a new
1011 "ERR_unload_strings" function.
1014 *) Add a copy() function to EVP_MD.
1017 *) Make EVP_MD routines take a context pointer instead of just the
1018 md_data void pointer.
1021 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1022 that the digest can only process a single chunk of data
1023 (typically because it is provided by a piece of
1024 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1025 is only going to provide a single chunk of data, and hence the
1026 framework needn't accumulate the data for oneshot drivers.
1029 *) As with "ERR", make it possible to replace the underlying "ex_data"
1030 functions. This change also alters the storage and management of global
1031 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1032 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1033 index counters. The API functions that use this state have been changed
1034 to take a "class_index" rather than pointers to the class's local STACK
1035 and counter, and there is now an API function to dynamically create new
1036 classes. This centralisation allows us to (a) plug a lot of the
1037 thread-safety problems that existed, and (b) makes it possible to clean
1038 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1039 such data would previously have always leaked in application code and
1040 workarounds were in place to make the memory debugging turn a blind eye
1041 to it. Application code that doesn't use this new function will still
1042 leak as before, but their memory debugging output will announce it now
1043 rather than letting it slide.
1045 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1046 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1047 has a return value to indicate success or failure.
1050 *) Make it possible to replace the underlying "ERR" functions such that the
1051 global state (2 LHASH tables and 2 locks) is only used by the "default"
1052 implementation. This change also adds two functions to "get" and "set"
1053 the implementation prior to it being automatically set the first time
1054 any other ERR function takes place. Ie. an application can call "get",
1055 pass the return value to a module it has just loaded, and that module
1056 can call its own "set" function using that value. This means the
1057 module's "ERR" operations will use (and modify) the error state in the
1058 application and not in its own statically linked copy of OpenSSL code.
1061 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1062 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1063 the operation, and provides a more encapsulated way for external code
1064 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1065 to use these functions rather than manually incrementing the counts.
1067 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1070 *) Add EVP test program.
1073 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1076 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1077 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1078 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1079 These allow a CRL to be built without having to access X509_CRL fields
1080 directly. Modify 'ca' application to use new functions.
1083 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1084 bug workarounds. Rollback attack detection is a security feature.
1085 The problem will only arise on OpenSSL servers when TLSv1 is not
1086 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1087 Software authors not wanting to support TLSv1 will have special reasons
1088 for their choice and can explicitly enable this option.
1089 [Bodo Moeller, Lutz Jaenicke]
1091 *) Rationalise EVP so it can be extended: don't include a union of
1092 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1093 (similar to those existing for EVP_CIPHER_CTX).
1098 EVP_MD_CTX_init(&md); /* new function call */
1099 EVP_DigestInit(&md, EVP_sha1());
1100 EVP_DigestUpdate(&md, in, len);
1101 EVP_DigestFinal(&md, out, NULL);
1102 EVP_MD_CTX_cleanup(&md); /* new function call */
1106 *) Make DES key schedule conform to the usual scheme, as well as
1107 correcting its structure. This means that calls to DES functions
1108 now have to pass a pointer to a des_key_schedule instead of a
1109 plain des_key_schedule (which was actually always a pointer
1112 des_key_schedule ks;
1114 des_set_key_checked(..., &ks);
1115 des_ncbc_encrypt(..., &ks, ...);
1117 (Note that a later change renames 'des_...' into 'DES_...'.)
1120 *) Initial reduction of linker bloat: the use of some functions, such as
1121 PEM causes large amounts of unused functions to be linked in due to
1122 poor organisation. For example pem_all.c contains every PEM function
1123 which has a knock on effect of linking in large amounts of (unused)
1124 ASN1 code. Grouping together similar functions and splitting unrelated
1125 functions prevents this.
1128 *) Cleanup of EVP macros.
1131 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1132 correct _ecb suffix.
1135 *) Add initial OCSP responder support to ocsp application. The
1136 revocation information is handled using the text based index
1137 use by the ca application. The responder can either handle
1138 requests generated internally, supplied in files (for example
1139 via a CGI script) or using an internal minimal server.
1142 *) Add configuration choices to get zlib compression for TLS.
1145 *) Changes to Kerberos SSL for RFC 2712 compliance:
1146 1. Implemented real KerberosWrapper, instead of just using
1147 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1148 2. Implemented optional authenticator field of KerberosWrapper.
1150 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1151 and authenticator structs; see crypto/krb5/.
1153 Generalized Kerberos calls to support multiple Kerberos libraries.
1154 [Vern Staats <staatsvr@asc.hpc.mil>,
1155 Jeffrey Altman <jaltman@columbia.edu>
1156 via Richard Levitte]
1158 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1159 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1160 values for each of the key sizes rather than having just
1161 parameters (and 'speed' generating keys each time).
1164 *) Speed up EVP routines.
1167 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1168 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1169 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1170 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1172 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1173 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1174 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1177 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1179 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1182 *) Added the OS2-EMX target.
1183 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1185 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1186 to support NCONF routines in extension code. New function CONF_set_nconf()
1187 to allow functions which take an NCONF to also handle the old LHASH
1188 structure: this means that the old CONF compatible routines can be
1189 retained (in particular wrt extensions) without having to duplicate the
1190 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1193 *) Enhance the general user interface with mechanisms for inner control
1194 and with possibilities to have yes/no kind of prompts.
1197 *) Change all calls to low level digest routines in the library and
1198 applications to use EVP. Add missing calls to HMAC_cleanup() and
1199 don't assume HMAC_CTX can be copied using memcpy().
1200 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1202 *) Add the possibility to control engines through control names but with
1203 arbitrary arguments instead of just a string.
1204 Change the key loaders to take a UI_METHOD instead of a callback
1205 function pointer. NOTE: this breaks binary compatibility with earlier
1206 versions of OpenSSL [engine].
1207 Adapt the nCipher code for these new conditions and add a card insertion
1211 *) Enhance the general user interface with mechanisms to better support
1212 dialog box interfaces, application-defined prompts, the possibility
1213 to use defaults (for example default passwords from somewhere else)
1214 and interrupts/cancellations.
1217 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1218 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1221 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1222 tidy up some unnecessarily weird code in 'sk_new()').
1223 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1225 *) Change the key loading routines for ENGINEs to use the same kind
1226 callback (pem_password_cb) as all other routines that need this
1230 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1231 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1232 than this minimum value is recommended.
1235 *) New random seeder for OpenVMS, using the system process statistics
1236 that are easily reachable.
1239 *) Windows apparently can't transparently handle global
1240 variables defined in DLLs. Initialisations such as:
1242 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1244 wont compile. This is used by the any applications that need to
1245 declare their own ASN1 modules. This was fixed by adding the option
1246 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1247 needed for static libraries under Win32.
1250 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1251 setting of purpose and trust fields. New X509_STORE trust and
1252 purpose functions and tidy up setting in other SSL functions.
1255 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1256 structure. These are inherited by X509_STORE_CTX when it is
1257 initialised. This allows various defaults to be set in the
1258 X509_STORE structure (such as flags for CRL checking and custom
1259 purpose or trust settings) for functions which only use X509_STORE_CTX
1260 internally such as S/MIME.
1262 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1263 trust settings if they are not set in X509_STORE. This allows X509_STORE
1264 purposes and trust (in S/MIME for example) to override any set by default.
1266 Add command line options for CRL checking to smime, s_client and s_server
1270 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1271 are set then the CRL is looked up in the X509_STORE structure and
1272 its validity and signature checked, then if the certificate is found
1273 in the CRL the verify fails with a revoked error.
1275 Various new CRL related callbacks added to X509_STORE_CTX structure.
1277 Command line options added to 'verify' application to support this.
1279 This needs some additional work, such as being able to handle multiple
1280 CRLs with different times, extension based lookup (rather than just
1281 by subject name) and ultimately more complete V2 CRL extension
1285 *) Add a general user interface API (crypto/ui/). This is designed
1286 to replace things like des_read_password and friends (backward
1287 compatibility functions using this new API are provided).
1288 The purpose is to remove prompting functions from the DES code
1289 section as well as provide for prompting through dialog boxes in
1290 a window system and the like.
1293 *) Add "ex_data" support to ENGINE so implementations can add state at a
1294 per-structure level rather than having to store it globally.
1297 *) Make it possible for ENGINE structures to be copied when retrieved by
1298 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1299 This causes the "original" ENGINE structure to act like a template,
1300 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1301 operational state can be localised to each ENGINE structure, despite the
1302 fact they all share the same "methods". New ENGINE structures returned in
1303 this case have no functional references and the return value is the single
1304 structural reference. This matches the single structural reference returned
1305 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1309 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1310 needs to match any other type at all we need to manually clear the
1314 *) Changes to the "openssl engine" utility to include;
1315 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1316 about an ENGINE's available control commands.
1317 - executing control commands from command line arguments using the
1318 '-pre' and '-post' switches. '-post' is only used if '-t' is
1319 specified and the ENGINE is successfully initialised. The syntax for
1320 the individual commands are colon-separated, for example;
1321 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1324 *) New dynamic control command support for ENGINEs. ENGINEs can now
1325 declare their own commands (numbers), names (strings), descriptions,
1326 and input types for run-time discovery by calling applications. A
1327 subset of these commands are implicitly classed as "executable"
1328 depending on their input type, and only these can be invoked through
1329 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1330 can be based on user input, config files, etc). The distinction is
1331 that "executable" commands cannot return anything other than a boolean
1332 result and can only support numeric or string input, whereas some
1333 discoverable commands may only be for direct use through
1334 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1335 pointers, or other custom uses. The "executable" commands are to
1336 support parameterisations of ENGINE behaviour that can be
1337 unambiguously defined by ENGINEs and used consistently across any
1338 OpenSSL-based application. Commands have been added to all the
1339 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1340 control over shared-library paths without source code alterations.
1343 *) Changed all ENGINE implementations to dynamically allocate their
1344 ENGINEs rather than declaring them statically. Apart from this being
1345 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1346 this also allows the implementations to compile without using the
1347 internal engine_int.h header.
1350 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1351 'const' value. Any code that should be able to modify a RAND_METHOD
1352 should already have non-const pointers to it (ie. they should only
1353 modify their own ones).
1356 *) Made a variety of little tweaks to the ENGINE code.
1357 - "atalla" and "ubsec" string definitions were moved from header files
1358 to C code. "nuron" string definitions were placed in variables
1359 rather than hard-coded - allowing parameterisation of these values
1360 later on via ctrl() commands.
1361 - Removed unused "#if 0"'d code.
1362 - Fixed engine list iteration code so it uses ENGINE_free() to release
1363 structural references.
1364 - Constified the RAND_METHOD element of ENGINE structures.
1365 - Constified various get/set functions as appropriate and added
1366 missing functions (including a catch-all ENGINE_cpy that duplicates
1367 all ENGINE values onto a new ENGINE except reference counts/state).
1368 - Removed NULL parameter checks in get/set functions. Setting a method
1369 or function to NULL is a way of cancelling out a previously set
1370 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1371 and doesn't justify the extra error symbols and code.
1372 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1373 flags from engine_int.h to engine.h.
1374 - Changed prototypes for ENGINE handler functions (init(), finish(),
1375 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1378 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1379 to the algorithm using long division. The binary algorithm can be
1380 used only if the modulus is odd. On 32-bit systems, it is faster
1381 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1382 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1383 up to 450 bits. In 64-bit environments, the binary algorithm
1384 appears to be advantageous for much longer moduli; here we use it
1385 for moduli up to 2048 bits.
1388 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1389 could not support the combine flag in choice fields.
1392 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1393 extensions from a certificate request to the certificate.
1396 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1397 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1398 file: this allows the display of the certificate about to be
1399 signed to be customised, to allow certain fields to be included
1400 or excluded and extension details. The old system didn't display
1401 multicharacter strings properly, omitted fields not in the policy
1402 and couldn't display additional details such as extensions.
1405 *) Function EC_POINTs_mul for multiple scalar multiplication
1406 of an arbitrary number of elliptic curve points
1407 \sum scalars[i]*points[i],
1408 optionally including the generator defined for the EC_GROUP:
1409 scalar*generator + \sum scalars[i]*points[i].
1411 EC_POINT_mul is a simple wrapper function for the typical case
1412 that the point list has just one item (besides the optional
1416 *) First EC_METHODs for curves over GF(p):
1418 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1419 operations and provides various method functions that can also
1420 operate with faster implementations of modular arithmetic.
1422 EC_GFp_mont_method() reuses most functions that are part of
1423 EC_GFp_simple_method, but uses Montgomery arithmetic.
1425 [Bodo Moeller; point addition and point doubling
1426 implementation directly derived from source code provided by
1427 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1429 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1430 crypto/ec/ec_lib.c):
1432 Curves are EC_GROUP objects (with an optional group generator)
1433 based on EC_METHODs that are built into the library.
1435 Points are EC_POINT objects based on EC_GROUP objects.
1437 Most of the framework would be able to handle curves over arbitrary
1438 finite fields, but as there are no obvious types for fields other
1439 than GF(p), some functions are limited to that for now.
1442 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1443 that the file contains a complete HTTP response.
1446 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1447 change the def and num file printf format specifier from "%-40sXXX"
1448 to "%-39s XXX". The latter will always guarantee a space after the
1449 field while the former will cause them to run together if the field
1450 is 40 of more characters long.
1453 *) Constify the cipher and digest 'method' functions and structures
1454 and modify related functions to take constant EVP_MD and EVP_CIPHER
1458 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1459 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1462 *) Modify EVP_Digest*() routines so they now return values. Although the
1463 internal software routines can never fail additional hardware versions
1467 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1469 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1470 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1473 ERR_R_NESTED_ASN1_ERROR
1475 ERR_R_MISSING_ASN1_EOS
1476 were 4 .. 9, conflicting with
1477 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1479 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1480 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1482 Add new error code 'ERR_R_INTERNAL_ERROR'.
1485 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1489 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1490 sets the subject name for a new request or supersedes the
1491 subject name in a given request. Formats that can be parsed are
1492 'CN=Some Name, OU=myOU, C=IT'
1494 'CN=Some Name/OU=myOU/C=IT'.
1496 Add options '-batch' and '-verbose' to 'openssl req'.
1497 [Massimiliano Pala <madwolf@hackmasters.net>]
1499 *) Introduce the possibility to access global variables through
1500 functions on platform were that's the best way to handle exporting
1501 global variables in shared libraries. To enable this functionality,
1502 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1503 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1504 is normally done by Configure or something similar).
1506 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1507 in the source file (foo.c) like this:
1509 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1510 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1512 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1513 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1515 OPENSSL_DECLARE_GLOBAL(int,foo);
1516 #define foo OPENSSL_GLOBAL_REF(foo)
1517 OPENSSL_DECLARE_GLOBAL(double,bar);
1518 #define bar OPENSSL_GLOBAL_REF(bar)
1520 The #defines are very important, and therefore so is including the
1521 header file everywhere where the defined globals are used.
1523 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1524 of ASN.1 items, but that structure is a bit different.
1526 The largest change is in util/mkdef.pl which has been enhanced with
1527 better and easier to understand logic to choose which symbols should
1528 go into the Windows .def files as well as a number of fixes and code
1529 cleanup (among others, algorithm keywords are now sorted
1530 lexicographically to avoid constant rewrites).
1533 *) In BN_div() keep a copy of the sign of 'num' before writing the
1534 result to 'rm' because if rm==num the value will be overwritten
1535 and produce the wrong result if 'num' is negative: this caused
1536 problems with BN_mod() and BN_nnmod().
1539 *) Function OCSP_request_verify(). This checks the signature on an
1540 OCSP request and verifies the signer certificate. The signer
1541 certificate is just checked for a generic purpose and OCSP request
1545 *) Add OCSP_check_validity() function to check the validity of OCSP
1546 responses. OCSP responses are prepared in real time and may only
1547 be a few seconds old. Simply checking that the current time lies
1548 between thisUpdate and nextUpdate max reject otherwise valid responses
1549 caused by either OCSP responder or client clock inaccuracy. Instead
1550 we allow thisUpdate and nextUpdate to fall within a certain period of
1551 the current time. The age of the response can also optionally be
1552 checked. Two new options -validity_period and -status_age added to
1556 *) If signature or public key algorithm is unrecognized print out its
1557 OID rather that just UNKNOWN.
1560 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1561 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1562 ID to be generated from the issuer certificate alone which can then be
1563 passed to OCSP_id_issuer_cmp().
1566 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1567 ASN1 modules to export functions returning ASN1_ITEM pointers
1568 instead of the ASN1_ITEM structures themselves. This adds several
1569 new macros which allow the underlying ASN1 function/structure to
1570 be accessed transparently. As a result code should not use ASN1_ITEM
1571 references directly (such as &X509_it) but instead use the relevant
1572 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1573 use of the new ASN1 code on platforms where exporting structures
1574 is problematical (for example in shared libraries) but exporting
1575 functions returning pointers to structures is not.
1578 *) Add support for overriding the generation of SSL/TLS session IDs.
1579 These callbacks can be registered either in an SSL_CTX or per SSL.
1580 The purpose of this is to allow applications to control, if they wish,
1581 the arbitrary values chosen for use as session IDs, particularly as it
1582 can be useful for session caching in multiple-server environments. A
1583 command-line switch for testing this (and any client code that wishes
1584 to use such a feature) has been added to "s_server".
1585 [Geoff Thorpe, Lutz Jaenicke]
1587 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1588 of the form '#if defined(...) || defined(...) || ...' and
1589 '#if !defined(...) && !defined(...) && ...'. This also avoids
1590 the growing number of special cases it was previously handling.
1593 *) Make all configuration macros available for application by making
1594 sure they are available in opensslconf.h, by giving them names starting
1595 with "OPENSSL_" to avoid conflicts with other packages and by making
1596 sure e_os2.h will cover all platform-specific cases together with
1598 Additionally, it is now possible to define configuration/platform-
1599 specific names (called "system identities"). In the C code, these
1600 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1601 macro with the name beginning with "OPENSSL_SYS_", which is determined
1602 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1606 *) New option -set_serial to 'req' and 'x509' this allows the serial
1607 number to use to be specified on the command line. Previously self
1608 signed certificates were hard coded with serial number 0 and the
1609 CA options of 'x509' had to use a serial number in a file which was
1613 *) New options to 'ca' utility to support V2 CRL entry extensions.
1614 Currently CRL reason, invalidity date and hold instruction are
1615 supported. Add new CRL extensions to V3 code and some new objects.
1618 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1619 disable standard block padding (aka PKCS#5 padding) in the EVP
1620 API, which was previously mandatory. This means that the data is
1621 not padded in any way and so the total length much be a multiple
1622 of the block size, otherwise an error occurs.
1625 *) Initial (incomplete) OCSP SSL support.
1628 *) New function OCSP_parse_url(). This splits up a URL into its host,
1629 port and path components: primarily to parse OCSP URLs. New -url
1630 option to ocsp utility.
1633 *) New nonce behavior. The return value of OCSP_check_nonce() now
1634 reflects the various checks performed. Applications can decide
1635 whether to tolerate certain situations such as an absent nonce
1636 in a response when one was present in a request: the ocsp application
1637 just prints out a warning. New function OCSP_add1_basic_nonce()
1638 this is to allow responders to include a nonce in a response even if
1639 the request is nonce-less.
1642 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1643 skipped when using openssl x509 multiple times on a single input file,
1644 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1647 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1648 set string type: to handle setting ASN1_TIME structures. Fix ca
1649 utility to correctly initialize revocation date of CRLs.
1652 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1653 the clients preferred ciphersuites and rather use its own preferences.
1654 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1655 Internet Explorer by ensuring unchanged hash method during stepup.
1656 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1659 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1660 to aes and add a new 'exist' option to print out symbols that don't
1664 *) Additional options to ocsp utility to allow flags to be set and
1665 additional certificates supplied.
1668 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1669 OCSP client a number of certificate to only verify the response
1673 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1674 handle the new API. Currently only ECB, CBC modes supported. Add new
1677 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1678 Encryption Standard (AES) Ciphersuites for Transport Layer
1679 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1680 not enabled by default and were not part of the "ALL" ciphersuite
1681 alias because they were not yet official; they could be
1682 explicitly requested by specifying the "AESdraft" ciphersuite
1683 group alias. In the final release of OpenSSL 0.9.7, the group
1684 alias is called "AES" and is part of "ALL".)
1685 [Ben Laurie, Steve Henson, Bodo Moeller]
1687 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1688 request to response.
1691 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1692 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1693 extract information from a certificate request. OCSP_response_create()
1694 creates a response and optionally adds a basic response structure.
1695 OCSP_basic_add1_status() adds a complete single response to a basic
1696 response and returns the OCSP_SINGLERESP structure just added (to allow
1697 extensions to be included for example). OCSP_basic_add1_cert() adds a
1698 certificate to a basic response and OCSP_basic_sign() signs a basic
1699 response with various flags. New helper functions ASN1_TIME_check()
1700 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1701 (converts ASN1_TIME to GeneralizedTime).
1704 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1705 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1706 structure from a certificate. X509_pubkey_digest() digests the public_key
1707 contents: this is used in various key identifiers.
1710 *) Make sk_sort() tolerate a NULL argument.
1711 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1713 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1714 passed by the function are trusted implicitly. If any of them signed the
1715 response then it is assumed to be valid and is not verified.
1718 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1719 to data. This was previously part of the PKCS7 ASN1 code. This
1720 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1721 [Steve Henson, reported by Kenneth R. Robinette
1722 <support@securenetterm.com>]
1724 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1725 routines: without these tracing memory leaks is very painful.
1726 Fix leaks in PKCS12 and PKCS7 routines.
1729 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1730 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1731 effectively meant GeneralizedTime would never be used. Now it
1732 is initialised to -1 but X509_time_adj() now has to check the value
1733 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1734 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1735 [Steve Henson, reported by Kenneth R. Robinette
1736 <support@securenetterm.com>]
1738 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1739 result in a zero length in the ASN1_INTEGER structure which was
1740 not consistent with the structure when d2i_ASN1_INTEGER() was used
1741 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1742 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1743 where it did not print out a minus for negative ASN1_INTEGER.
1746 *) Add summary printout to ocsp utility. The various functions which
1747 convert status values to strings have been renamed to:
1748 OCSP_response_status_str(), OCSP_cert_status_str() and
1749 OCSP_crl_reason_str() and are no longer static. New options
1750 to verify nonce values and to disable verification. OCSP response
1751 printout format cleaned up.
1754 *) Add additional OCSP certificate checks. These are those specified
1755 in RFC2560. This consists of two separate checks: the CA of the
1756 certificate being checked must either be the OCSP signer certificate
1757 or the issuer of the OCSP signer certificate. In the latter case the
1758 OCSP signer certificate must contain the OCSP signing extended key
1759 usage. This check is performed by attempting to match the OCSP
1760 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1761 in the OCSP_CERTID structures of the response.
1764 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1765 and related routines. This uses the standard OpenSSL certificate
1766 verify routines to perform initial checks (just CA validity) and
1767 to obtain the certificate chain. Then additional checks will be
1768 performed on the chain. Currently the root CA is checked to see
1769 if it is explicitly trusted for OCSP signing. This is used to set
1770 a root CA as a global signing root: that is any certificate that
1771 chains to that CA is an acceptable OCSP signing certificate.
1774 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1775 extensions from a separate configuration file.
1776 As when reading extensions from the main configuration file,
1777 the '-extensions ...' option may be used for specifying the
1779 [Massimiliano Pala <madwolf@comune.modena.it>]
1781 *) New OCSP utility. Allows OCSP requests to be generated or
1782 read. The request can be sent to a responder and the output
1783 parsed, outputed or printed in text form. Not complete yet:
1784 still needs to check the OCSP response validity.
1787 *) New subcommands for 'openssl ca':
1788 'openssl ca -status <serial>' prints the status of the cert with
1789 the given serial number (according to the index file).
1790 'openssl ca -updatedb' updates the expiry status of certificates
1792 [Massimiliano Pala <madwolf@comune.modena.it>]
1794 *) New '-newreq-nodes' command option to CA.pl. This is like
1795 '-newreq', but calls 'openssl req' with the '-nodes' option
1796 so that the resulting key is not encrypted.
1797 [Damien Miller <djm@mindrot.org>]
1799 *) New configuration for the GNU Hurd.
1800 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1802 *) Initial code to implement OCSP basic response verify. This
1803 is currently incomplete. Currently just finds the signer's
1804 certificate and verifies the signature on the response.
1807 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1808 value of OPENSSLDIR. This is available via the new '-d' option
1809 to 'openssl version', and is also included in 'openssl version -a'.
1812 *) Allowing defining memory allocation callbacks that will be given
1813 file name and line number information in additional arguments
1814 (a const char* and an int). The basic functionality remains, as
1815 well as the original possibility to just replace malloc(),
1816 realloc() and free() by functions that do not know about these
1817 additional arguments. To register and find out the current
1818 settings for extended allocation functions, the following
1819 functions are provided:
1821 CRYPTO_set_mem_ex_functions
1822 CRYPTO_set_locked_mem_ex_functions
1823 CRYPTO_get_mem_ex_functions
1824 CRYPTO_get_locked_mem_ex_functions
1826 These work the same way as CRYPTO_set_mem_functions and friends.
1827 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1828 extended allocation function is enabled.
1829 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1830 a conventional allocation function is enabled.
1831 [Richard Levitte, Bodo Moeller]
1833 *) Finish off removing the remaining LHASH function pointer casts.
1834 There should no longer be any prototype-casting required when using
1835 the LHASH abstraction, and any casts that remain are "bugs". See
1836 the callback types and macros at the head of lhash.h for details
1837 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1840 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1841 If /dev/[u]random devices are not available or do not return enough
1842 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1844 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1845 /etc/entropy will be queried once each in this sequence, quering stops
1846 when enough entropy was collected without querying more sockets.
1849 *) Change the Unix RAND_poll() variant to be able to poll several
1850 random devices, as specified by DEVRANDOM, until a sufficient amount
1851 of data has been collected. We spend at most 10 ms on each file
1852 (select timeout) and read in non-blocking mode. DEVRANDOM now
1853 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1854 (previously it was just the string "/dev/urandom"), so on typical
1855 platforms the 10 ms delay will never occur.
1856 Also separate out the Unix variant to its own file, rand_unix.c.
1857 For VMS, there's a currently-empty rand_vms.c.
1860 *) Move OCSP client related routines to ocsp_cl.c. These
1861 provide utility functions which an application needing
1862 to issue a request to an OCSP responder and analyse the
1863 response will typically need: as opposed to those which an
1864 OCSP responder itself would need which will be added later.
1866 OCSP_request_sign() signs an OCSP request with an API similar
1867 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1868 response. OCSP_response_get1_basic() extracts basic response
1869 from response. OCSP_resp_find_status(): finds and extracts status
1870 information from an OCSP_CERTID structure (which will be created
1871 when the request structure is built). These are built from lower
1872 level functions which work on OCSP_SINGLERESP structures but
1873 wont normally be used unless the application wishes to examine
1874 extensions in the OCSP response for example.
1876 Replace nonce routines with a pair of functions.
1877 OCSP_request_add1_nonce() adds a nonce value and optionally
1878 generates a random value. OCSP_check_nonce() checks the
1879 validity of the nonce in an OCSP response.
1882 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1883 This doesn't copy the supplied OCSP_CERTID and avoids the
1884 need to free up the newly created id. Change return type
1885 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1886 This can then be used to add extensions to the request.
1887 Deleted OCSP_request_new(), since most of its functionality
1888 is now in OCSP_REQUEST_new() (and the case insensitive name
1889 clash) apart from the ability to set the request name which
1890 will be added elsewhere.
1893 *) Update OCSP API. Remove obsolete extensions argument from
1894 various functions. Extensions are now handled using the new
1895 OCSP extension code. New simple OCSP HTTP function which
1896 can be used to send requests and parse the response.
1899 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1900 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1901 uses the special reorder version of SET OF to sort the attributes
1902 and reorder them to match the encoded order. This resolves a long
1903 standing problem: a verify on a PKCS7 structure just after signing
1904 it used to fail because the attribute order did not match the
1905 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1906 it uses the received order. This is necessary to tolerate some broken
1907 software that does not order SET OF. This is handled by encoding
1908 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1909 to produce the required SET OF.
1912 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1913 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1914 files to get correct declarations of the ASN.1 item variables.
1917 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1918 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1919 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1920 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1921 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1922 ASN1_ITEM and no wrapper functions.
1925 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1926 replace the old function pointer based I/O routines. Change most of
1927 the *_d2i_bio() and *_d2i_fp() functions to use these.
1930 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1931 lines, recognice more "algorithms" that can be deselected, and make
1932 it complain about algorithm deselection that isn't recognised.
1935 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1936 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1937 to use new functions. Add NO_ASN1_OLD which can be set to remove
1938 some old style ASN1 functions: this can be used to determine if old
1939 code will still work when these eventually go away.
1942 *) New extension functions for OCSP structures, these follow the
1943 same conventions as certificates and CRLs.
1946 *) New function X509V3_add1_i2d(). This automatically encodes and
1947 adds an extension. Its behaviour can be customised with various
1948 flags to append, replace or delete. Various wrappers added for
1949 certifcates and CRLs.
1952 *) Fix to avoid calling the underlying ASN1 print routine when
1953 an extension cannot be parsed. Correct a typo in the
1954 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1957 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1958 entries for variables.
1961 *) Add functionality to apps/openssl.c for detecting locking
1962 problems: As the program is single-threaded, all we have
1963 to do is register a locking callback using an array for
1964 storing which locks are currently held by the program.
1967 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1968 SSL_get_ex_data_X509_STORE_idx(), which is used in
1969 ssl_verify_cert_chain() and thus can be called at any time
1970 during TLS/SSL handshakes so that thread-safety is essential.
1971 Unfortunately, the ex_data design is not at all suited
1972 for multi-threaded use, so it probably should be abolished.
1975 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1976 [Broadcom, tweaked and integrated by Geoff Thorpe]
1978 *) Move common extension printing code to new function
1979 X509V3_print_extensions(). Reorganise OCSP print routines and
1980 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1983 *) New function X509_signature_print() to remove duplication in some
1987 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1988 set (this was treated exactly the same as SET OF previously). This
1989 is used to reorder the STACK representing the structure to match the
1990 encoding. This will be used to get round a problem where a PKCS7
1991 structure which was signed could not be verified because the STACK
1992 order did not reflect the encoded order.
1995 *) Reimplement the OCSP ASN1 module using the new code.
1998 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1999 for its ASN1 operations. The old style function pointers still exist
2000 for now but they will eventually go away.
2003 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2004 completely replaces the old ASN1 functionality with a table driven
2005 encoder and decoder which interprets an ASN1_ITEM structure describing
2006 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2007 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2008 has also been converted to the new form.
2011 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2012 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2013 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2014 for negative moduli.
2017 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2018 of not touching the result's sign bit.
2021 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2025 *) Changed the LHASH code to use prototypes for callbacks, and created
2026 macros to declare and implement thin (optionally static) functions
2027 that provide type-safety and avoid function pointer casting for the
2028 type-specific callbacks.
2031 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2033 [Veers Staats <staatsvr@asc.hpc.mil>,
2034 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2036 *) Reformat the FAQ so the different questions and answers can be divided
2037 in sections depending on the subject.
2040 *) Have the zlib compression code load ZLIB.DLL dynamically under
2044 *) New function BN_mod_sqrt for computing square roots modulo a prime
2045 (using the probabilistic Tonelli-Shanks algorithm unless
2046 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2047 be handled deterministically).
2048 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2050 *) Make BN_mod_inverse faster by explicitly handling small quotients
2051 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2052 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2055 *) New function BN_kronecker.
2058 *) Fix BN_gcd so that it works on negative inputs; the result is
2059 positive unless both parameters are zero.
2060 Previously something reasonably close to an infinite loop was
2061 possible because numbers could be growing instead of shrinking
2062 in the implementation of Euclid's algorithm.
2065 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2066 sign of the number in question.
2068 Fix BN_is_word(a,w) to work correctly for w == 0.
2070 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2071 because its test if the absolute value of 'a' equals 'w'.
2072 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2073 it exists mostly for use in the implementations of BN_is_zero(),
2074 BN_is_one(), and BN_is_word().
2077 *) New function BN_swap.
2080 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2081 the exponentiation functions are more likely to produce reasonable
2082 results on negative inputs.
2085 *) Change BN_mod_mul so that the result is always non-negative.
2086 Previously, it could be negative if one of the factors was negative;
2087 I don't think anyone really wanted that behaviour.
2090 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2091 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2092 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2093 and add new functions:
2102 BN_mod_lshift1_quick
2106 These functions always generate non-negative results.
2108 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2109 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2111 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2112 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2113 be reduced modulo m.
2114 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2117 The following entry accidentily appeared in the CHANGES file
2118 distributed with OpenSSL 0.9.7. The modifications described in
2119 it do *not* apply to OpenSSL 0.9.7.
2121 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2122 was actually never needed) and in BN_mul(). The removal in BN_mul()
2123 required a small change in bn_mul_part_recursive() and the addition
2124 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2125 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2126 bn_sub_words() and bn_add_words() except they take arrays with
2131 *) In 'openssl passwd', verify passwords read from the terminal
2132 unless the '-salt' option is used (which usually means that
2133 verification would just waste user's time since the resulting
2134 hash is going to be compared with some given password hash)
2135 or the new '-noverify' option is used.
2137 This is an incompatible change, but it does not affect
2138 non-interactive use of 'openssl passwd' (passwords on the command
2139 line, '-stdin' option, '-in ...' option) and thus should not
2143 *) Remove all references to RSAref, since there's no more need for it.
2146 *) Make DSO load along a path given through an environment variable
2147 (SHLIB_PATH) with shl_load().
2150 *) Constify the ENGINE code as a result of BIGNUM constification.
2151 Also constify the RSA code and most things related to it. In a
2152 few places, most notable in the depth of the ASN.1 code, ugly
2153 casts back to non-const were required (to be solved at a later
2157 *) Make it so the openssl application has all engines loaded by default.
2160 *) Constify the BIGNUM routines a little more.
2163 *) Add the following functions:
2165 ENGINE_load_cswift()
2167 ENGINE_load_atalla()
2169 ENGINE_load_builtin_engines()
2171 That way, an application can itself choose if external engines that
2172 are built-in in OpenSSL shall ever be used or not. The benefit is
2173 that applications won't have to be linked with libdl or other dso
2174 libraries unless it's really needed.
2176 Changed 'openssl engine' to load all engines on demand.
2177 Changed the engine header files to avoid the duplication of some
2178 declarations (they differed!).
2181 *) 'openssl engine' can now list capabilities.
2184 *) Better error reporting in 'openssl engine'.
2187 *) Never call load_dh_param(NULL) in s_server.
2190 *) Add engine application. It can currently list engines by name and
2191 identity, and test if they are actually available.
2194 *) Improve RPM specification file by forcing symbolic linking and making
2195 sure the installed documentation is also owned by root.root.
2196 [Damien Miller <djm@mindrot.org>]
2198 *) Give the OpenSSL applications more possibilities to make use of
2199 keys (public as well as private) handled by engines.
2202 *) Add OCSP code that comes from CertCo.
2205 *) Add VMS support for the Rijndael code.
2208 *) Added untested support for Nuron crypto accelerator.
2211 *) Add support for external cryptographic devices. This code was
2212 previously distributed separately as the "engine" branch.
2213 [Geoff Thorpe, Richard Levitte]
2215 *) Rework the filename-translation in the DSO code. It is now possible to
2216 have far greater control over how a "name" is turned into a filename
2217 depending on the operating environment and any oddities about the
2218 different shared library filenames on each system.
2221 *) Support threads on FreeBSD-elf in Configure.
2224 *) Fix for SHA1 assembly problem with MASM: it produces
2225 warnings about corrupt line number information when assembling
2226 with debugging information. This is caused by the overlapping
2228 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2231 NCONF_get_number() has no error checking at all. As a replacement,
2232 NCONF_get_number_e() is defined (_e for "error checking") and is
2233 promoted strongly. The old NCONF_get_number is kept around for
2234 binary backward compatibility.
2235 Make it possible for methods to load from something other than a BIO,
2236 by providing a function pointer that is given a name instead of a BIO.
2237 For example, this could be used to load configuration data from an
2241 *) Fix for non blocking accept BIOs. Added new I/O special reason
2242 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2243 with non blocking I/O was not possible because no retry code was
2244 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2248 *) Added the beginnings of Rijndael support.
2251 *) Fix for bug in DirectoryString mask setting. Add support for
2252 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2253 to allow certificate printing to more controllable, additional
2254 'certopt' option to 'x509' to allow new printing options to be
2258 *) Clean old EAY MD5 hack from e_os.h.
2261 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2263 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2264 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2265 [Joe Orton, Steve Henson]
2267 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2269 *) Fix additional bug revealed by the NISCC test suite:
2271 Stop bug triggering large recursion when presented with
2272 certain ASN.1 tags (CVE-2003-0851)
2275 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2277 *) Fix various bugs revealed by running the NISCC test suite:
2279 Stop out of bounds reads in the ASN1 code when presented with
2280 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2282 If verify callback ignores invalid public key errors don't try to check
2283 certificate signature with the NULL public key.
2287 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2288 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2292 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2293 extra data after the compression methods not only for TLS 1.0
2294 but also for SSL 3.0 (as required by the specification).
2295 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2297 *) Change X509_certificate_type() to mark the key as exported/exportable
2298 when it's 512 *bits* long, not 512 bytes.
2301 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2303 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2304 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2305 a protocol version number mismatch like a decryption error
2306 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2309 *) Turn on RSA blinding by default in the default implementation
2310 to avoid a timing attack. Applications that don't want it can call
2311 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2312 They would be ill-advised to do so in most cases.
2313 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2315 *) Change RSA blinding code so that it works when the PRNG is not
2316 seeded (in this case, the secret RSA exponent is abused as
2317 an unpredictable seed -- if it is not unpredictable, there
2318 is no point in blinding anyway). Make RSA blinding thread-safe
2319 by remembering the creator's thread ID in rsa->blinding and
2320 having all other threads use local one-time blinding factors
2321 (this requires more computation than sharing rsa->blinding, but
2322 avoids excessive locking; and if an RSA object is not shared
2323 between threads, blinding will still be very fast).
2326 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2328 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2329 via timing by performing a MAC computation even if incorrrect
2330 block cipher padding has been found. This is a countermeasure
2331 against active attacks where the attacker has to distinguish
2332 between bad padding and a MAC verification error. (CVE-2003-0078)
2334 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2335 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2336 Martin Vuagnoux (EPFL, Ilion)]
2338 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2340 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2341 memory from it's contents. This is done with a counter that will
2342 place alternating values in each byte. This can be used to solve
2343 two issues: 1) the removal of calls to memset() by highly optimizing
2344 compilers, and 2) cleansing with other values than 0, since those can
2345 be read through on certain media, for example a swap space on disk.
2348 *) Bugfix: client side session caching did not work with external caching,
2349 because the session->cipher setting was not restored when reloading
2350 from the external cache. This problem was masked, when
2351 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2352 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2355 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2356 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2357 [Zeev Lieber <zeev-l@yahoo.com>]
2359 *) Undo an undocumented change introduced in 0.9.6e which caused
2360 repeated calls to OpenSSL_add_all_ciphers() and
2361 OpenSSL_add_all_digests() to be ignored, even after calling
2365 *) Change the default configuration reader to deal with last line not
2366 being properly terminated.
2369 *) Change X509_NAME_cmp() so it applies the special rules on handling
2370 DN values that are of type PrintableString, as well as RDNs of type
2371 emailAddress where the value has the type ia5String.
2372 [stefank@valicert.com via Richard Levitte]
2374 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2375 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2376 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2377 the bitwise-OR of the two for use by the majority of applications
2378 wanting this behaviour, and update the docs. The documented
2379 behaviour and actual behaviour were inconsistent and had been
2380 changing anyway, so this is more a bug-fix than a behavioural
2382 [Geoff Thorpe, diagnosed by Nadav Har'El]
2384 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2385 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2388 *) Fix initialization code race conditions in
2389 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2390 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2391 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2392 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2393 ssl2_get_cipher_by_char(),
2394 ssl3_get_cipher_by_char().
2395 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2397 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2398 the cached sessions are flushed, as the remove_cb() might use ex_data
2399 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2400 (see [openssl.org #212]).
2401 [Geoff Thorpe, Lutz Jaenicke]
2403 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2404 length, instead of the encoding length to d2i_ASN1_OBJECT.
2407 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2409 *) [In 0.9.6g-engine release:]
2410 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2411 [Lynn Gazis <lgazis@rainbow.com>]
2413 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2415 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2416 and get fix the header length calculation.
2417 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2418 Alon Kantor <alonk@checkpoint.com> (and others),
2421 *) Use proper error handling instead of 'assertions' in buffer
2422 overflow checks added in 0.9.6e. This prevents DoS (the
2423 assertions could call abort()).
2424 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2426 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2428 *) Add various sanity checks to asn1_get_length() to reject
2429 the ASN1 length bytes if they exceed sizeof(long), will appear
2430 negative or the content length exceeds the length of the
2432 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2434 *) Fix cipher selection routines: ciphers without encryption had no flags
2435 for the cipher strength set and where therefore not handled correctly
2436 by the selection routines (PR #130).
2439 *) Fix EVP_dsa_sha macro.
2443 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2444 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2445 that was added in OpenSSL 0.9.6d.
2447 As the countermeasure turned out to be incompatible with some
2448 broken SSL implementations, the new option is part of SSL_OP_ALL.
2449 SSL_OP_ALL is usually employed when compatibility with weird SSL
2450 implementations is desired (e.g. '-bugs' option to 's_client' and
2451 's_server'), so the new option is automatically set in many
2455 *) Changes in security patch:
2457 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2458 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2459 Air Force Materiel Command, USAF, under agreement number
2462 *) Add various sanity checks to asn1_get_length() to reject
2463 the ASN1 length bytes if they exceed sizeof(long), will appear
2464 negative or the content length exceeds the length of the
2465 supplied buffer. (CVE-2002-0659)
2466 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2468 *) Assertions for various potential buffer overflows, not known to
2470 [Ben Laurie (CHATS)]
2472 *) Various temporary buffers to hold ASCII versions of integers were
2473 too small for 64 bit platforms. (CVE-2002-0655)
2474 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2476 *) Remote buffer overflow in SSL3 protocol - an attacker could
2477 supply an oversized session ID to a client. (CVE-2002-0656)
2478 [Ben Laurie (CHATS)]
2480 *) Remote buffer overflow in SSL2 protocol - an attacker could
2481 supply an oversized client master key. (CVE-2002-0656)
2482 [Ben Laurie (CHATS)]
2484 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2486 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2487 encoded as NULL) with id-dsa-with-sha1.
2488 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2490 *) Check various X509_...() return values in apps/req.c.
2491 [Nils Larsch <nla@trustcenter.de>]
2493 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2494 an end-of-file condition would erronously be flagged, when the CRLF
2495 was just at the end of a processed block. The bug was discovered when
2496 processing data through a buffering memory BIO handing the data to a
2497 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2498 <ptsekov@syntrex.com> and Nedelcho Stanev.
2501 *) Implement a countermeasure against a vulnerability recently found
2502 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2503 before application data chunks to avoid the use of known IVs
2504 with data potentially chosen by the attacker.
2507 *) Fix length checks in ssl3_get_client_hello().
2510 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2511 to prevent ssl3_read_internal() from incorrectly assuming that
2512 ssl3_read_bytes() found application data while handshake
2513 processing was enabled when in fact s->s3->in_read_app_data was
2514 merely automatically cleared during the initial handshake.
2515 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2517 *) Fix object definitions for Private and Enterprise: they were not
2518 recognized in their shortname (=lowercase) representation. Extend
2519 obj_dat.pl to issue an error when using undefined keywords instead
2520 of silently ignoring the problem (Svenning Sorensen
2521 <sss@sss.dnsalias.net>).
2524 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2525 generators, i.e. generators other than 2 and 5. (Previously, the
2526 code did not properly initialise the 'add' and 'rem' values to
2527 BN_generate_prime().)
2529 In the new general case, we do not insist that 'generator' is
2530 actually a primitive root: This requirement is rather pointless;
2531 a generator of the order-q subgroup is just as good, if not
2535 *) Map new X509 verification errors to alerts. Discovered and submitted by
2536 Tom Wu <tom@arcot.com>.
2539 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2540 returning non-zero before the data has been completely received
2541 when using non-blocking I/O.
2542 [Bodo Moeller; problem pointed out by John Hughes]
2544 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2545 [Ben Laurie, Lutz Jaenicke]
2547 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2548 Yoram Zahavi <YoramZ@gilian.com>).
2551 *) Add information about CygWin 1.3 and on, and preserve proper
2552 configuration for the versions before that.
2553 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2555 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2556 check whether we deal with a copy of a session and do not delete from
2557 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2558 <izhar@checkpoint.com>.
2561 *) Do not store session data into the internal session cache, if it
2562 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2563 flag is set). Proposed by Aslam <aslam@funk.com>.
2566 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2570 *) [In 0.9.6d-engine release:]
2571 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2572 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2574 *) Add the configuration target linux-s390x.
2575 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2577 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2578 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2579 variable as an indication that a ClientHello message has been
2580 received. As the flag value will be lost between multiple
2581 invocations of ssl3_accept when using non-blocking I/O, the
2582 function may not be aware that a handshake has actually taken
2583 place, thus preventing a new session from being added to the
2586 To avoid this problem, we now set s->new_session to 2 instead of
2587 using a local variable.
2588 [Lutz Jaenicke, Bodo Moeller]
2590 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2591 if the SSL_R_LENGTH_MISMATCH error is detected.
2592 [Geoff Thorpe, Bodo Moeller]
2594 *) New 'shared_ldflag' column in Configure platform table.
2597 *) Fix EVP_CIPHER_mode macro.
2598 ["Dan S. Camper" <dan@bti.net>]
2600 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2601 type, we must throw them away by setting rr->length to 0.
2602 [D P Chang <dpc@qualys.com>]
2604 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2606 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2607 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2608 worked incorrectly for those cases where range = 10..._2 and
2609 3*range is two bits longer than range.)
2612 *) Only add signing time to PKCS7 structures if it is not already
2616 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2617 OBJ_ld_ce should be OBJ_id_ce.
2618 Also some ip-pda OIDs in crypto/objects/objects.txt were
2619 incorrect (cf. RFC 3039).
2620 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2622 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2623 returns early because it has nothing to do.
2624 [Andy Schneider <andy.schneider@bjss.co.uk>]
2626 *) [In 0.9.6c-engine release:]
2627 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2628 [Andy Schneider <andy.schneider@bjss.co.uk>]
2630 *) [In 0.9.6c-engine release:]
2631 Add support for Cryptographic Appliance's keyserver technology.
2632 (Use engine 'keyclient')
2633 [Cryptographic Appliances and Geoff Thorpe]
2635 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2636 is called via tools/c89.sh because arguments have to be
2637 rearranged (all '-L' options must appear before the first object
2639 [Richard Shapiro <rshapiro@abinitio.com>]
2641 *) [In 0.9.6c-engine release:]
2642 Add support for Broadcom crypto accelerator cards, backported
2644 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2646 *) [In 0.9.6c-engine release:]
2647 Add support for SureWare crypto accelerator cards from
2648 Baltimore Technologies. (Use engine 'sureware')
2649 [Baltimore Technologies and Mark Cox]
2651 *) [In 0.9.6c-engine release:]
2652 Add support for crypto accelerator cards from Accelerated
2653 Encryption Processing, www.aep.ie. (Use engine 'aep')
2654 [AEP Inc. and Mark Cox]
2656 *) Add a configuration entry for gcc on UnixWare.
2657 [Gary Benson <gbenson@redhat.com>]
2659 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2660 messages are stored in a single piece (fixed-length part and
2661 variable-length part combined) and fix various bugs found on the way.
2664 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2665 instead. BIO_gethostbyname() does not know what timeouts are
2666 appropriate, so entries would stay in cache even when they have
2668 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2670 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2671 faced with a pathologically small ClientHello fragment that does
2672 not contain client_version: Instead of aborting with an error,
2673 simply choose the highest available protocol version (i.e.,
2674 TLS 1.0 unless it is disabled). In practice, ClientHello
2675 messages are never sent like this, but this change gives us
2676 strictly correct behaviour at least for TLS.
2679 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2680 never resets s->method to s->ctx->method when called from within
2681 one of the SSL handshake functions.
2682 [Bodo Moeller; problem pointed out by Niko Baric]
2684 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2685 (sent using the client's version number) if client_version is
2686 smaller than the protocol version in use. Also change
2687 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2688 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2689 the client will at least see that alert.
2692 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2696 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2697 client receives HelloRequest while in a handshake.
2698 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2700 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2701 should end in 'break', not 'goto end' which circuments various
2702 cleanups done in state SSL_ST_OK. But session related stuff
2703 must be disabled for SSL_ST_OK in the case that we just sent a
2706 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2707 before just sending a HelloRequest.
2708 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2710 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2711 reveal whether illegal block cipher padding was found or a MAC
2712 verification error occured. (Neither SSLerr() codes nor alerts
2713 are directly visible to potential attackers, but the information
2714 may leak via logfiles.)
2716 Similar changes are not required for the SSL 2.0 implementation
2717 because the number of padding bytes is sent in clear for SSL 2.0,
2718 and the extra bytes are just ignored. However ssl/s2_pkt.c
2719 failed to verify that the purported number of padding bytes is in
2723 *) Add OpenUNIX-8 support including shared libraries
2724 (Boyd Lynn Gerber <gerberb@zenez.com>).
2727 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2728 'wristwatch attack' using huge encoding parameters (cf.
2729 James H. Manger's CRYPTO 2001 paper). Note that the
2730 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2731 encoding parameters and hence was not vulnerable.
2734 *) BN_sqr() bug fix.
2735 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2737 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2738 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2739 followed by modular reduction.
2740 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2742 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2743 equivalent based on BN_pseudo_rand() instead of BN_rand().
2746 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2747 This function was broken, as the check for a new client hello message
2748 to handle SGC did not allow these large messages.
2749 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2752 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2755 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2756 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2759 *) Rework the configuration and shared library support for Tru64 Unix.
2760 The configuration part makes use of modern compiler features and
2761 still retains old compiler behavior for those that run older versions
2762 of the OS. The shared library support part includes a variant that
2763 uses the RPATH feature, and is available through the special
2764 configuration target "alpha-cc-rpath", which will never be selected
2766 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2768 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2769 with the same message size as in ssl3_get_certificate_request().
2770 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2771 messages might inadvertently be reject as too long.
2772 [Petr Lampa <lampa@fee.vutbr.cz>]
2774 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2777 *) Modified SSL library such that the verify_callback that has been set
2778 specificly for an SSL object with SSL_set_verify() is actually being
2779 used. Before the change, a verify_callback set with this function was
2780 ignored and the verify_callback() set in the SSL_CTX at the time of
2781 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2782 to allow the necessary settings.
2785 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2786 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2787 done automatically (in contradiction to the requirements of the C
2788 standard). This made problems when used from OpenSSH.
2791 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2792 dh->length and always used
2794 BN_rand_range(priv_key, dh->p).
2796 BN_rand_range() is not necessary for Diffie-Hellman, and this
2797 specific range makes Diffie-Hellman unnecessarily inefficient if
2798 dh->length (recommended exponent length) is much smaller than the
2799 length of dh->p. We could use BN_rand_range() if the order of
2800 the subgroup was stored in the DH structure, but we only have
2805 BN_rand(priv_key, l, ...)
2807 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2813 RSA_eay_public_encrypt
2814 RSA_eay_private_decrypt
2815 RSA_eay_private_encrypt (signing)
2816 RSA_eay_public_decrypt (signature verification)
2818 (default implementations for RSA_public_encrypt,
2819 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2820 always reject numbers >= n.
2823 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2824 to synchronize access to 'locking_thread'. This is necessary on
2825 systems where access to 'locking_thread' (an 'unsigned long'
2826 variable) is not atomic.
2829 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2830 *before* setting the 'crypto_lock_rand' flag. The previous code had
2831 a race condition if 0 is a valid thread ID.
2832 [Travis Vitek <vitek@roguewave.com>]
2834 *) Add support for shared libraries under Irix.
2835 [Albert Chin-A-Young <china@thewrittenword.com>]
2837 *) Add configuration option to build on Linux on both big-endian and
2839 [Ralf Baechle <ralf@uni-koblenz.de>]
2841 *) Add the possibility to create shared libraries on HP-UX.
2844 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2846 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2847 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2848 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2849 PRNG state recovery was possible based on the output of
2850 one PRNG request appropriately sized to gain knowledge on
2851 'md' followed by enough consecutive 1-byte PRNG requests
2852 to traverse all of 'state'.
2854 1. When updating 'md_local' (the current thread's copy of 'md')
2855 during PRNG output generation, hash all of the previous
2856 'md_local' value, not just the half used for PRNG output.
2858 2. Make the number of bytes from 'state' included into the hash
2859 independent from the number of PRNG bytes requested.
2861 The first measure alone would be sufficient to avoid
2862 Markku-Juhani's attack. (Actually it had never occurred
2863 to me that the half of 'md_local' used for chaining was the
2864 half from which PRNG output bytes were taken -- I had always
2865 assumed that the secret half would be used.) The second
2866 measure makes sure that additional data from 'state' is never
2867 mixed into 'md_local' in small portions; this heuristically
2868 further strengthens the PRNG.
2871 *) Fix crypto/bn/asm/mips3.s.
2874 *) When only the key is given to "enc", the IV is undefined. Print out
2875 an error message in this case.
2878 *) Handle special case when X509_NAME is empty in X509 printing routines.
2881 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2882 positive and less than q.
2885 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2886 used: it isn't thread safe and the add_lock_callback should handle
2888 [Paul Rose <Paul.Rose@bridge.com>]
2890 *) Verify that incoming data obeys the block size in
2891 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2895 [Ulf Möller, Bodo Möller]
2897 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2898 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2899 when fixing the server behaviour for backwards-compatible 'client
2900 hello' messages. (Note that the attack is impractical against
2901 SSL 3.0 and TLS 1.0 anyway because length and version checking
2902 means that the probability of guessing a valid ciphertext is
2903 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2906 Before 0.9.5, the countermeasure (hide the error by generating a
2907 random 'decryption result') did not work properly because
2908 ERR_clear_error() was missing, meaning that SSL_get_error() would
2909 detect the supposedly ignored error.
2911 Both problems are now fixed.
2914 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2915 (previously it was 1024).
2918 *) Fix for compatibility mode trust settings: ignore trust settings
2919 unless some valid trust or reject settings are present.
2922 *) Fix for blowfish EVP: its a variable length cipher.
2925 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2926 parameters in DSA public key structures and return an error in the
2927 DSA routines if parameters are absent.
2930 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2931 in the current directory if neither $RANDFILE nor $HOME was set.
2932 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2933 caused some confusion to Windows users who haven't defined $HOME.
2934 Thus RAND_file_name() is changed again: e_os.h can define a
2935 DEFAULT_HOME, which will be used if $HOME is not set.
2936 For Windows, we use "C:"; on other platforms, we still require
2937 environment variables.
2939 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2940 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2941 having multiple threads call RAND_poll() concurrently.
2944 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2945 combination of a flag and a thread ID variable.
2946 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2947 flag), *other* threads can enter ssleay_add_bytes without obeying
2948 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2949 that they do not hold after the first thread unsets add_do_not_lock).
2952 *) Change bctest again: '-x' expressions are not available in all
2956 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2958 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2959 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2961 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2962 the default extension for executables, if any. Also, make the perl
2963 scripts that use symlink() to test if it really exists and use "cp"
2964 if it doesn't. All this made OpenSSL compilable and installable in
2968 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2969 If SEQUENCE is length is indefinite just set c->slen to the total
2970 amount of data available.
2971 [Steve Henson, reported by shige@FreeBSD.org]
2972 [This change does not apply to 0.9.7.]
2974 *) Change bctest to avoid here-documents inside command substitution
2975 (workaround for FreeBSD /bin/sh bug).
2976 For compatibility with Ultrix, avoid shell functions (introduced
2977 in the bctest version that searches along $PATH).
2980 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2981 with des_encrypt() defined on some operating systems, like Solaris
2985 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2986 On the Importance of Eliminating Errors in Cryptographic
2987 Computations, J. Cryptology 14 (2001) 2, 101-119,
2988 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2991 *) MIPS assembler BIGNUM division bug fix.
2994 *) Disabled incorrect Alpha assembler code.
2997 *) Fix PKCS#7 decode routines so they correctly update the length
2998 after reading an EOC for the EXPLICIT tag.
3000 [This change does not apply to 0.9.7.]
3002 *) Fix bug in PKCS#12 key generation routines. This was triggered
3003 if a 3DES key was generated with a 0 initial byte. Include
3004 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3005 (but broken) behaviour.
3008 *) Enhance bctest to search for a working bc along $PATH and print
3010 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3012 *) Fix memory leaks in err.c: free err_data string if necessary;
3013 don't write to the wrong index in ERR_set_error_data.
3016 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3020 *) Replace rdtsc with _emit statements for VC++ version 5.
3021 [Jeremy Cooper <jeremy@baymoo.org>]
3023 *) Make it possible to reuse SSLv2 sessions.
3026 *) In copy_email() check for >= 0 as a return value for
3027 X509_NAME_get_index_by_NID() since 0 is a valid index.
3028 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3030 *) Avoid coredump with unsupported or invalid public keys by checking if
3031 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3032 PKCS7_verify() fails with non detached data.
3035 *) Don't use getenv in library functions when run as setuid/setgid.
3036 New function OPENSSL_issetugid().
3039 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3040 due to incorrect handling of multi-threading:
3042 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3044 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3046 3. Count how many times MemCheck_off() has been called so that
3047 nested use can be treated correctly. This also avoids
3048 inband-signalling in the previous code (which relied on the
3049 assumption that thread ID 0 is impossible).
3052 *) Add "-rand" option also to s_client and s_server.
3055 *) Fix CPU detection on Irix 6.x.
3056 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3057 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3059 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3062 [This change does not apply to 0.9.7.]
3064 *) Use the cached encoding of an X509_NAME structure rather than
3065 copying it. This is apparently the reason for the libsafe "errors"
3066 but the code is actually correct.
3069 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3070 Bleichenbacher's DSA attack.
3071 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3072 to be set and top=0 forces the highest bit to be set; top=-1 is new
3073 and leaves the highest bit random.
3074 [Ulf Moeller, Bodo Moeller]
3076 *) In the NCONF_...-based implementations for CONF_... queries
3077 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3078 a temporary CONF structure with the data component set to NULL
3079 (which gives segmentation faults in lh_retrieve).
3080 Instead, use NULL for the CONF pointer in CONF_get_string and
3081 CONF_get_number (which may use environment variables) and directly
3082 return NULL from CONF_get_section.
3085 *) Fix potential buffer overrun for EBCDIC.
3088 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3089 keyUsage if basicConstraints absent for a CA.
3092 *) Make SMIME_write_PKCS7() write mail header values with a format that
3093 is more generally accepted (no spaces before the semicolon), since
3094 some programs can't parse those values properly otherwise. Also make
3095 sure BIO's that break lines after each write do not create invalid
3099 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3100 macros previously used would not encode an empty SEQUENCE OF
3101 and break the signature.
3103 [This change does not apply to 0.9.7.]
3105 *) Zero the premaster secret after deriving the master secret in
3109 *) Add some EVP_add_digest_alias registrations (as found in
3110 OpenSSL_add_all_digests()) to SSL_library_init()
3111 aka OpenSSL_add_ssl_algorithms(). This provides improved
3112 compatibility with peers using X.509 certificates
3113 with unconventional AlgorithmIdentifier OIDs.
3116 *) Fix for Irix with NO_ASM.
3117 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3119 *) ./config script fixes.
3120 [Ulf Moeller, Richard Levitte]
3122 *) Fix 'openssl passwd -1'.
3125 *) Change PKCS12_key_gen_asc() so it can cope with non null
3126 terminated strings whose length is passed in the passlen
3127 parameter, for example from PEM callbacks. This was done
3128 by adding an extra length parameter to asc2uni().
3129 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3131 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3132 call failed, free the DSA structure.
3135 *) Fix to uni2asc() to cope with zero length Unicode strings.
3136 These are present in some PKCS#12 files.
3139 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3140 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3141 when writing a 32767 byte record.
3142 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3144 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3145 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3147 (RSA objects have a reference count access to which is protected
3148 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3149 so they are meant to be shared between threads.)
3150 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3151 "Reddie, Steven" <Steven.Reddie@ca.com>]
3153 *) Fix a deadlock in CRYPTO_mem_leaks().
3156 *) Use better test patterns in bntest.
3159 *) rand_win.c fix for Borland C.
3162 *) BN_rshift bugfix for n == 0.
3165 *) Add a 'bctest' script that checks for some known 'bc' bugs
3166 so that 'make test' does not abort just because 'bc' is broken.
3169 *) Store verify_result within SSL_SESSION also for client side to
3170 avoid potential security hole. (Re-used sessions on the client side
3171 always resulted in verify_result==X509_V_OK, not using the original
3172 result of the server certificate verification.)
3175 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3176 SSL3_RT_APPLICATION_DATA, return 0.
3177 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3181 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3182 releases, have been re-implemented by renaming the previous
3183 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3184 and ssl3_read_internal, respectively, and adding 'peek' parameters
3185 to them. The new ssl[23]_{read,peek} functions are calls to
3186 ssl[23]_read_internal with the 'peek' flag set appropriately.
3187 A 'peek' parameter has also been added to ssl3_read_bytes, which
3188 does the actual work for ssl3_read_internal.
3191 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3192 the method-specific "init()" handler. Also clean up ex_data after
3193 calling the method-specific "finish()" handler. Previously, this was
3194 happening the other way round.
3197 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3198 The previous value, 12, was not always sufficient for BN_mod_exp().
3201 *) Make sure that shared libraries get the internal name engine with
3202 the full version number and not just 0. This should mark the
3203 shared libraries as not backward compatible. Of course, this should
3204 be changed again when we can guarantee backward binary compatibility.
3207 *) Fix typo in get_cert_by_subject() in by_dir.c
3208 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3210 *) Rework the system to generate shared libraries:
3212 - Make note of the expected extension for the shared libraries and
3213 if there is a need for symbolic links from for example libcrypto.so.0
3214 to libcrypto.so.0.9.7. There is extended info in Configure for
3217 - Make as few rebuilds of the shared libraries as possible.
3219 - Still avoid linking the OpenSSL programs with the shared libraries.
3221 - When installing, install the shared libraries separately from the
3225 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3227 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3228 and not in SSL_clear because the latter is also used by the
3229 accept/connect functions; previously, the settings made by
3230 SSL_set_read_ahead would be lost during the handshake.
3231 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3233 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3234 Previously, it would create entries for disableed algorithms no
3238 *) Added several new manual pages for SSL_* function.
3241 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3243 *) In ssl23_get_client_hello, generate an error message when faced
3244 with an initial SSL 3.0/TLS record that is too small to contain the
3245 first two bytes of the ClientHello message, i.e. client_version.
3246 (Note that this is a pathologic case that probably has never happened
3247 in real life.) The previous approach was to use the version number
3248 from the record header as a substitute; but our protocol choice
3249 should not depend on that one because it is not authenticated
3250 by the Finished messages.
3253 *) More robust randomness gathering functions for Windows.
3254 [Jeffrey Altman <jaltman@columbia.edu>]
3256 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3257 not set then we don't setup the error code for issuer check errors
3258 to avoid possibly overwriting other errors which the callback does
3259 handle. If an application does set the flag then we assume it knows
3260 what it is doing and can handle the new informational codes
3264 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3265 a general "ANY" type, as such it should be able to decode anything
3266 including tagged types. However it didn't check the class so it would
3267 wrongly interpret tagged types in the same way as their universal
3268 counterpart and unknown types were just rejected. Changed so that the
3269 tagged and unknown types are handled in the same way as a SEQUENCE:
3270 that is the encoding is stored intact. There is also a new type
3271 "V_ASN1_OTHER" which is used when the class is not universal, in this
3272 case we have no idea what the actual type is so we just lump them all
3276 *) On VMS, stdout may very well lead to a file that is written to
3277 in a record-oriented fashion. That means that every write() will
3278 write a separate record, which will be read separately by the
3279 programs trying to read from it. This can be very confusing.
3281 The solution is to put a BIO filter in the way that will buffer
3282 text until a linefeed is reached, and then write everything a
3283 line at a time, so every record written will be an actual line,
3284 not chunks of lines and not (usually doesn't happen, but I've
3285 seen it once) several lines in one record. BIO_f_linebuffer() is
3288 Currently, it's a VMS-only method, because that's where it has
3289 been tested well enough.
3292 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3293 it can return incorrect results.
3294 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3295 but it was in 0.9.6-beta[12].)
3298 *) Disable the check for content being present when verifying detached
3299 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3300 include zero length content when signing messages.
3303 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3304 BIO_ctrl (for BIO pairs).
3307 *) Add DSO method for VMS.
3310 *) Bug fix: Montgomery multiplication could produce results with the
3314 *) Add RPM specification openssl.spec and modify it to build three
3315 packages. The default package contains applications, application
3316 documentation and run-time libraries. The devel package contains
3317 include files, static libraries and function documentation. The
3318 doc package contains the contents of the doc directory. The original
3319 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3322 *) Add a large number of documentation files for many SSL routines.
3323 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3325 *) Add a configuration entry for Sony News 4.
3326 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3328 *) Don't set the two most significant bits to one when generating a
3329 random number < q in the DSA library.
3332 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3333 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3334 the underlying transport is blocking) if a handshake took place.
3335 (The default behaviour is needed by applications such as s_client
3336 and s_server that use select() to determine when to use SSL_read;
3337 but for applications that know in advance when to expect data, it
3338 just makes things more complicated.)
3341 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3345 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3346 work better on such systems.
3347 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3349 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3350 Update PKCS12_parse() so it copies the friendlyName and the
3351 keyid to the certificates aux info.
3354 *) Fix bug in PKCS7_verify() which caused an infinite loop
3355 if there was more than one signature.
3356 [Sven Uszpelkat <su@celocom.de>]
3358 *) Major change in util/mkdef.pl to include extra information
3359 about each symbol, as well as presentig variables as well
3360 as functions. This change means that there's n more need
3361 to rebuild the .num files when some algorithms are excluded.
3364 *) Allow the verify time to be set by an application,
3365 rather than always using the current time.
3368 *) Phase 2 verify code reorganisation. The certificate
3369 verify code now looks up an issuer certificate by a
3370 number of criteria: subject name, authority key id
3371 and key usage. It also verifies self signed certificates
3372 by the same criteria. The main comparison function is
3373 X509_check_issued() which performs these checks.
3375 Lot of changes were necessary in order to support this
3376 without completely rewriting the lookup code.
3378 Authority and subject key identifier are now cached.
3380 The LHASH 'certs' is X509_STORE has now been replaced
3381 by a STACK_OF(X509_OBJECT). This is mainly because an
3382 LHASH can't store or retrieve multiple objects with
3383 the same hash value.
3385 As a result various functions (which were all internal
3386 use only) have changed to handle the new X509_STORE
3387 structure. This will break anything that messed round
3388 with X509_STORE internally.
3390 The functions X509_STORE_add_cert() now checks for an
3391 exact match, rather than just subject name.
3393 The X509_STORE API doesn't directly support the retrieval
3394 of multiple certificates matching a given criteria, however
3395 this can be worked round by performing a lookup first
3396 (which will fill the cache with candidate certificates)
3397 and then examining the cache for matches. This is probably
3398 the best we can do without throwing out X509_LOOKUP
3399 entirely (maybe later...).
3401 The X509_VERIFY_CTX structure has been enhanced considerably.
3403 All certificate lookup operations now go via a get_issuer()
3404 callback. Although this currently uses an X509_STORE it
3405 can be replaced by custom lookups. This is a simple way
3406 to bypass the X509_STORE hackery necessary to make this
3407 work and makes it possible to use more efficient techniques
3408 in future. A very simple version which uses a simple
3409 STACK for its trusted certificate store is also provided
3410 using X509_STORE_CTX_trusted_stack().
3412 The verify_cb() and verify() callbacks now have equivalents
3413 in the X509_STORE_CTX structure.
3415 X509_STORE_CTX also has a 'flags' field which can be used
3416 to customise the verify behaviour.
3419 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3420 excludes S/MIME capabilities.
3423 *) When a certificate request is read in keep a copy of the
3424 original encoding of the signed data and use it when outputing
3425 again. Signatures then use the original encoding rather than
3426 a decoded, encoded version which may cause problems if the
3427 request is improperly encoded.
3430 *) For consistency with other BIO_puts implementations, call
3431 buffer_write(b, ...) directly in buffer_puts instead of calling
3434 In BIO_puts, increment b->num_write as in BIO_write.
3435 [Peter.Sylvester@EdelWeb.fr]
3437 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3438 BN_zero, we may not return a BIGNUM with an array consisting of
3442 *) Avoid calling abort() from within the library when problems are
3443 detected, except if preprocessor symbols have been defined
3444 (such as REF_CHECK, BN_DEBUG etc.).
3447 *) New openssl application 'rsautl'. This utility can be
3448 used for low level RSA operations. DER public key
3449 BIO/fp routines also added.
3452 *) New Configure entry and patches for compiling on QNX 4.
3453 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3455 *) A demo state-machine implementation was sponsored by
3456 Nuron (http://www.nuron.com/) and is now available in
3457 demos/state_machine.
3460 *) New options added to the 'dgst' utility for signature
3461 generation and verification.
3464 *) Unrecognized PKCS#7 content types are now handled via a
3465 catch all ASN1_TYPE structure. This allows unsupported
3466 types to be stored as a "blob" and an application can
3467 encode and decode it manually.
3470 *) Fix various signed/unsigned issues to make a_strex.c
3472 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3474 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3475 length if passed a buffer. ASN1_INTEGER_to_BN failed
3476 if passed a NULL BN and its argument was negative.
3477 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3479 *) Modification to PKCS#7 encoding routines to output definite
3480 length encoding. Since currently the whole structures are in
3481 memory there's not real point in using indefinite length
3482 constructed encoding. However if OpenSSL is compiled with
3483 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3486 *) Added BIO_vprintf() and BIO_vsnprintf().
3489 *) Added more prefixes to parse for in the the strings written
3490 through a logging bio, to cover all the levels that are available
3491 through syslog. The prefixes are now:
3493 PANIC, EMERG, EMR => LOG_EMERG
3494 ALERT, ALR => LOG_ALERT
3495 CRIT, CRI => LOG_CRIT
3496 ERROR, ERR => LOG_ERR
3497 WARNING, WARN, WAR => LOG_WARNING
3498 NOTICE, NOTE, NOT => LOG_NOTICE
3499 INFO, INF => LOG_INFO
3500 DEBUG, DBG => LOG_DEBUG
3502 and as before, if none of those prefixes are present at the
3503 beginning of the string, LOG_ERR is chosen.
3505 On Win32, the LOG_* levels are mapped according to this:
3507 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3508 LOG_WARNING => EVENTLOG_WARNING_TYPE
3509 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3513 *) Made it possible to reconfigure with just the configuration
3514 argument "reconf" or "reconfigure". The command line arguments
3515 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3516 and are retrieved from there when reconfiguring.
3520 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3522 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3525 *) The obj_dat.pl script was messing up the sorting of object
3526 names. The reason was that it compared the quoted version
3527 of strings as a result "OCSP" > "OCSP Signing" because
3528 " > SPACE. Changed script to store unquoted versions of
3529 names and add quotes on output. It was also omitting some
3530 names from the lookup table if they were given a default
3531 value (that is if SN is missing it is given the same
3532 value as LN and vice versa), these are now added on the
3533 grounds that if an object has a name we should be able to
3534 look it up. Finally added warning output when duplicate
3535 short or long names are found.
3538 *) Changes needed for Tandem NSK.
3539 [Scott Uroff <scott@xypro.com>]
3541 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3542 RSA_padding_check_SSLv23(), special padding was never detected
3543 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3544 version rollback attacks was not effective.
3546 In s23_clnt.c, don't use special rollback-attack detection padding
3547 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3548 client; similarly, in s23_srvr.c, don't do the rollback check if
3549 SSL 2.0 is the only protocol enabled in the server.
3552 *) Make it possible to get hexdumps of unprintable data with 'openssl
3553 asn1parse'. By implication, the functions ASN1_parse_dump() and
3554 BIO_dump_indent() are added.
3557 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3558 these print out strings and name structures based on various
3559 flags including RFC2253 support and proper handling of
3560 multibyte characters. Added options to the 'x509' utility
3561 to allow the various flags to be set.
3564 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3565 Also change the functions X509_cmp_current_time() and
3566 X509_gmtime_adj() work with an ASN1_TIME structure,
3567 this will enable certificates using GeneralizedTime in validity
3568 dates to be checked.
3571 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3572 negative public key encodings) on by default,
3573 NO_NEG_PUBKEY_BUG can be set to disable it.
3576 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3577 content octets. An i2c_ASN1_OBJECT is unnecessary because
3578 the encoding can be trivially obtained from the structure.
3581 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3582 not read locks (CRYPTO_r_[un]lock).
3585 *) A first attempt at creating official support for shared
3586 libraries through configuration. I've kept it so the
3587 default is static libraries only, and the OpenSSL programs
3588 are always statically linked for now, but there are
3589 preparations for dynamic linking in place.
3590 This has been tested on Linux and Tru64.
3593 *) Randomness polling function for Win9x, as described in:
3594 Peter Gutmann, Software Generation of Practically Strong
3598 *) Fix so PRNG is seeded in req if using an already existing
3602 *) New options to smime application. -inform and -outform
3603 allow alternative formats for the S/MIME message including
3604 PEM and DER. The -content option allows the content to be
3605 specified separately. This should allow things like Netscape
3606 form signing output easier to verify.
3609 *) Fix the ASN1 encoding of tags using the 'long form'.
3612 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3613 STRING types. These convert content octets to and from the
3614 underlying type. The actual tag and length octets are
3615 already assumed to have been read in and checked. These
3616 are needed because all other string types have virtually
3617 identical handling apart from the tag. By having versions
3618 of the ASN1 functions that just operate on content octets
3619 IMPLICIT tagging can be handled properly. It also allows
3620 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3621 and ASN1_INTEGER are identical apart from the tag.
3624 *) Change the handling of OID objects as follows:
3626 - New object identifiers are inserted in objects.txt, following
3627 the syntax given in objects.README.
3628 - objects.pl is used to process obj_mac.num and create a new
3630 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3633 This is currently kind of a hack, and the perl code in objects.pl
3634 isn't very elegant, but it works as I intended. The simplest way
3635 to check that it worked correctly is to look in obj_dat.h and
3636 check the array nid_objs and make sure the objects haven't moved
3637 around (this is important!). Additions are OK, as well as
3638 consistent name changes.
3641 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3644 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3645 The given file adds to whatever has already been seeded into the
3646 random pool through the RANDFILE configuration file option or
3647 environment variable, or the default random state file.
3650 *) mkstack.pl now sorts each macro group into lexical order.
3651 Previously the output order depended on the order the files
3652 appeared in the directory, resulting in needless rewriting
3656 *) Patches to make OpenSSL compile under Win32 again. Mostly
3657 work arounds for the VC++ problem that it treats func() as
3658 func(void). Also stripped out the parts of mkdef.pl that
3659 added extra typesafe functions: these no longer exist.
3662 *) Reorganisation of the stack code. The macros are now all
3663 collected in safestack.h . Each macro is defined in terms of
3664 a "stack macro" of the form SKM_<name>(type, a, b). The
3665 DEBUG_SAFESTACK is now handled in terms of function casts,
3666 this has the advantage of retaining type safety without the
3667 use of additional functions. If DEBUG_SAFESTACK is not defined
3668 then the non typesafe macros are used instead. Also modified the
3669 mkstack.pl script to handle the new form. Needs testing to see
3670 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3671 the default if no major problems. Similar behaviour for ASN1_SET_OF
3672 and PKCS12_STACK_OF.
3675 *) When some versions of IIS use the 'NET' form of private key the
3676 key derivation algorithm is different. Normally MD5(password) is
3677 used as a 128 bit RC4 key. In the modified case
3678 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3679 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3680 as the old Netscape_RSA functions except they have an additional
3681 'sgckey' parameter which uses the modified algorithm. Also added
3682 an -sgckey command line option to the rsa utility. Thanks to
3683 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3684 algorithm to openssl-dev.
3687 *) The evp_local.h macros were using 'c.##kname' which resulted in
3688 invalid expansion on some systems (SCO 5.0.5 for example).
3689 Corrected to 'c.kname'.
3690 [Phillip Porch <root@theporch.com>]
3692 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3693 a STACK of email addresses from a certificate or request, these look
3694 in the subject name and the subject alternative name extensions and
3695 omit any duplicate addresses.
3698 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3699 This makes DSA verification about 2 % faster.
3702 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3703 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3704 plus overhead for 1024 bit moduli).
3705 This makes exponentiations about 0.5 % faster for 1024 bit
3706 exponents (as measured by "openssl speed rsa2048").
3709 *) Rename memory handling macros to avoid conflicts with other
3711 Malloc => OPENSSL_malloc
3712 Malloc_locked => OPENSSL_malloc_locked
3713 Realloc => OPENSSL_realloc
3714 Free => OPENSSL_free
3717 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3718 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3721 *) CygWin32 support.
3722 [John Jarvie <jjarvie@newsguy.com>]
3724 *) The type-safe stack code has been rejigged. It is now only compiled
3725 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3726 by default all type-specific stack functions are "#define"d back to
3727 standard stack functions. This results in more streamlined output
3728 but retains the type-safety checking possibilities of the original
3732 *) The STACK code has been cleaned up, and certain type declarations
3733 that didn't make a lot of sense have been brought in line. This has
3734 also involved a cleanup of sorts in safestack.h to more correctly
3735 map type-safe stack functions onto their plain stack counterparts.
3736 This work has also resulted in a variety of "const"ifications of
3737 lots of the code, especially "_cmp" operations which should normally
3738 be prototyped with "const" parameters anyway.
3741 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3742 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3743 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3744 where all of 'md' is used each time the PRNG is used, but 'state'
3745 is used only indexed by a cyclic counter. As entropy may not be
3746 well distributed from the beginning, 'md' is important as a
3747 chaining variable. However, the output function chains only half
3748 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3749 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3750 in all of 'state' being rewritten, with the new values depending
3751 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3754 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3755 the handshake is continued after ssl_verify_cert_chain();
3756 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3757 can lead to 'unexplainable' connection aborts later.
3758 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3760 *) Major EVP API cipher revision.
3761 Add hooks for extra EVP features. This allows various cipher
3762 parameters to be set in the EVP interface. Support added for variable
3763 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3764 setting of RC2 and RC5 parameters.
3766 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3769 Remove lots of duplicated code from the EVP library. For example *every*
3770 cipher init() function handles the 'iv' in the same way according to the
3771 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3772 for CFB and OFB modes they zero ctx->num.
3774 New functionality allows removal of S/MIME code RC2 hack.
3776 Most of the routines have the same form and so can be declared in terms
3779 By shifting this to the top level EVP_CipherInit() it can be removed from
3780 all individual ciphers. If the cipher wants to handle IVs or keys
3781 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3784 Change lots of functions like EVP_EncryptUpdate() to now return a
3785 value: although software versions of the algorithms cannot fail
3786 any installed hardware versions can.
3789 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3790 this option is set, tolerate broken clients that send the negotiated
3791 protocol version number instead of the requested protocol version
3795 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3796 i.e. non-zero for export ciphersuites, zero otherwise.
3797 Previous versions had this flag inverted, inconsistent with
3798 rsa_tmp_cb (..._TMP_RSA_CB).
3799 [Bodo Moeller; problem reported by Amit Chopra]
3801 *) Add missing DSA library text string. Work around for some IIS
3802 key files with invalid SEQUENCE encoding.
3805 *) Add a document (doc/standards.txt) that list all kinds of standards
3806 and so on that are implemented in OpenSSL.
3809 *) Enhance c_rehash script. Old version would mishandle certificates
3810 with the same subject name hash and wouldn't handle CRLs at all.
3811 Added -fingerprint option to crl utility, to support new c_rehash
3815 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3818 *) Fix for SSL server purpose checking. Server checking was
3819 rejecting certificates which had extended key usage present
3820 but no ssl client purpose.
3821 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3823 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3824 is a little unclear about how a blank password is handled.
3825 Since the password in encoded as a BMPString with terminating
3826 double NULL a zero length password would end up as just the
3827 double NULL. However no password at all is different and is
3828 handled differently in the PKCS#12 key generation code. NS
3829 treats a blank password as zero length. MSIE treats it as no
3830 password on export: but it will try both on import. We now do
3831 the same: PKCS12_parse() tries zero length and no password if
3832 the password is set to "" or NULL (NULL is now a valid password:
3833 it wasn't before) as does the pkcs12 application.
3836 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3837 perror when PEM_read_bio_X509_REQ fails, the error message must
3838 be obtained from the error queue.
3841 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3842 it in ERR_remove_state if appropriate, and change ERR_get_state
3843 accordingly to avoid race conditions (this is necessary because
3844 thread_hash is no longer constant once set).
3847 *) Bugfix for linux-elf makefile.one.
3850 *) RSA_get_default_method() will now cause a default
3851 RSA_METHOD to be chosen if one doesn't exist already.
3852 Previously this was only set during a call to RSA_new()
3853 or RSA_new_method(NULL) meaning it was possible for
3854 RSA_get_default_method() to return NULL.
3857 *) Added native name translation to the existing DSO code
3858 that will convert (if the flag to do so is set) filenames
3859 that are sufficiently small and have no path information
3860 into a canonical native form. Eg. "blah" converted to
3861 "libblah.so" or "blah.dll" etc.
3864 *) New function ERR_error_string_n(e, buf, len) which is like
3865 ERR_error_string(e, buf), but writes at most 'len' bytes
3866 including the 0 terminator. For ERR_error_string_n, 'buf'
3868 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3870 *) CONF library reworked to become more general. A new CONF
3871 configuration file reader "class" is implemented as well as a
3872 new functions (NCONF_*, for "New CONF") to handle it. The now
3873 old CONF_* functions are still there, but are reimplemented to
3874 work in terms of the new functions. Also, a set of functions
3875 to handle the internal storage of the configuration data is
3876 provided to make it easier to write new configuration file
3877 reader "classes" (I can definitely see something reading a
3878 configuration file in XML format, for example), called _CONF_*,
3879 or "the configuration storage API"...
3881 The new configuration file reading functions are:
3883 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3884 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3886 NCONF_default, NCONF_WIN32
3888 NCONF_dump_fp, NCONF_dump_bio
3890 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3891 NCONF_new creates a new CONF object. This works in the same way
3892 as other interfaces in OpenSSL, like the BIO interface.
3893 NCONF_dump_* dump the internal storage of the configuration file,
3894 which is useful for debugging. All other functions take the same
3895 arguments as the old CONF_* functions wth the exception of the
3896 first that must be a `CONF *' instead of a `LHASH *'.
3898 To make it easer to use the new classes with the old CONF_* functions,
3899 the function CONF_set_default_method is provided.
3902 *) Add '-tls1' option to 'openssl ciphers', which was already
3903 mentioned in the documentation but had not been implemented.
3904 (This option is not yet really useful because even the additional
3905 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3908 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3909 OpenSSL-based applications) load shared libraries and bind to
3910 them in a portable way.
3911 [Geoff Thorpe, with contributions from Richard Levitte]
3913 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3915 *) Make sure _lrotl and _lrotr are only used with MSVC.
3917 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3918 (the default implementation of RAND_status).
3920 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3921 to '-clrext' (= clear extensions), as intended and documented.
3922 [Bodo Moeller; inconsistency pointed out by Michael Attili
3925 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3926 was larger than the MD block size.
3927 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3929 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3930 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3931 using the passed key: if the passed key was a private key the result
3932 of X509_print(), for example, would be to print out all the private key
3936 *) des_quad_cksum() byte order bug fix.
3937 [Ulf Möller, using the problem description in krb4-0.9.7, where
3938 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3940 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3942 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3944 *) For easily testing in shell scripts whether some command
3945 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3946 returns with exit code 0 iff no command of the given name is available.
3947 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3948 the output goes to stdout and nothing is printed to stderr.
3949 Additional arguments are always ignored.
3951 Since for each cipher there is a command of the same name,
3952 the 'no-cipher' compilation switches can be tested this way.
3954 ('openssl no-XXX' is not able to detect pseudo-commands such
3955 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3958 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3961 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3962 is set; it will be thrown away anyway because each handshake creates
3964 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3965 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3966 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3967 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3970 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3971 'Q' and 'R' lose their special meanings (quit/renegotiate).
3972 This is part of what -quiet does; unlike -quiet, -ign_eof
3973 does not suppress any output.
3976 *) Add compatibility options to the purpose and trust code. The
3977 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3978 accepts a certificate or CA, this was the previous behaviour,
3979 with all the associated security issues.
3981 X509_TRUST_COMPAT is the old trust behaviour: only and
3982 automatically trust self signed roots in certificate store. A
3983 new trust setting X509_TRUST_DEFAULT is used to specify that
3984 a purpose has no associated trust setting and it should instead
3985 use the value in the default purpose.
3988 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3989 and fix a memory leak.
3992 *) In util/mkerr.pl (which implements 'make errors'), preserve
3993 reason strings from the previous version of the .c file, as
3994 the default to have only downcase letters (and digits) in
3995 automatically generated reasons codes is not always appropriate.
3998 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3999 using strerror. Previously, ERR_reason_error_string() returned
4000 library names as reason strings for SYSerr; but SYSerr is a special
4001 case where small numbers are errno values, not library numbers.
4004 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4005 converts DSA parameters into DH parameters. (When creating parameters,
4006 DSA_generate_parameters is used.)
4009 *) Include 'length' (recommended exponent length) in C code generated
4010 by 'openssl dhparam -C'.
4013 *) The second argument to set_label in perlasm was already being used
4014 so couldn't be used as a "file scope" flag. Moved to third argument
4018 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4019 instead of RAND_bytes for encryption IVs and salts.
4022 *) Include RAND_status() into RAND_METHOD instead of implementing
4023 it only for md_rand.c Otherwise replacing the PRNG by calling
4024 RAND_set_rand_method would be impossible.
4027 *) Don't let DSA_generate_key() enter an infinite loop if the random
4028 number generation fails.
4031 *) New 'rand' application for creating pseudo-random output.
4034 *) Added configuration support for Linux/IA64
4035 [Rolf Haberrecker <rolf@suse.de>]
4037 *) Assembler module support for Mingw32.
4040 *) Shared library support for HPUX (in shlib/).
4041 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4043 *) Shared library support for Solaris gcc.
4044 [Lutz Behnke <behnke@trustcenter.de>]
4046 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4048 *) PKCS7_encrypt() was adding text MIME headers twice because they
4049 were added manually and by SMIME_crlf_copy().
4052 *) In bntest.c don't call BN_rand with zero bits argument.
4053 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4055 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4056 case was implemented. This caused BN_div_recp() to fail occasionally.
4059 *) Add an optional second argument to the set_label() in the perl
4060 assembly language builder. If this argument exists and is set
4061 to 1 it signals that the assembler should use a symbol whose
4062 scope is the entire file, not just the current function. This
4063 is needed with MASM which uses the format label:: for this scope.
4064 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4066 *) Change the ASN1 types so they are typedefs by default. Before
4067 almost all types were #define'd to ASN1_STRING which was causing
4068 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4072 *) Change names of new functions to the new get1/get0 naming
4073 convention: After 'get1', the caller owns a reference count
4074 and has to call ..._free; 'get0' returns a pointer to some
4075 data structure without incrementing reference counters.
4076 (Some of the existing 'get' functions increment a reference
4077 counter, some don't.)
4078 Similarly, 'set1' and 'add1' functions increase reference
4079 counters or duplicate objects.
4082 *) Allow for the possibility of temp RSA key generation failure:
4083 the code used to assume it always worked and crashed on failure.
4086 *) Fix potential buffer overrun problem in BIO_printf().
4087 [Ulf Möller, using public domain code by Patrick Powell; problem
4088 pointed out by David Sacerdote <das33@cornell.edu>]
4090 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4091 RAND_egd() and RAND_status(). In the command line application,
4092 the EGD socket can be specified like a seed file using RANDFILE
4096 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4097 Some CAs (e.g. Verisign) distribute certificates in this form.
4100 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4101 list to exclude them. This means that no special compilation option
4102 is needed to use anonymous DH: it just needs to be included in the
4106 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4107 EVP_MD_type. The old functionality is available in a new macro called
4108 EVP_MD_md(). Change code that uses it and update docs.
4111 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4112 where the 'void *' argument is replaced by a function pointer argument.
4113 Previously 'void *' was abused to point to functions, which works on
4114 many platforms, but is not correct. As these functions are usually
4115 called by macros defined in OpenSSL header files, most source code
4116 should work without changes.
4119 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4120 sections with information on -D... compiler switches used for
4121 compiling the library so that applications can see them. To enable
4122 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4123 must be defined. E.g.,
4124 #define OPENSSL_ALGORITHM_DEFINES
4125 #include <openssl/opensslconf.h>
4126 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4127 [Richard Levitte, Ulf and Bodo Möller]
4129 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4133 *) Change the 'other' type in certificate aux info to a STACK_OF
4134 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4135 the required ASN1 format: arbitrary types determined by an OID.
4138 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4139 argument to 'req'. This is not because the function is newer or
4140 better than others it just uses the work 'NEW' in the certificate
4141 request header lines. Some software needs this.
4144 *) Reorganise password command line arguments: now passwords can be
4145 obtained from various sources. Delete the PEM_cb function and make
4146 it the default behaviour: i.e. if the callback is NULL and the
4147 usrdata argument is not NULL interpret it as a null terminated pass
4148 phrase. If usrdata and the callback are NULL then the pass phrase
4149 is prompted for as usual.
4152 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4153 the support is automatically enabled. The resulting binaries will
4154 autodetect the card and use it if present.
4155 [Ben Laurie and Compaq Inc.]
4157 *) Work around for Netscape hang bug. This sends certificate request
4158 and server done in one record. Since this is perfectly legal in the
4159 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4160 the bugs/SSLv3 entry for more info.
4163 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4166 *) Add -rand argument to smime and pkcs12 applications and read/write
4170 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4173 *) Add command line password options to the remaining applications.
4176 *) Bug fix for BN_div_recp() for numerators with an even number of
4180 *) More tests in bntest.c, and changed test_bn output.
4183 *) ./config recognizes MacOS X now.
4186 *) Bug fix for BN_div() when the first words of num and divsor are
4187 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4190 *) Add support for various broken PKCS#8 formats, and command line
4191 options to produce them.
4194 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4195 get temporary BIGNUMs from a BN_CTX.
4198 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4202 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4203 include a #define from the old name to the new. The original intent
4204 was that statically linked binaries could for example just call
4205 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4206 link with digests. This never worked becayse SSLeay_add_all_digests()
4207 and SSLeay_add_all_ciphers() were in the same source file so calling
4208 one would link with the other. They are now in separate source files.
4211 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4214 *) Use a less unusual form of the Miller-Rabin primality test (it used
4215 a binary algorithm for exponentiation integrated into the Miller-Rabin
4216 loop, our standard modexp algorithms are faster).
4219 *) Support for the EBCDIC character set completed.
4220 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4222 *) Source code cleanups: use const where appropriate, eliminate casts,
4223 use void * instead of char * in lhash.
4226 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4227 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4228 this the server could overwrite ephemeral keys that the client
4232 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4233 using 50 iterations of the Rabin-Miller test.
4235 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4236 iterations of the Rabin-Miller test as required by the appendix
4237 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4238 As BN_is_prime_fasttest includes trial division, DSA parameter
4239 generation becomes much faster.
4241 This implies a change for the callback functions in DSA_is_prime
4242 and DSA_generate_parameters: The callback function is called once
4243 for each positive witness in the Rabin-Miller test, not just
4244 occasionally in the inner loop; and the parameters to the
4245 callback function now provide an iteration count for the outer
4246 loop rather than for the current invocation of the inner loop.
4247 DSA_generate_parameters additionally can call the callback
4248 function with an 'iteration count' of -1, meaning that a
4249 candidate has passed the trial division test (when q is generated
4250 from an application-provided seed, trial division is skipped).
4253 *) New function BN_is_prime_fasttest that optionally does trial
4254 division before starting the Rabin-Miller test and has
4255 an additional BN_CTX * argument (whereas BN_is_prime always
4256 has to allocate at least one BN_CTX).
4257 'callback(1, -1, cb_arg)' is called when a number has passed the
4258 trial division stage.
4261 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4265 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4268 *) New function BN_pseudo_rand().
4271 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4272 bignum version of BN_from_montgomery() with the working code from
4273 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4277 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4278 made it impossible to use the same SSL_SESSION data structure in
4279 SSL2 clients in multiple threads.
4282 *) The return value of RAND_load_file() no longer counts bytes obtained
4283 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4284 to seed the PRNG (previously an explicit byte count was required).
4285 [Ulf Möller, Bodo Möller]
4287 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4288 used (char *) instead of (void *) and had casts all over the place.
4291 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4294 *) Retain source code compatibility for BN_prime_checks macro:
4295 BN_is_prime(..., BN_prime_checks, ...) now uses
4296 BN_prime_checks_for_size to determine the appropriate number of
4297 Rabin-Miller iterations.
4300 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4301 DH_CHECK_P_NOT_SAFE_PRIME.
4302 (Check if this is true? OpenPGP calls them "strong".)
4305 *) Merge the functionality of "dh" and "gendh" programs into a new program
4306 "dhparam". The old programs are retained for now but will handle DH keys
4307 (instead of parameters) in future.
4310 *) Make the ciphers, s_server and s_client programs check the return values
4311 when a new cipher list is set.
4314 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4315 ciphers. Before when the 56bit ciphers were enabled the sorting was
4318 The syntax for the cipher sorting has been extended to support sorting by
4319 cipher-strength (using the strength_bits hard coded in the tables).
4320 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4322 Fix a bug in the cipher-command parser: when supplying a cipher command
4323 string with an "undefined" symbol (neither command nor alphanumeric
4324 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4325 an error is flagged.
4327 Due to the strength-sorting extension, the code of the
4328 ssl_create_cipher_list() function was completely rearranged. I hope that
4329 the readability was also increased :-)
4330 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4332 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4333 for the first serial number and places 2 in the serial number file. This
4334 avoids problems when the root CA is created with serial number zero and
4335 the first user certificate has the same issuer name and serial number
4339 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4340 the new code. Add documentation for this stuff.
4343 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4344 X509_*() to X509at_*() on the grounds that they don't handle X509
4345 structures and behave in an analagous way to the X509v3 functions:
4346 they shouldn't be called directly but wrapper functions should be used
4349 So we also now have some wrapper functions that call the X509at functions
4350 when passed certificate requests. (TO DO: similar things can be done with
4351 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4352 things. Some of these need some d2i or i2d and print functionality
4353 because they handle more complex structures.)
4356 *) Add missing #ifndefs that caused missing symbols when building libssl
4357 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4358 NO_RSA in ssl/s2*.c.
4359 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4361 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4362 has a return value which indicates the quality of the random data
4363 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4364 error queue. New function RAND_pseudo_bytes() generates output that is
4365 guaranteed to be unique but not unpredictable. RAND_add is like
4366 RAND_seed, but takes an extra argument for an entropy estimate
4367 (RAND_seed always assumes full entropy).
4370 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4371 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4372 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4373 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4374 false-positive rate of at most 2^-80 for random input.
4377 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4380 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4381 in the 0.9.5 release), this returns the chain
4382 from an X509_CTX structure with a dup of the stack and all
4383 the X509 reference counts upped: so the stack will exist
4384 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4387 Also make SSL_SESSION_print() print out the verify return
4391 *) Add manpage for the pkcs12 command. Also change the default
4392 behaviour so MAC iteration counts are used unless the new
4393 -nomaciter option is used. This improves file security and
4394 only older versions of MSIE (4.0 for example) need it.
4397 *) Honor the no-xxx Configure options when creating .DEF files.
4400 *) Add PKCS#10 attributes to field table: challengePassword,
4401 unstructuredName and unstructuredAddress. These are taken from
4402 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4403 international characters are used.
4405 More changes to X509_ATTRIBUTE code: allow the setting of types
4406 based on strings. Remove the 'loc' parameter when adding
4407 attributes because these will be a SET OF encoding which is sorted
4411 *) Initial changes to the 'req' utility to allow request generation
4412 automation. This will allow an application to just generate a template
4413 file containing all the field values and have req construct the
4416 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4417 used all over the place including certificate requests and PKCS#7
4418 structures. They are currently handled manually where necessary with
4419 some primitive wrappers for PKCS#7. The new functions behave in a
4420 manner analogous to the X509 extension functions: they allow
4421 attributes to be looked up by NID and added.
4423 Later something similar to the X509V3 code would be desirable to
4424 automatically handle the encoding, decoding and printing of the
4425 more complex types. The string types like challengePassword can
4426 be handled by the string table functions.
4428 Also modified the multi byte string table handling. Now there is
4429 a 'global mask' which masks out certain types. The table itself
4430 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4431 is useful when for example there is only one permissible type
4432 (as in countryName) and using the mask might result in no valid
4436 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4437 SSL_get_peer_finished to allow applications to obtain the latest
4438 Finished messages sent to the peer or expected from the peer,
4439 respectively. (SSL_get_peer_finished is usually the Finished message
4440 actually received from the peer, otherwise the protocol will be aborted.)
4442 As the Finished message are message digests of the complete handshake
4443 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4444 be used for external authentication procedures when the authentication
4445 provided by SSL/TLS is not desired or is not enough.
4448 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4449 the host supports BWX extension and if Compaq C is present on the
4450 $PATH. Just exploiting of the BWX extension results in 20-30%
4451 performance kick for some algorithms, e.g. DES and RC4 to mention
4452 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4456 *) Add support for MS "fast SGC". This is arguably a violation of the
4457 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4458 weak crypto and after checking the certificate is SGC a second one
4459 with strong crypto. MS SGC stops the first handshake after receiving
4460 the server certificate message and sends a second client hello. Since
4461 a server will typically do all the time consuming operations before
4462 expecting any further messages from the client (server key exchange
4463 is the most expensive) there is little difference between the two.
4465 To get OpenSSL to support MS SGC we have to permit a second client
4466 hello message after we have sent server done. In addition we have to
4467 reset the MAC if we do get this second client hello.
4470 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4471 if a DER encoded private key is RSA or DSA traditional format. Changed
4472 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4473 format DER encoded private key. Newer code should use PKCS#8 format which
4474 has the key type encoded in the ASN1 structure. Added DER private key
4475 support to pkcs8 application.
4478 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4479 ciphersuites has been selected (as required by the SSL 3/TLS 1
4480 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4481 is set, we interpret this as a request to violate the specification
4482 (the worst that can happen is a handshake failure, and 'correct'
4483 behaviour would result in a handshake failure anyway).
4486 *) In SSL_CTX_add_session, take into account that there might be multiple
4487 SSL_SESSION structures with the same session ID (e.g. when two threads
4488 concurrently obtain them from an external cache).
4489 The internal cache can handle only one SSL_SESSION with a given ID,
4490 so if there's a conflict, we now throw out the old one to achieve
4494 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4495 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4496 some routines that use cipher OIDs: some ciphers do not have OIDs
4497 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4501 *) Simplify the trust setting structure and code. Now we just have
4502 two sequences of OIDs for trusted and rejected settings. These will
4503 typically have values the same as the extended key usage extension
4504 and any application specific purposes.
4506 The trust checking code now has a default behaviour: it will just
4507 check for an object with the same NID as the passed id. Functions can
4508 be provided to override either the default behaviour or the behaviour
4509 for a given id. SSL client, server and email already have functions
4510 in place for compatibility: they check the NID and also return "trusted"
4511 if the certificate is self signed.
4514 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4515 traditional format into an EVP_PKEY structure.
4518 *) Add a password callback function PEM_cb() which either prompts for
4519 a password if usr_data is NULL or otherwise assumes it is a null
4520 terminated password. Allow passwords to be passed on command line
4521 environment or config files in a few more utilities.
4524 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4525 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4526 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4527 Update documentation.
4530 *) Support for ASN1 "NULL" type. This could be handled before by using
4531 ASN1_TYPE but there wasn't any function that would try to read a NULL
4532 and produce an error if it couldn't. For compatibility we also have
4533 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4534 don't allocate anything because they don't need to.
4537 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4539 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4541 *) Rebuild of the memory allocation routines used by OpenSSL code and
4542 possibly others as well. The purpose is to make an interface that
4543 provide hooks so anyone can build a separate set of allocation and
4544 deallocation routines to be used by OpenSSL, for example memory
4545 pool implementations, or something else, which was previously hard
4546 since Malloc(), Realloc() and Free() were defined as macros having
4547 the values malloc, realloc and free, respectively (except for Win32
4548 compilations). The same is provided for memory debugging code.
4549 OpenSSL already comes with functionality to find memory leaks, but
4550 this gives people a chance to debug other memory problems.
4552 With these changes, a new set of functions and macros have appeared:
4554 CRYPTO_set_mem_debug_functions() [F]
4555 CRYPTO_get_mem_debug_functions() [F]
4556 CRYPTO_dbg_set_options() [F]
4557 CRYPTO_dbg_get_options() [F]
4558 CRYPTO_malloc_debug_init() [M]
4560 The memory debug functions are NULL by default, unless the library
4561 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4562 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4563 gives the standard debugging functions that come with OpenSSL) or
4564 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4565 provided by the library user) must be used. When the standard
4566 debugging functions are used, CRYPTO_dbg_set_options can be used to
4567 request additional information:
4568 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4569 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4571 Also, things like CRYPTO_set_mem_functions will always give the
4572 expected result (the new set of functions is used for allocation
4573 and deallocation) at all times, regardless of platform and compiler
4576 To finish it up, some functions that were never use in any other
4577 way than through macros have a new API and new semantic:
4580 CRYPTO_dbg_realloc()
4583 All macros of value have retained their old syntax.
4584 [Richard Levitte and Bodo Moeller]
4586 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4587 ordering of SMIMECapabilities wasn't in "strength order" and there
4588 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4592 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4593 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4594 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4596 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4597 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4598 functionality to handle multipart/signed properly) and a utility
4599 called 'smime' to call all this stuff. This is based on code I
4600 originally wrote for Celo who have kindly allowed it to be
4601 included in OpenSSL.
4604 *) Add variants des_set_key_checked and des_set_key_unchecked of
4605 des_set_key (aka des_key_sched). Global variable des_check_key
4606 decides which of these is called by des_set_key; this way
4607 des_check_key behaves as it always did, but applications and
4608 the library itself, which was buggy for des_check_key == 1,
4609 have a cleaner way to pick the version they need.
4612 *) New function PKCS12_newpass() which changes the password of a
4616 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4617 dynamic mix. In both cases the ids can be used as an index into the
4618 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4619 functions so they accept a list of the field values and the
4620 application doesn't need to directly manipulate the X509_TRUST
4624 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4628 *) Modify the way the V3 extension code looks up extensions. This now
4629 works in a similar way to the object code: we have some "standard"
4630 extensions in a static table which is searched with OBJ_bsearch()
4631 and the application can add dynamic ones if needed. The file
4632 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4633 updated whenever a new extension is added to the core code and kept
4634 in ext_nid order. There is a simple program 'tabtest.c' which checks
4635 this. New extensions are not added too often so this file can readily
4636 be maintained manually.
4638 There are two big advantages in doing things this way. The extensions
4639 can be looked up immediately and no longer need to be "added" using
4640 X509V3_add_standard_extensions(): this function now does nothing.
4641 [Side note: I get *lots* of email saying the extension code doesn't
4642 work because people forget to call this function]
4643 Also no dynamic allocation is done unless new extensions are added:
4644 so if we don't add custom extensions there is no need to call
4645 X509V3_EXT_cleanup().
4648 *) Modify enc utility's salting as follows: make salting the default. Add a
4649 magic header, so unsalted files fail gracefully instead of just decrypting
4650 to garbage. This is because not salting is a big security hole, so people
4651 should be discouraged from doing it.
4654 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4655 digest to be passed on the command line but it only used this
4656 parameter when signing a certificate. Modified so all relevant
4657 operations are affected by the digest parameter including the
4658 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4659 DSA key was used because it didn't fix the digest.
4662 *) Initial certificate chain verify code. Currently tests the untrusted
4663 certificates for consistency with the verify purpose (which is set
4664 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4666 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4667 this is because it will reject chains with invalid extensions whereas
4668 every previous version of OpenSSL and SSLeay made no checks at all.
4670 Trust code: checks the root CA for the relevant trust settings. Trust
4671 settings have an initial value consistent with the verify purpose: e.g.
4672 if the verify purpose is for SSL client use it expects the CA to be
4673 trusted for SSL client use. However the default value can be changed to
4674 permit custom trust settings: one example of this would be to only trust
4675 certificates from a specific "secure" set of CAs.
4677 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4678 which should be used for version portability: especially since the
4679 verify structure is likely to change more often now.
4681 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4682 to set them. If not set then assume SSL clients will verify SSL servers
4685 Two new options to the verify program: -untrusted allows a set of
4686 untrusted certificates to be passed in and -purpose which sets the
4687 intended purpose of the certificate. If a purpose is set then the
4688 new chain verify code is used to check extension consistency.
4691 *) Support for the authority information access extension.
4694 *) Modify RSA and DSA PEM read routines to transparently handle
4695 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4696 public keys in a format compatible with certificate
4697 SubjectPublicKeyInfo structures. Unfortunately there were already
4698 functions called *_PublicKey_* which used various odd formats so
4699 these are retained for compatibility: however the DSA variants were
4700 never in a public release so they have been deleted. Changed dsa/rsa
4701 utilities to handle the new format: note no releases ever handled public
4702 keys so we should be OK.
4704 The primary motivation for this change is to avoid the same fiasco
4705 that dogs private keys: there are several incompatible private key
4706 formats some of which are standard and some OpenSSL specific and
4707 require various evil hacks to allow partial transparent handling and
4708 even then it doesn't work with DER formats. Given the option anything
4709 other than PKCS#8 should be dumped: but the other formats have to
4710 stay in the name of compatibility.
4712 With public keys and the benefit of hindsight one standard format
4713 is used which works with EVP_PKEY, RSA or DSA structures: though
4714 it clearly returns an error if you try to read the wrong kind of key.
4716 Added a -pubkey option to the 'x509' utility to output the public key.
4717 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4718 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4719 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4720 that do the same as the EVP_PKEY_assign_*() except they up the
4721 reference count of the added key (they don't "swallow" the
4725 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4726 CRLs would fail if the file contained no certificates or no CRLs:
4727 added a new function to read in both types and return the number
4728 read: this means that if none are read it will be an error. The
4729 DER versions of the certificate and CRL reader would always fail
4730 because it isn't possible to mix certificates and CRLs in DER format
4731 without choking one or the other routine. Changed this to just read
4732 a certificate: this is the best we can do. Also modified the code
4733 in apps/verify.c to take notice of return codes: it was previously
4734 attempting to read in certificates from NULL pointers and ignoring
4735 any errors: this is one reason why the cert and CRL reader seemed
4736 to work. It doesn't check return codes from the default certificate
4737 routines: these may well fail if the certificates aren't installed.
4740 *) Code to support otherName option in GeneralName.
4743 *) First update to verify code. Change the verify utility
4744 so it warns if it is passed a self signed certificate:
4745 for consistency with the normal behaviour. X509_verify
4746 has been modified to it will now verify a self signed
4747 certificate if *exactly* the same certificate appears
4748 in the store: it was previously impossible to trust a
4749 single self signed certificate. This means that:
4750 openssl verify ss.pem
4751 now gives a warning about a self signed certificate but
4752 openssl verify -CAfile ss.pem ss.pem
4756 *) For servers, store verify_result in SSL_SESSION data structure
4757 (and add it to external session representation).
4758 This is needed when client certificate verifications fails,
4759 but an application-provided verification callback (set by
4760 SSL_CTX_set_cert_verify_callback) allows accepting the session
4761 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4762 but returns 1): When the session is reused, we have to set
4763 ssl->verify_result to the appropriate error code to avoid
4765 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4767 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4768 case in PKCS7_dataInit() where the signed PKCS7 structure
4769 didn't contain any existing data because it was being created.
4770 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4772 *) Add a salt to the key derivation routines in enc.c. This
4773 forms the first 8 bytes of the encrypted file. Also add a
4774 -S option to allow a salt to be input on the command line.
4777 *) New function X509_cmp(). Oddly enough there wasn't a function
4778 to compare two certificates. We do this by working out the SHA1
4779 hash and comparing that. X509_cmp() will be needed by the trust
4783 *) SSL_get1_session() is like SSL_get_session(), but increments
4784 the reference count in the SSL_SESSION returned.
4785 [Geoff Thorpe <geoff@eu.c2.net>]
4787 *) Fix for 'req': it was adding a null to request attributes.
4788 Also change the X509_LOOKUP and X509_INFO code to handle
4789 certificate auxiliary information.
4792 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4796 *) Add the possibility to add extra information to the memory leak
4797 detecting output, to form tracebacks, showing from where each
4798 allocation was originated: CRYPTO_push_info("constant string") adds
4799 the string plus current file name and line number to a per-thread
4800 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4801 is like calling CYRPTO_pop_info() until the stack is empty.
4802 Also updated memory leak detection code to be multi-thread-safe.
4805 *) Add options -text and -noout to pkcs7 utility and delete the
4806 encryption options which never did anything. Update docs.
4809 *) Add options to some of the utilities to allow the pass phrase
4810 to be included on either the command line (not recommended on
4811 OSes like Unix) or read from the environment. Update the
4812 manpages and fix a few bugs.
4815 *) Add a few manpages for some of the openssl commands.
4818 *) Fix the -revoke option in ca. It was freeing up memory twice,
4819 leaking and not finding already revoked certificates.
4822 *) Extensive changes to support certificate auxiliary information.
4823 This involves the use of X509_CERT_AUX structure and X509_AUX
4824 functions. An X509_AUX function such as PEM_read_X509_AUX()
4825 can still read in a certificate file in the usual way but it
4826 will also read in any additional "auxiliary information". By
4827 doing things this way a fair degree of compatibility can be
4828 retained: existing certificates can have this information added
4829 using the new 'x509' options.
4831 Current auxiliary information includes an "alias" and some trust
4832 settings. The trust settings will ultimately be used in enhanced
4833 certificate chain verification routines: currently a certificate
4834 can only be trusted if it is self signed and then it is trusted
4838 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4839 The problem was that one of the replacement routines had not been working
4840 since SSLeay releases. For now the offending routine has been replaced
4841 with non-optimised assembler. Even so, this now gives around 95%
4842 performance improvement for 1024 bit RSA signs.
4845 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4846 handling. Most clients have the effective key size in bits equal to
4847 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4848 A few however don't do this and instead use the size of the decrypted key
4849 to determine the RC2 key length and the AlgorithmIdentifier to determine
4850 the effective key length. In this case the effective key length can still
4851 be 40 bits but the key length can be 168 bits for example. This is fixed
4852 by manually forcing an RC2 key into the EVP_PKEY structure because the
4853 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4854 the key length and effective key length are equal.
4857 *) Add a bunch of functions that should simplify the creation of
4858 X509_NAME structures. Now you should be able to do:
4859 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4860 and have it automatically work out the correct field type and fill in
4861 the structures. The more adventurous can try:
4862 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4863 and it will (hopefully) work out the correct multibyte encoding.
4866 *) Change the 'req' utility to use the new field handling and multibyte
4867 copy routines. Before the DN field creation was handled in an ad hoc
4868 way in req, ca, and x509 which was rather broken and didn't support
4869 BMPStrings or UTF8Strings. Since some software doesn't implement
4870 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4871 using the dirstring_type option. See the new comment in the default
4872 openssl.cnf for more info.
4875 *) Make crypto/rand/md_rand.c more robust:
4876 - Assure unique random numbers after fork().
4877 - Make sure that concurrent threads access the global counter and
4878 md serializably so that we never lose entropy in them
4879 or use exactly the same state in multiple threads.
4880 Access to the large state is not always serializable because
4881 the additional locking could be a performance killer, and
4882 md should be large enough anyway.
4885 *) New file apps/app_rand.c with commonly needed functionality
4886 for handling the random seed file.
4888 Use the random seed file in some applications that previously did not:
4890 dsaparam -genkey (which also ignored its '-rand' option),
4893 x509 (when signing).
4894 Except on systems with /dev/urandom, it is crucial to have a random
4895 seed file at least for key creation, DSA signing, and for DH exchanges;
4896 for RSA signatures we could do without one.
4898 gendh and gendsa (unlike genrsa) used to read only the first byte
4899 of each file listed in the '-rand' option. The function as previously
4900 found in genrsa is now in app_rand.c and is used by all programs
4901 that support '-rand'.
4904 *) In RAND_write_file, use mode 0600 for creating files;
4905 don't just chmod when it may be too late.
4908 *) Report an error from X509_STORE_load_locations
4909 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4912 *) New function ASN1_mbstring_copy() this copies a string in either
4913 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4914 into an ASN1_STRING type. A mask of permissible types is passed
4915 and it chooses the "minimal" type to use or an error if not type
4919 *) Add function equivalents to the various macros in asn1.h. The old
4920 macros are retained with an M_ prefix. Code inside the library can
4921 use the M_ macros. External code (including the openssl utility)
4922 should *NOT* in order to be "shared library friendly".
4925 *) Add various functions that can check a certificate's extensions
4926 to see if it usable for various purposes such as SSL client,
4927 server or S/MIME and CAs of these types. This is currently
4928 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4929 verification. Also added a -purpose flag to x509 utility to
4930 print out all the purposes.
4933 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4937 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4938 for, obtain and decode and extension and obtain its critical flag.
4939 This allows all the necessary extension code to be handled in a
4940 single function call.
4943 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4944 platforms. See crypto/rc4/rc4_enc.c for further details.
4947 *) New -noout option to asn1parse. This causes no output to be produced
4948 its main use is when combined with -strparse and -out to extract data
4949 from a file (which may not be in ASN.1 format).
4952 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4953 when producing the local key id.
4954 [Richard Levitte <levitte@stacken.kth.se>]
4956 *) New option -dhparam in s_server. This allows a DH parameter file to be
4957 stated explicitly. If it is not stated then it tries the first server
4958 certificate file. The previous behaviour hard coded the filename
4962 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4963 a public key to be input or output. For example:
4964 openssl rsa -in key.pem -pubout -out pubkey.pem
4965 Also added necessary DSA public key functions to handle this.
4968 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4969 in the message. This was handled by allowing
4970 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4971 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4973 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4974 to the end of the strings whereas this didn't. This would cause problems
4975 if strings read with d2i_ASN1_bytes() were later modified.
4976 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4978 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4979 data and it contains EOF it will end up returning an error. This is
4980 caused by input 46 bytes long. The cause is due to the way base64
4981 BIOs find the start of base64 encoded data. They do this by trying a
4982 trial decode on each line until they find one that works. When they
4983 do a flag is set and it starts again knowing it can pass all the
4984 data directly through the decoder. Unfortunately it doesn't reset
4985 the context it uses. This means that if EOF is reached an attempt
4986 is made to pass two EOFs through the context and this causes the
4987 resulting error. This can also cause other problems as well. As is
4988 usual with these problems it takes *ages* to find and the fix is
4989 trivial: move one line.
4990 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4992 *) Ugly workaround to get s_client and s_server working under Windows. The
4993 old code wouldn't work because it needed to select() on sockets and the
4994 tty (for keypresses and to see if data could be written). Win32 only
4995 supports select() on sockets so we select() with a 1s timeout on the
4996 sockets and then see if any characters are waiting to be read, if none
4997 are present then we retry, we also assume we can always write data to
4998 the tty. This isn't nice because the code then blocks until we've
4999 received a complete line of data and it is effectively polling the
5000 keyboard at 1s intervals: however it's quite a bit better than not
5001 working at all :-) A dedicated Windows application might handle this
5002 with an event loop for example.
5005 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5006 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5007 will be called when RSA_sign() and RSA_verify() are used. This is useful
5008 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5009 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5010 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5011 This necessitated the support of an extra signature type NID_md5_sha1
5012 for SSL signatures and modifications to the SSL library to use it instead
5013 of calling RSA_public_decrypt() and RSA_private_encrypt().
5016 *) Add new -verify -CAfile and -CApath options to the crl program, these
5017 will lookup a CRL issuers certificate and verify the signature in a
5018 similar way to the verify program. Tidy up the crl program so it
5019 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5020 less strict. It will now permit CRL extensions even if it is not
5021 a V2 CRL: this will allow it to tolerate some broken CRLs.
5024 *) Initialize all non-automatic variables each time one of the openssl
5025 sub-programs is started (this is necessary as they may be started
5026 multiple times from the "OpenSSL>" prompt).
5027 [Lennart Bang, Bodo Moeller]
5029 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5030 removing all other RSA functionality (this is what NO_RSA does). This
5031 is so (for example) those in the US can disable those operations covered
5032 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5036 *) Non-copying interface to BIO pairs.
5037 (still largely untested)
5040 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5041 ASCII string. This was handled independently in various places before.
5044 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5045 UTF8 strings a character at a time.
5048 *) Use client_version from client hello to select the protocol
5049 (s23_srvr.c) and for RSA client key exchange verification
5050 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5053 *) Add various utility functions to handle SPKACs, these were previously
5054 handled by poking round in the structure internals. Added new function
5055 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5056 print, verify and generate SPKACs. Based on an original idea from
5057 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5060 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5063 *) Allow the config file extension section to be overwritten on the
5064 command line. Based on an original idea from Massimiliano Pala
5065 <madwolf@comune.modena.it>. The new option is called -extensions
5066 and can be applied to ca, req and x509. Also -reqexts to override
5067 the request extensions in req and -crlexts to override the crl extensions
5071 *) Add new feature to the SPKAC handling in ca. Now you can include
5072 the same field multiple times by preceding it by "XXXX." for example:
5075 this is the same syntax as used in the req config file.
5078 *) Allow certificate extensions to be added to certificate requests. These
5079 are specified in a 'req_extensions' option of the req section of the
5080 config file. They can be printed out with the -text option to req but
5081 are otherwise ignored at present.
5084 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5085 data read consists of only the final block it would not decrypted because
5086 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5087 A misplaced 'break' also meant the decrypted final block might not be
5088 copied until the next read.
5091 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5092 a few extra parameters to the DH structure: these will be useful if
5093 for example we want the value of 'q' or implement X9.42 DH.
5096 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5097 provides hooks that allow the default DSA functions or functions on a
5098 "per key" basis to be replaced. This allows hardware acceleration and
5099 hardware key storage to be handled without major modification to the
5100 library. Also added low level modexp hooks and CRYPTO_EX structure and
5101 associated functions.
5104 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5105 as "read only": it can't be written to and the buffer it points to will
5106 not be freed. Reading from a read only BIO is much more efficient than
5107 a normal memory BIO. This was added because there are several times when
5108 an area of memory needs to be read from a BIO. The previous method was
5109 to create a memory BIO and write the data to it, this results in two
5110 copies of the data and an O(n^2) reading algorithm. There is a new
5111 function BIO_new_mem_buf() which creates a read only memory BIO from
5112 an area of memory. Also modified the PKCS#7 routines to use read only
5116 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5117 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5118 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5119 but a retry condition occured while trying to read the rest.
5122 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5123 NID_pkcs7_encrypted by default: this was wrong since this should almost
5124 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5125 the encrypted data type: this is a more sensible place to put it and it
5126 allows the PKCS#12 code to be tidied up that duplicated this
5130 *) Changed obj_dat.pl script so it takes its input and output files on
5131 the command line. This should avoid shell escape redirection problems
5135 *) Initial support for certificate extension requests, these are included
5136 in things like Xenroll certificate requests. Included functions to allow
5137 extensions to be obtained and added.
5140 *) -crlf option to s_client and s_server for sending newlines as
5141 CRLF (as required by many protocols).
5144 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5146 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5147 [Ralf S. Engelschall]
5149 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5150 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5152 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5156 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5157 DH parameters/keys (q is lost during that conversion, but the resulting
5158 DH parameters contain its length).
5160 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5161 much faster than DH_generate_parameters (which creates parameters
5162 where p = 2*q + 1), and also the smaller q makes DH computations
5163 much more efficient (160-bit exponentiation instead of 1024-bit
5164 exponentiation); so this provides a convenient way to support DHE
5165 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5166 utter importance to use
5167 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5169 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5170 when such DH parameters are used, because otherwise small subgroup
5171 attacks may become possible!
5174 *) Avoid memory leak in i2d_DHparams.
5177 *) Allow the -k option to be used more than once in the enc program:
5178 this allows the same encrypted message to be read by multiple recipients.
5181 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5182 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5183 it will always use the numerical form of the OID, even if it has a short
5187 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5188 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5189 otherwise bn_mod_exp was called. In the case of hardware keys for example
5190 no private key components need be present and it might store extra data
5191 in the RSA structure, which cannot be accessed from bn_mod_exp.
5192 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5193 private key operations.
5196 *) Added support for SPARC Linux.
5199 *) pem_password_cb function type incompatibly changed from
5200 typedef int pem_password_cb(char *buf, int size, int rwflag);
5202 ....(char *buf, int size, int rwflag, void *userdata);
5203 so that applications can pass data to their callbacks:
5204 The PEM[_ASN1]_{read,write}... functions and macros now take an
5205 additional void * argument, which is just handed through whenever
5206 the password callback is called.
5207 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5209 New function SSL_CTX_set_default_passwd_cb_userdata.
5211 Compatibility note: As many C implementations push function arguments
5212 onto the stack in reverse order, the new library version is likely to
5213 interoperate with programs that have been compiled with the old
5214 pem_password_cb definition (PEM_whatever takes some data that
5215 happens to be on the stack as its last argument, and the callback
5216 just ignores this garbage); but there is no guarantee whatsoever that
5219 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5220 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5221 problems not only on Windows, but also on some Unix platforms.
5222 To avoid problematic command lines, these definitions are now in an
5223 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5224 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5227 *) MIPS III/IV assembler module is reimplemented.
5230 *) More DES library cleanups: remove references to srand/rand and
5231 delete an unused file.
5234 *) Add support for the the free Netwide assembler (NASM) under Win32,
5235 since not many people have MASM (ml) and it can be hard to obtain.
5236 This is currently experimental but it seems to work OK and pass all
5237 the tests. Check out INSTALL.W32 for info.
5240 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5241 without temporary keys kept an extra copy of the server key,
5242 and connections with temporary keys did not free everything in case
5246 *) New function RSA_check_key and new openssl rsa option -check
5247 for verifying the consistency of RSA keys.
5248 [Ulf Moeller, Bodo Moeller]
5250 *) Various changes to make Win32 compile work:
5251 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5252 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5253 comparison" warnings.
5254 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5257 *) Add a debugging option to PKCS#5 v2 key generation function: when
5258 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5259 derived keys are printed to stderr.
5262 *) Copy the flags in ASN1_STRING_dup().
5263 [Roman E. Pavlov <pre@mo.msk.ru>]
5265 *) The x509 application mishandled signing requests containing DSA
5266 keys when the signing key was also DSA and the parameters didn't match.
5268 It was supposed to omit the parameters when they matched the signing key:
5269 the verifying software was then supposed to automatically use the CA's
5270 parameters if they were absent from the end user certificate.
5272 Omitting parameters is no longer recommended. The test was also
5273 the wrong way round! This was probably due to unusual behaviour in
5274 EVP_cmp_parameters() which returns 1 if the parameters match.
5275 This meant that parameters were omitted when they *didn't* match and
5276 the certificate was useless. Certificates signed with 'ca' didn't have
5278 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5280 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5281 The interface is as follows:
5282 Applications can use
5283 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5284 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5285 "off" is now the default.
5286 The library internally uses
5287 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5288 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5289 to disable memory-checking temporarily.
5291 Some inconsistent states that previously were possible (and were
5292 even the default) are now avoided.
5294 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5295 with each memory chunk allocated; this is occasionally more helpful
5296 than just having a counter.
5298 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5300 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5304 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5305 which largely parallels "options", but is for changing API behaviour,
5306 whereas "options" are about protocol behaviour.
5307 Initial "mode" flags are:
5309 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5310 a single record has been written.
5311 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5312 retries use the same buffer location.
5313 (But all of the contents must be
5317 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5320 *) Fix problems with no-hmac etc.
5321 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5323 *) New functions RSA_get_default_method(), RSA_set_method() and
5324 RSA_get_method(). These allows replacement of RSA_METHODs without having
5325 to mess around with the internals of an RSA structure.
5328 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5329 Also really enable memory leak checks in openssl.c and in some
5331 [Chad C. Mulligan, Bodo Moeller]
5333 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5334 up the length of negative integers. This has now been simplified to just
5335 store the length when it is first determined and use it later, rather
5336 than trying to keep track of where data is copied and updating it to
5338 [Steve Henson, reported by Brien Wheeler
5339 <bwheeler@authentica-security.com>]
5341 *) Add a new function PKCS7_signatureVerify. This allows the verification
5342 of a PKCS#7 signature but with the signing certificate passed to the
5343 function itself. This contrasts with PKCS7_dataVerify which assumes the
5344 certificate is present in the PKCS#7 structure. This isn't always the
5345 case: certificates can be omitted from a PKCS#7 structure and be
5346 distributed by "out of band" means (such as a certificate database).
5349 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5350 function prototypes in pem.h, also change util/mkdef.pl to add the
5351 necessary function names.
5354 *) mk1mf.pl (used by Windows builds) did not properly read the
5355 options set by Configure in the top level Makefile, and Configure
5356 was not even able to write more than one option correctly.
5357 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5360 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5361 file to be loaded from a BIO or FILE pointer. The BIO version will
5362 for example allow memory BIOs to contain config info.
5365 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5366 Whoever hopes to achieve shared-library compatibility across versions
5367 must use this, not the compile-time macro.
5368 (Exercise 0.9.4: Which is the minimum library version required by
5370 Note: All this applies only to multi-threaded programs, others don't
5374 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5375 through a BIO pair triggered the default case, i.e.
5376 SSLerr(...,SSL_R_UNKNOWN_STATE).
5379 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5380 can use the SSL library even if none of the specific BIOs is
5384 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5385 for the encoded length.
5386 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5388 *) Add initial documentation of the X509V3 functions.
5391 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5392 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5393 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5394 secure PKCS#8 private key format with a high iteration count.
5397 *) Fix determination of Perl interpreter: A perl or perl5
5398 _directory_ in $PATH was also accepted as the interpreter.
5399 [Ralf S. Engelschall]
5401 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5402 wrong with it but it was very old and did things like calling
5403 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5407 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5408 to use the new extension code.
5411 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5412 with macros. This should make it easier to change their form, add extra
5413 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5417 *) Add to configuration table a new entry that can specify an alternative
5418 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5419 according to Mark Crispin <MRC@Panda.COM>.
5423 *) DES CBC did not update the IV. Weird.
5426 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5427 Changing the behaviour of the former might break existing programs --
5428 where IV updating is needed, des_ncbc_encrypt can be used.
5431 *) When bntest is run from "make test" it drives bc to check its
5432 calculations, as well as internally checking them. If an internal check
5433 fails, it needs to cause bc to give a non-zero result or make test carries
5434 on without noticing the failure. Fixed.
5437 *) DES library cleanups.
5440 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5441 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5442 ciphers. NOTE: although the key derivation function has been verified
5443 against some published test vectors it has not been extensively tested
5444 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5448 *) Instead of "mkdir -p", which is not fully portable, use new
5449 Perl script "util/mkdir-p.pl".
5452 *) Rewrite the way password based encryption (PBE) is handled. It used to
5453 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5454 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5455 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5456 the 'parameter' field of the AlgorithmIdentifier is passed to the
5457 underlying key generation function so it must do its own ASN1 parsing.
5458 This has also changed the EVP_PBE_CipherInit() function which now has a
5459 'parameter' argument instead of literal salt and iteration count values
5460 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5463 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5464 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5465 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5466 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5467 value was just used as a "magic string" and not used directly its
5468 value doesn't matter.
5471 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5475 *) "linux-sparc64" configuration (ultrapenguin).
5476 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5477 "linux-sparc" configuration.
5478 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5480 *) config now generates no-xxx options for missing ciphers.
5483 *) Support the EBCDIC character set (work in progress).
5484 File ebcdic.c not yet included because it has a different license.
5485 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5487 *) Support BS2000/OSD-POSIX.
5488 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5490 *) Make callbacks for key generation use void * instead of char *.
5493 *) Make S/MIME samples compile (not yet tested).
5496 *) Additional typesafe stacks.
5499 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5503 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5505 *) New configuration variant "sco5-gcc".
5507 *) Updated some demos.
5508 [Sean O Riordain, Wade Scholine]
5510 *) Add missing BIO_free at exit of pkcs12 application.
5513 *) Fix memory leak in conf.c.
5516 *) Updates for Win32 to assembler version of MD5.
5519 *) Set #! path to perl in apps/der_chop to where we found it
5520 instead of using a fixed path.
5523 *) SHA library changes for irix64-mips4-cc.
5526 *) Improvements for VMS support.
5530 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5532 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5533 This also avoids the problems with SC4.2 and unpatched SC5.
5534 [Andy Polyakov <appro@fy.chalmers.se>]
5536 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5537 These are required because of the typesafe stack would otherwise break
5538 existing code. If old code used a structure member which used to be STACK
5539 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5540 sk_num or sk_value it would produce an error because the num, data members
5541 are not present in STACK_OF. Now it just produces a warning. sk_set
5542 replaces the old method of assigning a value to sk_value
5543 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5544 that does this will no longer work (and should use sk_set instead) but
5545 this could be regarded as a "questionable" behaviour anyway.
5548 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5549 correctly handle encrypted S/MIME data.
5552 *) Change type of various DES function arguments from des_cblock
5553 (which means, in function argument declarations, pointer to char)
5554 to des_cblock * (meaning pointer to array with 8 char elements),
5555 which allows the compiler to do more typechecking; it was like
5556 that back in SSLeay, but with lots of ugly casts.
5558 Introduce new type const_des_cblock.
5561 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5562 problems: find RecipientInfo structure that matches recipient certificate
5563 and initialise the ASN1 structures properly based on passed cipher.
5566 *) Belatedly make the BN tests actually check the results.
5569 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5570 to and from BNs: it was completely broken. New compilation option
5571 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5572 key elements as negative integers.
5575 *) Reorganize and speed up MD5.
5576 [Andy Polyakov <appro@fy.chalmers.se>]
5579 [Richard Levitte <richard@levitte.org>]
5581 *) New option -out to asn1parse to allow the parsed structure to be
5582 output to a file. This is most useful when combined with the -strparse
5583 option to examine the output of things like OCTET STRINGS.
5586 *) Make SSL library a little more fool-proof by not requiring any longer
5587 that SSL_set_{accept,connect}_state be called before
5588 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5589 in many applications because usually everything *appeared* to work as
5590 intended anyway -- now it really works as intended).
5593 *) Move openssl.cnf out of lib/.
5596 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5597 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5598 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5599 [Ralf S. Engelschall]
5601 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5602 handle PKCS#7 enveloped data properly.
5603 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5605 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5606 copying pointers. The cert_st handling is changed by this in
5607 various ways (and thus what used to be known as ctx->default_cert
5608 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5609 any longer when s->cert does not give us what we need).
5610 ssl_cert_instantiate becomes obsolete by this change.
5611 As soon as we've got the new code right (possibly it already is?),
5612 we have solved a couple of bugs of the earlier code where s->cert
5613 was used as if it could not have been shared with other SSL structures.
5615 Note that using the SSL API in certain dirty ways now will result
5616 in different behaviour than observed with earlier library versions:
5617 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5618 does not influence s as it used to.
5620 In order to clean up things more thoroughly, inside SSL_SESSION
5621 we don't use CERT any longer, but a new structure SESS_CERT
5622 that holds per-session data (if available); currently, this is
5623 the peer's certificate chain and, for clients, the server's certificate
5624 and temporary key. CERT holds only those values that can have
5625 meaningful defaults in an SSL_CTX.
5628 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5629 from the internal representation. Various PKCS#7 fixes: remove some
5630 evil casts and set the enc_dig_alg field properly based on the signing
5634 *) Allow PKCS#12 password to be set from the command line or the
5635 environment. Let 'ca' get its config file name from the environment
5636 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5640 *) Allow certificate policies extension to use an IA5STRING for the
5641 organization field. This is contrary to the PKIX definition but
5642 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5646 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5647 without disallowing inline assembler and the like for non-pedantic builds.
5650 *) Support Borland C++ builder.
5651 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5656 *) SHA-1 cleanups and performance enhancements.
5657 [Andy Polyakov <appro@fy.chalmers.se>]
5659 *) Sparc v8plus assembler for the bignum library.
5660 [Andy Polyakov <appro@fy.chalmers.se>]
5662 *) Accept any -xxx and +xxx compiler options in Configure.
5665 *) Update HPUX configuration.
5668 *) Add missing sk_<type>_unshift() function to safestack.h
5669 [Ralf S. Engelschall]
5671 *) New function SSL_CTX_use_certificate_chain_file that sets the
5672 "extra_cert"s in addition to the certificate. (This makes sense
5673 only for "PEM" format files, as chains as a whole are not
5677 *) Support verify_depth from the SSL API.
5678 x509_vfy.c had what can be considered an off-by-one-error:
5679 Its depth (which was not part of the external interface)
5680 was actually counting the number of certificates in a chain;
5681 now it really counts the depth.
5684 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5685 instead of X509err, which often resulted in confusing error
5686 messages since the error codes are not globally unique
5687 (e.g. an alleged error in ssl3_accept when a certificate
5688 didn't match the private key).
5690 *) New function SSL_CTX_set_session_id_context that allows to set a default
5691 value (so that you don't need SSL_set_session_id_context for each
5692 connection using the SSL_CTX).
5695 *) OAEP decoding bug fix.
5698 *) Support INSTALL_PREFIX for package builders, as proposed by
5702 *) New Configure options "threads" and "no-threads". For systems
5703 where the proper compiler options are known (currently Solaris
5704 and Linux), "threads" is the default.
5707 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5710 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5711 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5712 such as /usr/local/bin.
5715 *) "make linux-shared" to build shared libraries.
5716 [Niels Poppe <niels@netbox.org>]
5718 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5721 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5722 extension adding in x509 utility.
5725 *) Remove NOPROTO sections and error code comments.
5728 *) Partial rewrite of the DEF file generator to now parse the ANSI
5732 *) New Configure options --prefix=DIR and --openssldir=DIR.
5735 *) Complete rewrite of the error code script(s). It is all now handled
5736 by one script at the top level which handles error code gathering,
5737 header rewriting and C source file generation. It should be much better
5738 than the old method: it now uses a modified version of Ulf's parser to
5739 read the ANSI prototypes in all header files (thus the old K&R definitions
5740 aren't needed for error creation any more) and do a better job of
5741 translating function codes into names. The old 'ASN1 error code imbedded
5742 in a comment' is no longer necessary and it doesn't use .err files which
5743 have now been deleted. Also the error code call doesn't have to appear all
5744 on one line (which resulted in some large lines...).
5747 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5750 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5751 0 (which usually indicates a closed connection), but continue reading.
5754 *) Fix some race conditions.
5757 *) Add support for CRL distribution points extension. Add Certificate
5758 Policies and CRL distribution points documentation.
5761 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5764 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5765 8 of keying material. Merlin has also confirmed interop with this fix
5766 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5767 [Merlin Hughes <merlin@baltimore.ie>]
5769 *) Fix lots of warnings.
5770 [Richard Levitte <levitte@stacken.kth.se>]
5772 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5773 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5774 [Richard Levitte <levitte@stacken.kth.se>]
5776 *) Fix problems with sizeof(long) == 8.
5777 [Andy Polyakov <appro@fy.chalmers.se>]
5779 *) Change functions to ANSI C.
5782 *) Fix typos in error codes.
5783 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5785 *) Remove defunct assembler files from Configure.
5788 *) SPARC v8 assembler BIGNUM implementation.
5789 [Andy Polyakov <appro@fy.chalmers.se>]
5791 *) Support for Certificate Policies extension: both print and set.
5792 Various additions to support the r2i method this uses.
5795 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5796 return a const string when you are expecting an allocated buffer.
5799 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5800 types DirectoryString and DisplayText.
5803 *) Add code to allow r2i extensions to access the configuration database,
5804 add an LHASH database driver and add several ctx helper functions.
5807 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5808 fail when they extended the size of a BIGNUM.
5811 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5812 support typesafe stack.
5815 *) Fix typo in SSL_[gs]et_options().
5816 [Nils Frostberg <nils@medcom.se>]
5818 *) Delete various functions and files that belonged to the (now obsolete)
5819 old X509V3 handling code.
5822 *) New Configure option "rsaref".
5825 *) Don't auto-generate pem.h.
5828 *) Introduce type-safe ASN.1 SETs.
5831 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5832 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5834 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5835 that links with OpenSSL (well at least cause lots of warnings), but fear
5836 not: the conversion is trivial, and it eliminates loads of evil casts. A
5837 few STACKed things have been converted already. Feel free to convert more.
5838 In the fullness of time, I'll do away with the STACK type altogether.
5841 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5842 specified in <certfile> by updating the entry in the index.txt file.
5843 This way one no longer has to edit the index.txt file manually for
5844 revoking a certificate. The -revoke option does the gory details now.
5845 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5847 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5848 `-text' option at all and this way the `-noout -text' combination was
5849 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5850 [Ralf S. Engelschall]
5852 *) Make sure a corresponding plain text error message exists for the
5853 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5854 verify callback function determined that a certificate was revoked.
5855 [Ralf S. Engelschall]
5857 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5858 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5859 all available cipers including rc5, which was forgotten until now.
5860 In order to let the testing shell script know which algorithms
5861 are available, a new (up to now undocumented) command
5862 "openssl list-cipher-commands" is used.
5865 *) Bugfix: s_client occasionally would sleep in select() when
5866 it should have checked SSL_pending() first.
5869 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5870 the raw DSA values prior to ASN.1 encoding.
5873 *) Tweaks to Configure
5874 [Niels Poppe <niels@netbox.org>]
5876 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5880 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5883 *) New config option to avoid instructions that are illegal on the 80386.
5884 The default code is faster, but requires at least a 486.
5887 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5888 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5889 same as SSL2_VERSION anyway.
5892 *) New "-showcerts" option for s_client.
5895 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5896 application. Various cleanups and fixes.
5899 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5900 modify error routines to work internally. Add error codes and PBE init
5901 to library startup routines.
5904 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5905 packing functions to asn1 and evp. Changed function names and error
5906 codes along the way.
5909 *) PKCS12 integration: and so it begins... First of several patches to
5910 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5911 objects to objects.h
5914 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5915 and display support for Thawte strong extranet extension.
5918 *) Add LinuxPPC support.
5919 [Jeff Dubrule <igor@pobox.org>]
5921 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5922 bn_div_words in alpha.s.
5923 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5925 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5926 OAEP isn't supported when OpenSSL is built with RSAref.
5927 [Ulf Moeller <ulf@fitug.de>]
5929 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5930 so they no longer are missing under -DNOPROTO.
5931 [Soren S. Jorvang <soren@t.dk>]
5934 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5936 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5937 doesn't work when the session is reused. Coming soon!
5940 *) Fix a security hole, that allows sessions to be reused in the wrong
5941 context thus bypassing client cert protection! All software that uses
5942 client certs and session caches in multiple contexts NEEDS PATCHING to
5943 allow session reuse! A fuller solution is in the works.
5944 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5946 *) Some more source tree cleanups (removed obsolete files
5947 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5948 permission on "config" script to be executable) and a fix for the INSTALL
5950 [Ulf Moeller <ulf@fitug.de>]
5952 *) Remove some legacy and erroneous uses of malloc, free instead of
5954 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5956 *) Make rsa_oaep_test return non-zero on error.
5957 [Ulf Moeller <ulf@fitug.de>]
5959 *) Add support for native Solaris shared libraries. Configure
5960 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5961 if someone would make that last step automatic.
5962 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5964 *) ctx_size was not built with the right compiler during "make links". Fixed.
5967 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5968 except NULL ciphers". This means the default cipher list will no longer
5969 enable NULL ciphers. They need to be specifically enabled e.g. with
5970 the string "DEFAULT:eNULL".
5973 *) Fix to RSA private encryption routines: if p < q then it would
5974 occasionally produce an invalid result. This will only happen with
5975 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5978 *) Be less restrictive and allow also `perl util/perlpath.pl
5979 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5980 because this way one can also use an interpreter named `perl5' (which is
5981 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5982 installed as `perl').
5983 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5985 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5986 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5988 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5989 advapi32.lib to Win32 build and change the pem test comparision
5990 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5991 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5992 and crypto/des/ede_cbcm_enc.c.
5995 *) DES quad checksum was broken on big-endian architectures. Fixed.
5998 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5999 Win32 test batch file so it (might) work again. The Win32 test batch file
6000 is horrible: I feel ill....
6003 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6004 in e_os.h. Audit of header files to check ANSI and non ANSI
6005 sections: 10 functions were absent from non ANSI section and not exported
6006 from Windows DLLs. Fixed up libeay.num for new functions.
6009 *) Make `openssl version' output lines consistent.
6010 [Ralf S. Engelschall]
6012 *) Fix Win32 symbol export lists for BIO functions: Added
6013 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6014 to ms/libeay{16,32}.def.
6015 [Ralf S. Engelschall]
6017 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6018 fine under Unix and passes some trivial tests I've now added. But the
6019 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6020 added to make sure no one expects that this stuff really works in the
6021 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6022 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6024 [Ralf S. Engelschall]
6026 *) Fix the generation of two part addresses in perl.
6027 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6029 *) Add config entry for Linux on MIPS.
6030 [John Tobey <jtobey@channel1.com>]
6032 *) Make links whenever Configure is run, unless we are on Windoze.
6035 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6036 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6040 *) Add a useful kludge to allow package maintainers to specify compiler and
6041 other platforms details on the command line without having to patch the
6042 Configure script everytime: One now can use ``perl Configure
6043 <id>:<details>'', i.e. platform ids are allowed to have details appended
6044 to them (seperated by colons). This is treated as there would be a static
6045 pre-configured entry in Configure's %table under key <id> with value
6046 <details> and ``perl Configure <id>'' is called. So, when you want to
6047 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6048 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6049 now, which overrides the FreeBSD-elf entry on-the-fly.
6050 [Ralf S. Engelschall]
6052 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6055 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6056 on the `perl Configure ...' command line. This way one can compile
6057 OpenSSL libraries with Position Independent Code (PIC) which is needed
6058 for linking it into DSOs.
6059 [Ralf S. Engelschall]
6061 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6065 *) Cleaned up the LICENSE document: The official contact for any license
6066 questions now is the OpenSSL core team under openssl-core@openssl.org.
6067 And add a paragraph about the dual-license situation to make sure people
6068 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6069 to the OpenSSL toolkit.
6070 [Ralf S. Engelschall]
6072 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6073 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6074 Additonally cleaned up the `make links' target: Remove unnecessary
6075 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6076 to speed processing and no longer clutter the display with confusing
6077 stuff. Instead only the actually done links are displayed.
6078 [Ralf S. Engelschall]
6080 *) Permit null encryption ciphersuites, used for authentication only. It used
6081 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6082 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6086 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6087 signed attributes when verifying signatures (this would break them),
6088 the detached data encoding was wrong and public keys obtained using
6089 X509_get_pubkey() weren't freed.
6092 *) Add text documentation for the BUFFER functions. Also added a work around
6093 to a Win95 console bug. This was triggered by the password read stuff: the
6094 last character typed gets carried over to the next fread(). If you were
6095 generating a new cert request using 'req' for example then the last
6096 character of the passphrase would be CR which would then enter the first
6100 *) Added the new `Includes OpenSSL Cryptography Software' button as
6101 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6102 button and can be used by applications based on OpenSSL to show the
6103 relationship to the OpenSSL project.
6104 [Ralf S. Engelschall]
6106 *) Remove confusing variables in function signatures in files
6107 ssl/ssl_lib.c and ssl/ssl.h.
6108 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6110 *) Don't install bss_file.c under PREFIX/include/
6111 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6113 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6114 functions that return function pointers and has support for NT specific
6115 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6116 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6117 unsigned to signed types: this was killing the Win32 compile.
6120 *) Add new certificate file to stack functions,
6121 SSL_add_dir_cert_subjects_to_stack() and
6122 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6123 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6124 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6125 This means that Apache-SSL and similar packages don't have to mess around
6126 to add as many CAs as they want to the preferred list.
6129 *) Experiment with doxygen documentation. Currently only partially applied to
6131 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6132 openssl.doxy as the configuration file.
6135 *) Get rid of remaining C++-style comments which strict C compilers hate.
6136 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6138 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6139 compiled in by default: it has problems with large keys.
6142 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6143 DH private keys and/or callback functions which directly correspond to
6144 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6145 is needed for applications which have to configure certificates on a
6146 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6148 For the RSA certificate situation is makes no difference, but
6149 for the DSA certificate situation this fixes the "no shared cipher"
6150 problem where the OpenSSL cipher selection procedure failed because the
6151 temporary keys were not overtaken from the context and the API provided
6152 no way to reconfigure them.
6153 The new functions now let applications reconfigure the stuff and they
6154 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6155 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6156 non-public-API function ssl_cert_instantiate() is used as a helper
6157 function and also to reduce code redundancy inside ssl_rsa.c.
6158 [Ralf S. Engelschall]
6160 *) Move s_server -dcert and -dkey options out of the undocumented feature
6161 area because they are useful for the DSA situation and should be
6162 recognized by the users.
6163 [Ralf S. Engelschall]
6165 *) Fix the cipher decision scheme for export ciphers: the export bits are
6166 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6167 SSL_EXP_MASK. So, the original variable has to be used instead of the
6168 already masked variable.
6169 [Richard Levitte <levitte@stacken.kth.se>]
6171 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6172 [Richard Levitte <levitte@stacken.kth.se>]
6174 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6175 from `int' to `unsigned int' because it's a length and initialized by
6176 EVP_DigestFinal() which expects an `unsigned int *'.
6177 [Richard Levitte <levitte@stacken.kth.se>]
6179 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6180 script. Instead use the usual Shell->Perl transition trick.
6181 [Ralf S. Engelschall]
6183 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6184 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6185 -noout -modulus' as it's already the case for `openssl rsa -noout
6186 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6187 currently the public key is printed (a decision which was already done by
6188 `openssl dsa -modulus' in the past) which serves a similar purpose.
6189 Additionally the NO_RSA no longer completely removes the whole -modulus
6190 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6192 [Ralf S. Engelschall]
6194 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6195 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6196 [Arne Ansper <arne@ats.cyber.ee>]
6198 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6199 to be added. Now both 'req' and 'ca' can use new objects defined in the
6203 *) Add cool BIO that does syslog (or event log on NT).
6204 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6206 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6207 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6208 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6209 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6212 *) Add preliminary config info for new extension code.
6215 *) Make RSA_NO_PADDING really use no padding.
6216 [Ulf Moeller <ulf@fitug.de>]
6218 *) Generate errors when private/public key check is done.
6221 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6222 for some CRL extensions and new objects added.
6225 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6226 key usage extension and fuller support for authority key id.
6229 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6230 padding method for RSA, which is recommended for new applications in PKCS
6231 #1 v2.0 (RFC 2437, October 1998).
6232 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6233 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6234 against Bleichbacher's attack on RSA.
6235 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6238 *) Updates to the new SSL compression code
6239 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6241 *) Fix so that the version number in the master secret, when passed
6242 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6243 (because the server will not accept higher), that the version number
6244 is 0x03,0x01, not 0x03,0x00
6245 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6247 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6248 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6249 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6252 *) Support for RAW extensions where an arbitrary extension can be
6253 created by including its DER encoding. See apps/openssl.cnf for
6257 *) Make sure latest Perl versions don't interpret some generated C array
6258 code as Perl array code in the crypto/err/err_genc.pl script.
6259 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6261 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6262 not many people have the assembler. Various Win32 compilation fixes and
6263 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6267 *) Modify configure script 'Configure' to automatically create crypto/date.h
6268 file under Win32 and also build pem.h from pem.org. New script
6269 util/mkfiles.pl to create the MINFO file on environments that can't do a
6270 'make files': perl util/mkfiles.pl >MINFO should work.
6273 *) Major rework of DES function declarations, in the pursuit of correctness
6274 and purity. As a result, many evil casts evaporated, and some weirdness,
6275 too. You may find this causes warnings in your code. Zapping your evil
6276 casts will probably fix them. Mostly.
6279 *) Fix for a typo in asn1.h. Bug fix to object creation script
6280 obj_dat.pl. It considered a zero in an object definition to mean
6281 "end of object": none of the objects in objects.h have any zeros
6282 so it wasn't spotted.
6283 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6285 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6286 Masking (CBCM). In the absence of test vectors, the best I have been able
6287 to do is check that the decrypt undoes the encrypt, so far. Send me test
6288 vectors if you have them.
6291 *) Correct calculation of key length for export ciphers (too much space was
6292 allocated for null ciphers). This has not been tested!
6295 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6296 message is now correct (it understands "crypto" and "ssl" on its
6297 command line). There is also now an "update" option. This will update
6298 the util/ssleay.num and util/libeay.num files with any new functions.
6300 perl util/mkdef.pl crypto ssl update
6301 it will update them.
6304 *) Overhauled the Perl interface (perl/*):
6305 - ported BN stuff to OpenSSL's different BN library
6306 - made the perl/ source tree CVS-aware
6307 - renamed the package from SSLeay to OpenSSL (the files still contain
6308 their history because I've copied them in the repository)
6309 - removed obsolete files (the test scripts will be replaced
6310 by better Test::Harness variants in the future)
6311 [Ralf S. Engelschall]
6313 *) First cut for a very conservative source tree cleanup:
6314 1. merge various obsolete readme texts into doc/ssleay.txt
6315 where we collect the old documents and readme texts.
6316 2. remove the first part of files where I'm already sure that we no
6317 longer need them because of three reasons: either they are just temporary
6318 files which were left by Eric or they are preserved original files where
6319 I've verified that the diff is also available in the CVS via "cvs diff
6320 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6321 the crypto/md/ stuff).
6322 [Ralf S. Engelschall]
6324 *) More extension code. Incomplete support for subject and issuer alt
6325 name, issuer and authority key id. Change the i2v function parameters
6326 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6327 what that's for :-) Fix to ASN1 macro which messed up
6328 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6331 *) Preliminary support for ENUMERATED type. This is largely copied from the
6335 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6336 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6338 *) Make sure `make rehash' target really finds the `openssl' program.
6339 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6341 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6342 like to hear about it if this slows down other processors.
6345 *) Add CygWin32 platform information to Configure script.
6346 [Alan Batie <batie@aahz.jf.intel.com>]
6348 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6349 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6351 *) New program nseq to manipulate netscape certificate sequences
6354 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6358 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6359 but the BN code had some problems that would cause failures when
6360 doing certificate verification and some other functions.
6361 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6363 *) Add ASN1 and PEM code to support netscape certificate sequences.
6366 *) Add ASN1 and PEM code to support netscape certificate sequences.
6369 *) Add several PKIX and private extended key usage OIDs.
6372 *) Modify the 'ca' program to handle the new extension code. Modify
6373 openssl.cnf for new extension format, add comments.
6376 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6377 and add a sample to openssl.cnf so req -x509 now adds appropriate
6381 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6382 error code, add initial support to X509_print() and x509 application.
6385 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6386 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6387 stuff is currently isolated and isn't even compiled yet.
6390 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6391 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6392 Removed the versions check from X509 routines when loading extensions:
6393 this allows certain broken certificates that don't set the version
6394 properly to be processed.
6397 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6398 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6399 can still be regenerated with "make depend".
6402 *) Spelling mistake in C version of CAST-128.
6403 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6405 *) Changes to the error generation code. The perl script err-code.pl
6406 now reads in the old error codes and retains the old numbers, only
6407 adding new ones if necessary. It also only changes the .err files if new
6408 codes are added. The makefiles have been modified to only insert errors
6409 when needed (to avoid needlessly modifying header files). This is done
6410 by only inserting errors if the .err file is newer than the auto generated
6411 C file. To rebuild all the error codes from scratch (the old behaviour)
6412 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6413 or delete all the .err files.
6416 *) CAST-128 was incorrectly implemented for short keys. The C version has
6417 been fixed, but is untested. The assembler versions are also fixed, but
6418 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6419 to regenerate it if needed.
6420 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6421 Hagino <itojun@kame.net>]
6423 *) File was opened incorrectly in randfile.c.
6424 [Ulf Möller <ulf@fitug.de>]
6426 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6427 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6428 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6429 al: it's just almost always a UTCTime. Note this patch adds new error
6430 codes so do a "make errors" if there are problems.
6433 *) Correct Linux 1 recognition in config.
6434 [Ulf Möller <ulf@fitug.de>]
6436 *) Remove pointless MD5 hash when using DSA keys in ca.
6437 [Anonymous <nobody@replay.com>]
6439 *) Generate an error if given an empty string as a cert directory. Also
6440 generate an error if handed NULL (previously returned 0 to indicate an
6441 error, but didn't set one).
6442 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6444 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6447 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6448 parameters. This was causing a warning which killed off the Win32 compile.
6451 *) Remove C++ style comments from crypto/bn/bn_local.h.
6452 [Neil Costigan <neil.costigan@celocom.com>]
6454 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6455 based on a text string, looking up short and long names and finally
6456 "dot" format. The "dot" format stuff didn't work. Added new function
6457 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6458 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6459 OID is not part of the table.
6462 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6463 X509_LOOKUP_by_alias().
6466 *) Sort openssl functions by name.
6469 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6470 encryption from sample DSA keys (in case anyone is interested the password
6474 *) Make _all_ *_free functions accept a NULL pointer.
6475 [Frans Heymans <fheymans@isaserver.be>]
6477 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6479 [Anonymous <nobody@replay.com>]
6481 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6482 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6484 *) Don't blow it for numeric -newkey arguments to apps/req.
6485 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6487 *) Temp key "for export" tests were wrong in s3_srvr.c.
6488 [Anonymous <nobody@replay.com>]
6490 *) Add prototype for temp key callback functions
6491 SSL_CTX_set_tmp_{rsa,dh}_callback().
6494 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6495 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6498 *) X509_name_add_entry() freed the wrong thing after an error.
6499 [Arne Ansper <arne@ats.cyber.ee>]
6501 *) rsa_eay.c would attempt to free a NULL context.
6502 [Arne Ansper <arne@ats.cyber.ee>]
6504 *) BIO_s_socket() had a broken should_retry() on Windoze.
6505 [Arne Ansper <arne@ats.cyber.ee>]
6507 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6508 [Arne Ansper <arne@ats.cyber.ee>]
6510 *) Make sure the already existing X509_STORE->depth variable is initialized
6511 in X509_STORE_new(), but document the fact that this variable is still
6512 unused in the certificate verification process.
6513 [Ralf S. Engelschall]
6515 *) Fix the various library and apps files to free up pkeys obtained from
6516 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6519 *) Fix reference counting in X509_PUBKEY_get(). This makes
6520 demos/maurice/example2.c work, amongst others, probably.
6521 [Steve Henson and Ben Laurie]
6523 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6524 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6525 are no longer created. This way we have a single and consistent command
6526 line interface `openssl <command>', similar to `cvs <command>'.
6527 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6529 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6530 BIT STRING wrapper always have zero unused bits.
6533 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6536 *) Make the top-level INSTALL documentation easier to understand.
6539 *) Makefiles updated to exit if an error occurs in a sub-directory
6540 make (including if user presses ^C) [Paul Sutton]
6542 *) Make Montgomery context stuff explicit in RSA data structure.
6545 *) Fix build order of pem and err to allow for generated pem.h.
6548 *) Fix renumbering bug in X509_NAME_delete_entry().
6551 *) Enhanced the err-ins.pl script so it makes the error library number
6552 global and can add a library name. This is needed for external ASN1 and
6553 other error libraries.
6556 *) Fixed sk_insert which never worked properly.
6559 *) Fix ASN1 macros so they can handle indefinite length construted
6560 EXPLICIT tags. Some non standard certificates use these: they can now
6564 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6565 into a single doc/ssleay.txt bundle. This way the information is still
6566 preserved but no longer messes up this directory. Now it's new room for
6567 the new set of documenation files.
6568 [Ralf S. Engelschall]
6570 *) SETs were incorrectly DER encoded. This was a major pain, because they
6571 shared code with SEQUENCEs, which aren't coded the same. This means that
6572 almost everything to do with SETs or SEQUENCEs has either changed name or
6573 number of arguments.
6574 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6576 *) Fix test data to work with the above.
6579 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6580 was already fixed by Eric for 0.9.1 it seems.
6581 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6583 *) Autodetect FreeBSD3.
6586 *) Fix various bugs in Configure. This affects the following platforms:
6590 unixware-2.0-pentium
6594 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6595 before they are needed.
6598 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6602 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6604 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6605 changed SSLeay to OpenSSL in version strings.
6606 [Ralf S. Engelschall]
6608 *) Some fixups to the top-level documents.
6611 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6612 because the symlink to include/ was missing.
6613 [Ralf S. Engelschall]
6615 *) Incorporated the popular no-RSA/DSA-only patches
6616 which allow to compile a RSA-free SSLeay.
6617 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6619 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6620 when "ssleay" is still not found.
6621 [Ralf S. Engelschall]
6623 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6624 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6626 *) Updated the README file.
6627 [Ralf S. Engelschall]
6629 *) Added various .cvsignore files in the CVS repository subdirs
6630 to make a "cvs update" really silent.
6631 [Ralf S. Engelschall]
6633 *) Recompiled the error-definition header files and added
6634 missing symbols to the Win32 linker tables.
6635 [Ralf S. Engelschall]
6637 *) Cleaned up the top-level documents;
6638 o new files: CHANGES and LICENSE
6639 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6640 o merged COPYRIGHT into LICENSE
6641 o removed obsolete TODO file
6642 o renamed MICROSOFT to INSTALL.W32
6643 [Ralf S. Engelschall]
6645 *) Removed dummy files from the 0.9.1b source tree:
6646 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6647 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6648 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6649 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6650 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6651 [Ralf S. Engelschall]
6653 *) Added various platform portability fixes.
6656 *) The Genesis of the OpenSSL rpject:
6657 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6658 Young and Tim J. Hudson created while they were working for C2Net until
6660 [The OpenSSL Project]
6663 Changes between 0.9.0b and 0.9.1b [not released]
6665 *) Updated a few CA certificates under certs/
6668 *) Changed some BIGNUM api stuff.
6671 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6672 DGUX x86, Linux Alpha, etc.
6675 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6676 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6680 *) Add -strparse option to asn1pars program which parses nested
6682 [Dr Stephen Henson <shenson@bigfoot.com>]
6684 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6687 *) DSA fix for "ca" program.
6690 *) Added "-genkey" option to "dsaparam" program.
6693 *) Added RIPE MD160 (rmd160) message digest.
6696 *) Added -a (all) option to "ssleay version" command.
6699 *) Added PLATFORM define which is the id given to Configure.
6702 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6705 *) Extended the ASN.1 parser routines.
6708 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6711 *) Added a BN_CTX to the BN library.
6714 *) Fixed the weak key values in DES library
6717 *) Changed API in EVP library for cipher aliases.
6720 *) Added support for RC2/64bit cipher.
6723 *) Converted the lhash library to the crypto/mem.c functions.
6726 *) Added more recognized ASN.1 object ids.
6729 *) Added more RSA padding checks for SSL/TLS.
6732 *) Added BIO proxy/filter functionality.
6735 *) Added extra_certs to SSL_CTX which can be used
6736 send extra CA certificates to the client in the CA cert chain sending
6737 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6740 *) Now Fortezza is denied in the authentication phase because
6741 this is key exchange mechanism is not supported by SSLeay at all.
6744 *) Additional PKCS1 checks.
6747 *) Support the string "TLSv1" for all TLS v1 ciphers.
6750 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6751 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6754 *) Fixed a few memory leaks.
6757 *) Fixed various code and comment typos.
6760 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6761 bytes sent in the client random.
6762 [Edward Bishop <ebishop@spyglass.com>]