5 Changes between 0.9.7a and 0.9.8 [xx XXX xxxx]
7 *) Support for policyConstraints certificate extension.
10 *) Support for policyMappings certificate extension.
13 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
14 ENGINE as defaults for all supported algorithms irrespective of
15 the 'flags' parameter. 'flags' is now honoured, so applications
16 should make sure they are passing it correctly.
19 *) Make sure the default DSA_METHOD implementation only uses its
20 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
21 and change its own handlers to be NULL so as to remove unnecessary
22 indirection. This lets alternative implementations fallback to the
23 default implementation more easily.
26 *) Support for directoryName in GeneralName related extensions
30 *) Make it possible to link applications using Makefile.shared.
31 Make that possible even when linking against static libraries!
34 *) Various fixes to base64 BIO and non blocking I/O. On write
35 flushes were not handled properly if the BIO retried. On read
36 data was not being buffered properly and had various logic bugs.
39 *) Support for single pass processing for S/MIME signing. This now
40 means that S/MIME signing can be done from a pipe, in addition
41 cleartext signing (multipart/signed type) is effectively streaming
42 and the signed data does not need to be all held in memory.
44 This is done with a new flag PKCS7_STREAM. When this flag is set
45 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
46 is done after the data is output (and digests calculated) in
50 *) Add full support for -rpath/-R, both in shared libraries and
51 applications, at least on the platforms where it's known how
55 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
56 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
57 will now compute a table of multiples of the generator that
58 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
59 faster (notably in the case of a single point multiplication,
61 [Nils Larsch, Bodo Moeller]
63 *) IPv6 support for certificate extensions. The various extensions
64 which use the IP:a.b.c.d can now take IPv6 addresses using the
65 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
69 *) Added an ENGINE that implements RSA by performing private key
70 exponentiations with the GMP library. The conversions to and from
71 GMP's mpz_t format aren't optimised nor are any montgomery forms
72 cached, and on x86 it appears OpenSSL's own performance has caught up.
73 However there are likely to be other architectures where GMP could
74 provide a boost. This ENGINE is not built in by default, but it can be
75 specified at Configure time and should be accompanied by the necessary
77 ./config -DOPENSSL_USE_GMP -lgmp
80 *) "openssl engine" will not display ENGINE/DSO load failure errors when
81 testing availability of engines with "-t" - the old behaviour is
82 produced by increasing the feature's verbosity with "-tt".
85 *) ECDSA routines: under certain error conditions uninitialized BN objects
86 could be freed. Solution: make sure initialization is performed early
87 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
91 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
92 and DH_METHOD (eg. by ENGINE implementations) to override the normal
93 software implementations. For DSA and DH, parameter generation can
94 also be overriden by providing the appropriate method callbacks.
97 *) Change the "progress" mechanism used in key-generation and
98 primality testing to functions that take a new BN_GENCB pointer in
99 place of callback/argument pairs. The new API functions have "_ex"
100 postfixes and the older functions are reimplemented as wrappers for
101 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
102 declarations of the old functions to help (graceful) attempts to
103 migrate to the new functions. Also, the new key-generation API
104 functions operate on a caller-supplied key-structure and return
105 success/failure rather than returning a key or NULL - this is to
106 help make "keygen" another member function of RSA_METHOD etc.
108 Example for using the new callback interface:
110 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
114 BN_GENCB_set(&my_cb, my_callback, my_arg);
116 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
117 /* For the meaning of a, b in calls to my_callback(), see the
118 * documentation of the function that calls the callback.
119 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
120 * my_callback should return 1 if it wants BN_is_prime_ex()
121 * to continue, or 0 to stop.
126 *) Change the ZLIB compression method to be stateful, and make it
127 available to TLS with the number defined in
128 draft-ietf-tls-compression-04.txt.
131 *) Add the ASN.1 structures and functions for CertificatePair, which
132 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
134 CertificatePair ::= SEQUENCE {
135 forward [0] Certificate OPTIONAL,
136 reverse [1] Certificate OPTIONAL,
137 -- at least one of the pair shall be present -- }
139 Also implement the PEM functions to read and write certificate
140 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
142 This needed to be defined, mostly for the sake of the LDAP
143 attribute crossCertificatePair, but may prove useful elsewhere as
147 *) Make it possible to inhibit symlinking of shared libraries in
148 Makefile.shared, for Cygwin's sake.
151 *) Extend the BIGNUM API by creating new macros that behave like
154 void BN_set_sign(BIGNUM *a, int neg);
155 int BN_get_sign(const BIGNUM *a);
157 and avoid the need to access 'a->neg' directly in applications.
158 [Nils Larsch <nla@trustcenter.de>]
160 *) Implement fast modular reduction for pseudo-Mersenne primes
161 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
162 EC_GROUP_new_curve_GFp() will now automatically use this
164 [Nils Larsch <nla@trustcenter.de>]
166 *) Add new lock type (CRYPTO_LOCK_BN).
169 *) Change the ENGINE framework to automatically load engines
170 dynamically from specific directories unless they could be
171 found to already be built in or loaded. Move all the
172 current engines except for the cryptodev one to a new
174 The engines in engines/ are built as shared libraries if
175 the "shared" options was given to ./Configure or ./config.
176 Otherwise, they are inserted in libcrypto.a.
177 /usr/local/ssl/engines is the default directory for dynamic
178 engines, but that can be overriden at configure time through
179 the usual use of --prefix and/or --openssldir, and at run
180 time with the environment variable OPENSSL_ENGINES.
181 [Geoff Thorpe and Richard Levitte]
183 *) Add Makefile.shared, a helper makefile to build shared
184 libraries. Addapt Makefile.org.
187 *) Add version info to Win32 DLLs.
188 [Peter 'Luna' Runestig" <peter@runestig.com>]
190 *) Add new 'medium level' PKCS#12 API. Certificates and keys
191 can be added using this API to created arbitrary PKCS#12
192 files while avoiding the low level API.
194 New options to PKCS12_create(), key or cert can be NULL and
195 will then be omitted from the output file. The encryption
196 algorithm NIDs can be set to -1 for no encryption, the mac
197 iteration count can be set to 0 to omit the mac.
199 Enhance pkcs12 utility by making the -nokeys and -nocerts
200 options work when creating a PKCS#12 file. New option -nomac
201 to omit the mac, NONE can be set for an encryption algorithm.
202 New code is modified to use the enhanced PKCS12_create()
203 instead of the low level API.
206 *) Extend ASN1 encoder to support indefinite length constructed
207 encoding. This can output sequences tags and octet strings in
208 this form. Modify pk7_asn1.c to support indefinite length
209 encoding. This is experimental and needs additional code to
210 be useful, such as an ASN1 bio and some enhanced streaming
213 Extend template encode functionality so that tagging is passed
214 down to the template encoder.
217 *) Let 'openssl req' fail if an argument to '-newkey' is not
218 recognized instead of using RSA as a default.
221 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
222 As these are not official, they are not included in "ALL";
223 the "ECCdraft" ciphersuite group alias can be used to select them.
224 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
226 *) Add ECDH engine support.
227 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
229 *) Add ECDH in new directory crypto/ecdh/.
230 [Douglas Stebila (Sun Microsystems Laboratories)]
232 *) Let BN_rand_range() abort with an error after 100 iterations
233 without success (which indicates a broken PRNG).
236 *) Change BN_mod_sqrt() so that it verifies that the input value
237 is really the square of the return value. (Previously,
238 BN_mod_sqrt would show GIGO behaviour.)
241 *) Add named elliptic curves over binary fields from X9.62, SECG,
242 and WAP/WTLS; add OIDs that were still missing.
244 [Sheueling Chang Shantz and Douglas Stebila
245 (Sun Microsystems Laboratories)]
247 *) Extend the EC library for elliptic curves over binary fields
248 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
251 EC_GF2m_simple_method
255 EC_GROUP_new_curve_GF2m
256 EC_GROUP_set_curve_GF2m
257 EC_GROUP_get_curve_GF2m
258 EC_POINT_set_affine_coordinates_GF2m
259 EC_POINT_get_affine_coordinates_GF2m
260 EC_POINT_set_compressed_coordinates_GF2m
262 Point compression for binary fields is disabled by default for
263 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
266 As binary polynomials are represented as BIGNUMs, various members
267 of the EC_GROUP and EC_POINT data structures can be shared
268 between the implementations for prime fields and binary fields;
269 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
270 are essentially identical to their ..._GFp counterparts.
271 (For simplicity, the '..._GFp' prefix has been dropped from
272 various internal method names.)
274 An internal 'field_div' method (similar to 'field_mul' and
275 'field_sqr') has been added; this is used only for binary fields.
277 [Sheueling Chang Shantz and Douglas Stebila
278 (Sun Microsystems Laboratories)]
280 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
281 through methods ('mul', 'precompute_mult').
283 The generic implementations (now internally called 'ec_wNAF_mul'
284 and 'ec_wNAF_precomputed_mult') remain the default if these
285 methods are undefined.
287 [Sheueling Chang Shantz and Douglas Stebila
288 (Sun Microsystems Laboratories)]
290 *) New function EC_GROUP_get_degree, which is defined through
291 EC_METHOD. For curves over prime fields, this returns the bit
292 length of the modulus.
294 [Sheueling Chang Shantz and Douglas Stebila
295 (Sun Microsystems Laboratories)]
297 *) New functions EC_GROUP_dup, EC_POINT_dup.
298 (These simply call ..._new and ..._copy).
300 [Sheueling Chang Shantz and Douglas Stebila
301 (Sun Microsystems Laboratories)]
303 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
304 Polynomials are represented as BIGNUMs (where the sign bit is not
305 used) in the following functions [macros]:
308 BN_GF2m_sub [= BN_GF2m_add]
309 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
310 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
311 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
313 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
314 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
315 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
316 BN_GF2m_cmp [= BN_ucmp]
318 (Note that only the 'mod' functions are actually for fields GF(2^m).
319 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
321 For some functions, an the irreducible polynomial defining a
322 field can be given as an 'unsigned int[]' with strictly
323 decreasing elements giving the indices of those bits that are set;
324 i.e., p[] represents the polynomial
325 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
327 p[0] > p[1] > ... > p[k] = 0.
328 This applies to the following functions:
333 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
334 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
337 BN_GF2m_mod_solve_quad_arr
341 Conversion can be performed by the following functions:
346 bntest.c has additional tests for binary polynomial arithmetic.
348 Two implementations for BN_GF2m_mod_div() are available.
349 The default algorithm simply uses BN_GF2m_mod_inv() and
350 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
351 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
352 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
354 [Sheueling Chang Shantz and Douglas Stebila
355 (Sun Microsystems Laboratories)]
357 *) Add new error code 'ERR_R_DISABLED' that can be used when some
358 functionality is disabled at compile-time.
359 [Douglas Stebila <douglas.stebila@sun.com>]
361 *) Change default behaviour of 'openssl asn1parse' so that more
362 information is visible when viewing, e.g., a certificate:
364 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
365 mode the content of non-printable OCTET STRINGs is output in a
366 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
367 avoid the appearance of a printable string.
368 [Nils Larsch <nla@trustcenter.de>]
370 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
372 EC_GROUP_set_asn1_flag()
373 EC_GROUP_get_asn1_flag()
374 EC_GROUP_set_point_conversion_form()
375 EC_GROUP_get_point_conversion_form()
376 These control ASN1 encoding details:
377 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
378 has been set to OPENSSL_EC_NAMED_CURVE.
379 - Points are encoded in uncompressed form by default; options for
380 asn1_for are as for point2oct, namely
381 POINT_CONVERSION_COMPRESSED
382 POINT_CONVERSION_UNCOMPRESSED
383 POINT_CONVERSION_HYBRID
385 Also add 'seed' and 'seed_len' members to EC_GROUP with access
389 EC_GROUP_get_seed_len()
390 This is used only for ASN1 purposes (so far).
391 [Nils Larsch <nla@trustcenter.de>]
393 *) Add 'field_type' member to EC_METHOD, which holds the NID
394 of the appropriate field type OID. The new function
395 EC_METHOD_get_field_type() returns this value.
396 [Nils Larsch <nla@trustcenter.de>]
403 providing useful interfaces to EC_POINT_point2oct() and
404 EC_POINT_oct2point().
405 [Nils Larsch <nla@trustcenter.de>]
407 *) Change internals of the EC library so that the functions
408 EC_GROUP_set_generator()
409 EC_GROUP_get_generator()
411 EC_GROUP_get_cofactor()
412 are implemented directly in crypto/ec/ec_lib.c and not dispatched
413 to methods, which would lead to unnecessary code duplication when
414 adding different types of curves.
415 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
417 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
418 arithmetic, and such that modified wNAFs are generated
419 (which avoid length expansion in many cases).
422 *) Add a function EC_GROUP_check_discriminant() (defined via
423 EC_METHOD) that verifies that the curve discriminant is non-zero.
425 Add a function EC_GROUP_check() that makes some sanity tests
426 on a EC_GROUP, its generator and order. This includes
427 EC_GROUP_check_discriminant().
428 [Nils Larsch <nla@trustcenter.de>]
430 *) Add ECDSA in new directory crypto/ecdsa/.
432 Add applications 'openssl ecparam' and 'openssl ecdsa'
433 (these are based on 'openssl dsaparam' and 'openssl dsa').
435 ECDSA support is also included in various other files across the
436 library. Most notably,
437 - 'openssl req' now has a '-newkey ecdsa:file' option;
438 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
439 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
440 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
441 them suitable for ECDSA where domain parameters must be
442 extracted before the specific public key;
443 - ECDSA engine support has been added.
444 [Nils Larsch <nla@trustcenter.de>]
446 *) Include some named elliptic curves, and add OIDs from X9.62,
447 SECG, and WAP/WTLS. Each curve can be obtained from the new
449 EC_GROUP_new_by_nid(),
450 and the list of available named curves can be obtained with
451 EC_get_builtin_curves().
452 Also add a 'curve_name' member to EC_GROUP objects, which can be
456 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
458 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
459 was actually never needed) and in BN_mul(). The removal in BN_mul()
460 required a small change in bn_mul_part_recursive() and the addition
461 of the functions bn_cmp_part_words(), bn_sub_part_words() and
462 bn_add_part_words(), which do the same thing as bn_cmp_words(),
463 bn_sub_words() and bn_add_words() except they take arrays with
467 Changes between 0.9.7a and 0.9.7b [xx XXX 2003]
469 *) Countermeasure against the Klima-Pokorny-Rosa extension of
470 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
471 a protocol version number mismatch like a decryption error
472 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
475 *) Turn on RSA blinding by default in the default implementation
476 to avoid a timing attack. Applications that don't want it can call
477 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
478 They would be ill-advised to do so in most cases.
479 [Ben Laurie, Steve Henson, Geoff Thorpe]
481 *) Change RSA blinding code so that it works when the PRNG is not
482 seeded (in this case, the secret RSA exponent is abused as
483 an unpredictable seed -- if it is not unpredictable, there
484 is no point in blinding anyway).
487 yet to be integrated into this CVS branch:
488 - Geoff's ENGINE_set_default() fix
490 *) Target "mingw" now allows native Windows code to be generated in
491 the Cygwin environment as well as with the MinGW compiler.
494 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
496 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
497 via timing by performing a MAC computation even if incorrrect
498 block cipher padding has been found. This is a countermeasure
499 against active attacks where the attacker has to distinguish
500 between bad padding and a MAC verification error. (CAN-2003-0078)
502 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
503 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
504 Martin Vuagnoux (EPFL, Ilion)]
506 *) Make the no-err option work as intended. The intention with no-err
507 is not to have the whole error stack handling routines removed from
508 libcrypto, it's only intended to remove all the function name and
509 reason texts, thereby removing some of the footprint that may not
510 be interesting if those errors aren't displayed anyway.
512 NOTE: it's still possible for any application or module to have it's
513 own set of error texts inserted. The routines are there, just not
514 used by default when no-err is given.
517 *) Add support for FreeBSD on IA64.
518 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
520 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
521 Kerberos function mit_des_cbc_cksum(). Before this change,
522 the value returned by DES_cbc_cksum() was like the one from
523 mit_des_cbc_cksum(), except the bytes were swapped.
524 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
526 *) Allow an application to disable the automatic SSL chain building.
527 Before this a rather primitive chain build was always performed in
528 ssl3_output_cert_chain(): an application had no way to send the
529 correct chain if the automatic operation produced an incorrect result.
531 Now the chain builder is disabled if either:
533 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
535 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
537 The reasoning behind this is that an application would not want the
538 auto chain building to take place if extra chain certificates are
539 present and it might also want a means of sending no additional
540 certificates (for example the chain has two certificates and the
544 *) Add the possibility to build without the ENGINE framework.
545 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
547 *) Under Win32 gmtime() can return NULL: check return value in
548 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
551 *) DSA routines: under certain error conditions uninitialized BN objects
552 could be freed. Solution: make sure initialization is performed early
553 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
554 Nils Larsch <nla@trustcenter.de> via PR#459)
557 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
558 checked on reconnect on the client side, therefore session resumption
559 could still fail with a "ssl session id is different" error. This
560 behaviour is masked when SSL_OP_ALL is used due to
561 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
562 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
566 *) IA-32 assembler support enhancements: unified ELF targets, support
567 for SCO/Caldera platforms, fix for Cygwin shared build.
570 *) Add support for FreeBSD on sparc64. As a consequence, support for
571 FreeBSD on non-x86 processors is separate from x86 processors on
572 the config script, much like the NetBSD support.
573 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
575 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
577 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
578 code (06) was taken as the first octet of the session ID and the last
579 octet was ignored consequently. As a result SSLv2 client side session
580 caching could not have worked due to the session ID mismatch between
582 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
586 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
587 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
591 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
592 seems that in spite of existing for more than a year, many application
593 author have done nothing to provide the necessary callbacks, which
594 means that this particular engine will not work properly anywhere.
595 This is a very unfortunate situation which forces us, in the name
596 of usability, to give the hw_ncipher.c a static lock, which is part
598 NOTE: This is for the 0.9.7 series ONLY. This hack will never
599 appear in 0.9.8 or later. We EXPECT application authors to have
600 dealt properly with this when 0.9.8 is released (unless we actually
601 make such changes in the libcrypto locking code that changes will
602 have to be made anyway).
605 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
606 octets have been read, EOF or an error occurs. Without this change
607 some truncated ASN1 structures will not produce an error.
610 *) Disable Heimdal support, since it hasn't been fully implemented.
611 Still give the possibility to force the use of Heimdal, but with
612 warnings and a request that patches get sent to openssl-dev.
615 *) Add the VC-CE target, introduce the WINCE sysname, and add
616 INSTALL.WCE and appropriate conditionals to make it build.
617 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
619 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
620 cygssl-x.y.z.dll, where x, y and z are the major, minor and
621 edit numbers of the version.
622 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
624 *) Introduce safe string copy and catenation functions
625 (BUF_strlcpy() and BUF_strlcat()).
626 [Ben Laurie (CHATS) and Richard Levitte]
628 *) Avoid using fixed-size buffers for one-line DNs.
631 *) Add BUF_MEM_grow_clean() to avoid information leakage when
632 resizing buffers containing secrets, and use where appropriate.
635 *) Avoid using fixed size buffers for configuration file location.
638 *) Avoid filename truncation for various CA files.
641 *) Use sizeof in preference to magic numbers.
644 *) Avoid filename truncation in cert requests.
647 *) Add assertions to check for (supposedly impossible) buffer
651 *) Don't cache truncated DNS entries in the local cache (this could
652 potentially lead to a spoofing attack).
655 *) Fix various buffers to be large enough for hex/decimal
656 representations in a platform independent manner.
659 *) Add CRYPTO_realloc_clean() to avoid information leakage when
660 resizing buffers containing secrets, and use where appropriate.
663 *) Add BIO_indent() to avoid much slightly worrying code to do
667 *) Convert sprintf()/BIO_puts() to BIO_printf().
670 *) buffer_gets() could terminate with the buffer only half
674 *) Add assertions to prevent user-supplied crypto functions from
675 overflowing internal buffers by having large block sizes, etc.
678 *) New OPENSSL_assert() macro (similar to assert(), but enabled
682 *) Eliminate unused copy of key in RC4.
685 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
688 *) Fix off-by-one error in EGD path.
691 *) If RANDFILE path is too long, ignore instead of truncating.
694 *) Eliminate unused and incorrectly sized X.509 structure
698 *) Eliminate unused and dangerous function knumber().
701 *) Eliminate unused and dangerous structure, KSSL_ERR.
704 *) Protect against overlong session ID context length in an encoded
705 session object. Since these are local, this does not appear to be
709 *) Change from security patch (see 0.9.6e below) that did not affect
710 the 0.9.6 release series:
712 Remote buffer overflow in SSL3 protocol - an attacker could
713 supply an oversized master key in Kerberos-enabled versions.
717 *) Change the SSL kerb5 codes to match RFC 2712.
720 *) Make -nameopt work fully for req and add -reqopt switch.
721 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
723 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
724 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
726 *) Make sure tests can be performed even if the corresponding algorithms
727 have been removed entirely. This was also the last step to make
728 OpenSSL compilable with DJGPP under all reasonable conditions.
729 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
731 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
732 to allow version independent disabling of normally unselected ciphers,
733 which may be activated as a side-effect of selecting a single cipher.
735 (E.g., cipher list string "RSA" enables ciphersuites that are left
736 out of "ALL" because they do not provide symmetric encryption.
737 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
738 [Lutz Jaenicke, Bodo Moeller]
740 *) Add appropriate support for separate platform-dependent build
741 directories. The recommended way to make a platform-dependent
742 build directory is the following (tested on Linux), maybe with
745 # Place yourself outside of the OpenSSL source tree. In
746 # this example, the environment variable OPENSSL_SOURCE
747 # is assumed to contain the absolute OpenSSL source directory.
748 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
749 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
750 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
751 mkdir -p `dirname $F`
752 ln -s $OPENSSL_SOURCE/$F $F
755 To be absolutely sure not to disturb the source tree, a "make clean"
756 is a good thing. If it isn't successfull, don't worry about it,
757 it probably means the source directory is very clean.
760 *) Make sure any ENGINE control commands make local copies of string
761 pointers passed to them whenever necessary. Otherwise it is possible
762 the caller may have overwritten (or deallocated) the original string
763 data when a later ENGINE operation tries to use the stored values.
764 [Götz Babin-Ebell <babinebell@trustcenter.de>]
766 *) Improve diagnostics in file reading and command-line digests.
767 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
769 *) Add AES modes CFB and OFB to the object database. Correct an
770 error in AES-CFB decryption.
773 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
774 allows existing EVP_CIPHER_CTX structures to be reused after
775 calling EVP_*Final(). This behaviour is used by encryption
776 BIOs and some applications. This has the side effect that
777 applications must explicitly clean up cipher contexts with
778 EVP_CIPHER_CTX_cleanup() or they will leak memory.
781 *) Check the values of dna and dnb in bn_mul_recursive before calling
782 bn_mul_comba (a non zero value means the a or b arrays do not contain
783 n2 elements) and fallback to bn_mul_normal if either is not zero.
786 *) Fix escaping of non-ASCII characters when using the -subj option
787 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
790 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
791 form for "surname", serialNumber has no short form.
792 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
793 therefore remove "mail" short name for "internet 7".
794 The OID for unique identifiers in X509 certificates is
795 x500UniqueIdentifier, not uniqueIdentifier.
796 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
799 *) Add an "init" command to the ENGINE config module and auto initialize
800 ENGINEs. Without any "init" command the ENGINE will be initialized
801 after all ctrl commands have been executed on it. If init=1 the
802 ENGINE is initailized at that point (ctrls before that point are run
803 on the uninitialized ENGINE and after on the initialized one). If
804 init=0 then the ENGINE will not be iniatialized at all.
807 *) Fix the 'app_verify_callback' interface so that the user-defined
808 argument is actually passed to the callback: In the
809 SSL_CTX_set_cert_verify_callback() prototype, the callback
810 declaration has been changed from
813 int (*cb)(X509_STORE_CTX *,void *);
814 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
815 i=s->ctx->app_verify_callback(&ctx)
816 has been changed into
817 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
819 To update applications using SSL_CTX_set_cert_verify_callback(),
820 a dummy argument can be added to their callback functions.
821 [D. K. Smetters <smetters@parc.xerox.com>]
823 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
824 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
826 *) Add and OPENSSL_LOAD_CONF define which will cause
827 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
828 This allows older applications to transparently support certain
829 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
830 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
831 load the config file and OPENSSL_add_all_algorithms_conf() which will
832 always load it have also been added.
835 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
836 Adjust NIDs and EVP layer.
837 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
839 *) Config modules support in openssl utility.
841 Most commands now load modules from the config file,
842 though in a few (such as version) this isn't done
843 because it couldn't be used for anything.
845 In the case of ca and req the config file used is
846 the same as the utility itself: that is the -config
847 command line option can be used to specify an
851 *) Move default behaviour from OPENSSL_config(). If appname is NULL
852 use "openssl_conf" if filename is NULL use default openssl config file.
855 *) Add an argument to OPENSSL_config() to allow the use of an alternative
856 config section name. Add a new flag to tolerate a missing config file
857 and move code to CONF_modules_load_file().
860 *) Support for crypto accelerator cards from Accelerated Encryption
861 Processing, www.aep.ie. (Use engine 'aep')
862 The support was copied from 0.9.6c [engine] and adapted/corrected
863 to work with the new engine framework.
864 [AEP Inc. and Richard Levitte]
866 *) Support for SureWare crypto accelerator cards from Baltimore
867 Technologies. (Use engine 'sureware')
868 The support was copied from 0.9.6c [engine] and adapted
869 to work with the new engine framework.
872 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
873 make the newer ENGINE framework commands for the CHIL engine work.
874 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
876 *) Make it possible to produce shared libraries on ReliantUNIX.
877 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
879 *) Add the configuration target debug-linux-ppro.
880 Make 'openssl rsa' use the general key loading routines
881 implemented in apps.c, and make those routines able to
882 handle the key format FORMAT_NETSCAPE and the variant
884 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
886 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
887 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
889 *) Add -keyform to rsautl, and document -engine.
890 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
892 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
893 BIO_R_NO_SUCH_FILE error code rather than the generic
894 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
899 ERR_peek_last_error_line
900 ERR_peek_last_error_line_data.
904 ERR_peek_error_line_data,
905 but report on the latest error recorded rather than the first one
906 still in the error queue.
907 [Ben Laurie, Bodo Moeller]
909 *) default_algorithms option in ENGINE config module. This allows things
911 default_algorithms = ALL
912 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
915 *) Prelminary ENGINE config module.
918 *) New experimental application configuration code.
921 *) Change the AES code to follow the same name structure as all other
922 symmetric ciphers, and behave the same way. Move everything to
923 the directory crypto/aes, thereby obsoleting crypto/rijndael.
924 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
926 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
927 [Ben Laurie and Theo de Raadt]
929 *) Add option to output public keys in req command.
930 [Massimiliano Pala madwolf@openca.org]
932 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
933 (up to about 10% better than before for P-192 and P-224).
936 *) New functions/macros
938 SSL_CTX_set_msg_callback(ctx, cb)
939 SSL_CTX_set_msg_callback_arg(ctx, arg)
940 SSL_set_msg_callback(ssl, cb)
941 SSL_set_msg_callback_arg(ssl, arg)
943 to request calling a callback function
945 void cb(int write_p, int version, int content_type,
946 const void *buf, size_t len, SSL *ssl, void *arg)
948 whenever a protocol message has been completely received
949 (write_p == 0) or sent (write_p == 1). Here 'version' is the
950 protocol version according to which the SSL library interprets
951 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
952 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
953 the content type as defined in the SSL 3.0/TLS 1.0 protocol
954 specification (change_cipher_spec(20), alert(21), handshake(22)).
955 'buf' and 'len' point to the actual message, 'ssl' to the
956 SSL object, and 'arg' is the application-defined value set by
957 SSL[_CTX]_set_msg_callback_arg().
959 'openssl s_client' and 'openssl s_server' have new '-msg' options
960 to enable a callback that displays all protocol messages.
963 *) Change the shared library support so shared libraries are built as
964 soon as the corresponding static library is finished, and thereby get
965 openssl and the test programs linked against the shared library.
966 This still only happens when the keyword "shard" has been given to
967 the configuration scripts.
969 NOTE: shared library support is still an experimental thing, and
970 backward binary compatibility is still not guaranteed.
971 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
973 *) Add support for Subject Information Access extension.
974 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
976 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
977 additional bytes when new memory had to be allocated, not just
978 when reusing an existing buffer.
981 *) New command line and configuration option 'utf8' for the req command.
982 This allows field values to be specified as UTF8 strings.
985 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
986 runs for the former and machine-readable output for the latter.
989 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
990 of the e-mail address in the DN (i.e., it will go into a certificate
991 extension only). The new configuration file option 'email_in_dn = no'
993 [Massimiliano Pala madwolf@openca.org]
995 *) Change all functions with names starting with des_ to be starting
996 with DES_ instead. Add wrappers that are compatible with libdes,
997 but are named _ossl_old_des_*. Finally, add macros that map the
998 des_* symbols to the corresponding _ossl_old_des_* if libdes
999 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1000 desired, the des_* symbols will be mapped to DES_*, with one
1003 Since we provide two compatibility mappings, the user needs to
1004 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1005 compatibility is desired. The default (i.e., when that macro
1006 isn't defined) is OpenSSL 0.9.6c compatibility.
1008 There are also macros that enable and disable the support of old
1009 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1010 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1011 are defined, the default will apply: to support the old des routines.
1013 In either case, one must include openssl/des.h to get the correct
1014 definitions. Do not try to just include openssl/des_old.h, that
1017 NOTE: This is a major break of an old API into a new one. Software
1018 authors are encouraged to switch to the DES_ style functions. Some
1019 time in the future, des_old.h and the libdes compatibility functions
1020 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1021 default), and then completely removed.
1024 *) Test for certificates which contain unsupported critical extensions.
1025 If such a certificate is found during a verify operation it is
1026 rejected by default: this behaviour can be overridden by either
1027 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1028 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1029 X509_supported_extension() has also been added which returns 1 if a
1030 particular extension is supported.
1033 *) Modify the behaviour of EVP cipher functions in similar way to digests
1034 to retain compatibility with existing code.
1037 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1038 compatibility with existing code. In particular the 'ctx' parameter does
1039 not have to be to be initialized before the call to EVP_DigestInit() and
1040 it is tidied up after a call to EVP_DigestFinal(). New function
1041 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1042 EVP_MD_CTX_copy() changed to not require the destination to be
1043 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1044 requires the destination to be valid.
1046 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1047 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1050 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1051 so that complete 'Handshake' protocol structures are kept in memory
1052 instead of overwriting 'msg_type' and 'length' with 'body' data.
1055 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1056 [Massimo Santin via Richard Levitte]
1058 *) Major restructuring to the underlying ENGINE code. This includes
1059 reduction of linker bloat, separation of pure "ENGINE" manipulation
1060 (initialisation, etc) from functionality dealing with implementations
1061 of specific crypto iterfaces. This change also introduces integrated
1062 support for symmetric ciphers and digest implementations - so ENGINEs
1063 can now accelerate these by providing EVP_CIPHER and EVP_MD
1064 implementations of their own. This is detailed in crypto/engine/README
1065 as it couldn't be adequately described here. However, there are a few
1066 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1067 were changed in the original introduction of ENGINE code have now
1068 reverted back - the hooking from this code to ENGINE is now a good
1069 deal more passive and at run-time, operations deal directly with
1070 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1071 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1072 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1073 they were not being used by the framework as there is no concept of a
1074 BIGNUM_METHOD and they could not be generalised to the new
1075 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1076 ENGINE_cpy() has been removed as it cannot be consistently defined in
1080 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1083 *) Change mkdef.pl to sort symbols that get the same entry number,
1084 and make sure the automatically generated functions ERR_load_*
1085 become part of libeay.num as well.
1088 *) New function SSL_renegotiate_pending(). This returns true once
1089 renegotiation has been requested (either SSL_renegotiate() call
1090 or HelloRequest/ClientHello receveived from the peer) and becomes
1091 false once a handshake has been completed.
1092 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1093 sends a HelloRequest, but does not ensure that a handshake takes
1094 place. SSL_renegotiate_pending() is useful for checking if the
1095 client has followed the request.)
1098 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1099 By default, clients may request session resumption even during
1100 renegotiation (if session ID contexts permit); with this option,
1101 session resumption is possible only in the first handshake.
1103 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1104 more bits available for options that should not be part of
1105 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1108 *) Add some demos for certificate and certificate request creation.
1111 *) Make maximum certificate chain size accepted from the peer application
1112 settable (SSL*_get/set_max_cert_list()), as proposed by
1113 "Douglas E. Engert" <deengert@anl.gov>.
1116 *) Add support for shared libraries for Unixware-7
1117 (Boyd Lynn Gerber <gerberb@zenez.com>).
1120 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1121 be done prior to destruction. Use this to unload error strings from
1122 ENGINEs that load their own error strings. NB: This adds two new API
1123 functions to "get" and "set" this destroy handler in an ENGINE.
1126 *) Alter all existing ENGINE implementations (except "openssl" and
1127 "openbsd") to dynamically instantiate their own error strings. This
1128 makes them more flexible to be built both as statically-linked ENGINEs
1129 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1130 Also, add stub code to each that makes building them as self-contained
1131 shared-libraries easier (see README.ENGINE).
1134 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1135 implementations into applications that are completely implemented in
1136 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1137 commands that can be used to configure what shared-library to load and
1138 to control aspects of the way it is handled. Also, made an update to
1139 the README.ENGINE file that brings its information up-to-date and
1140 provides some information and instructions on the "dynamic" ENGINE
1141 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1144 *) Make it possible to unload ranges of ERR strings with a new
1145 "ERR_unload_strings" function.
1148 *) Add a copy() function to EVP_MD.
1151 *) Make EVP_MD routines take a context pointer instead of just the
1152 md_data void pointer.
1155 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1156 that the digest can only process a single chunk of data
1157 (typically because it is provided by a piece of
1158 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1159 is only going to provide a single chunk of data, and hence the
1160 framework needn't accumulate the data for oneshot drivers.
1163 *) As with "ERR", make it possible to replace the underlying "ex_data"
1164 functions. This change also alters the storage and management of global
1165 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1166 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1167 index counters. The API functions that use this state have been changed
1168 to take a "class_index" rather than pointers to the class's local STACK
1169 and counter, and there is now an API function to dynamically create new
1170 classes. This centralisation allows us to (a) plug a lot of the
1171 thread-safety problems that existed, and (b) makes it possible to clean
1172 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1173 such data would previously have always leaked in application code and
1174 workarounds were in place to make the memory debugging turn a blind eye
1175 to it. Application code that doesn't use this new function will still
1176 leak as before, but their memory debugging output will announce it now
1177 rather than letting it slide.
1179 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1180 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1181 has a return value to indicate success or failure.
1184 *) Make it possible to replace the underlying "ERR" functions such that the
1185 global state (2 LHASH tables and 2 locks) is only used by the "default"
1186 implementation. This change also adds two functions to "get" and "set"
1187 the implementation prior to it being automatically set the first time
1188 any other ERR function takes place. Ie. an application can call "get",
1189 pass the return value to a module it has just loaded, and that module
1190 can call its own "set" function using that value. This means the
1191 module's "ERR" operations will use (and modify) the error state in the
1192 application and not in its own statically linked copy of OpenSSL code.
1195 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1196 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1197 the operation, and provides a more encapsulated way for external code
1198 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1199 to use these functions rather than manually incrementing the counts.
1201 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1204 *) Add EVP test program.
1207 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1210 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1211 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1212 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1213 These allow a CRL to be built without having to access X509_CRL fields
1214 directly. Modify 'ca' application to use new functions.
1217 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1218 bug workarounds. Rollback attack detection is a security feature.
1219 The problem will only arise on OpenSSL servers when TLSv1 is not
1220 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1221 Software authors not wanting to support TLSv1 will have special reasons
1222 for their choice and can explicitly enable this option.
1223 [Bodo Moeller, Lutz Jaenicke]
1225 *) Rationalise EVP so it can be extended: don't include a union of
1226 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1227 (similar to those existing for EVP_CIPHER_CTX).
1232 EVP_MD_CTX_init(&md); /* new function call */
1233 EVP_DigestInit(&md, EVP_sha1());
1234 EVP_DigestUpdate(&md, in, len);
1235 EVP_DigestFinal(&md, out, NULL);
1236 EVP_MD_CTX_cleanup(&md); /* new function call */
1240 *) Make DES key schedule conform to the usual scheme, as well as
1241 correcting its structure. This means that calls to DES functions
1242 now have to pass a pointer to a des_key_schedule instead of a
1243 plain des_key_schedule (which was actually always a pointer
1246 des_key_schedule ks;
1248 des_set_key_checked(..., &ks);
1249 des_ncbc_encrypt(..., &ks, ...);
1251 (Note that a later change renames 'des_...' into 'DES_...'.)
1254 *) Initial reduction of linker bloat: the use of some functions, such as
1255 PEM causes large amounts of unused functions to be linked in due to
1256 poor organisation. For example pem_all.c contains every PEM function
1257 which has a knock on effect of linking in large amounts of (unused)
1258 ASN1 code. Grouping together similar functions and splitting unrelated
1259 functions prevents this.
1262 *) Cleanup of EVP macros.
1265 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1266 correct _ecb suffix.
1269 *) Add initial OCSP responder support to ocsp application. The
1270 revocation information is handled using the text based index
1271 use by the ca application. The responder can either handle
1272 requests generated internally, supplied in files (for example
1273 via a CGI script) or using an internal minimal server.
1276 *) Add configuration choices to get zlib compression for TLS.
1279 *) Changes to Kerberos SSL for RFC 2712 compliance:
1280 1. Implemented real KerberosWrapper, instead of just using
1281 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1282 2. Implemented optional authenticator field of KerberosWrapper.
1284 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1285 and authenticator structs; see crypto/krb5/.
1287 Generalized Kerberos calls to support multiple Kerberos libraries.
1288 [Vern Staats <staatsvr@asc.hpc.mil>,
1289 Jeffrey Altman <jaltman@columbia.edu>
1290 via Richard Levitte]
1292 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1293 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1294 values for each of the key sizes rather than having just
1295 parameters (and 'speed' generating keys each time).
1298 *) Speed up EVP routines.
1301 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1302 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1303 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1304 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1306 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1307 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1308 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1311 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1313 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1316 *) Added the OS2-EMX target.
1317 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1319 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1320 to support NCONF routines in extension code. New function CONF_set_nconf()
1321 to allow functions which take an NCONF to also handle the old LHASH
1322 structure: this means that the old CONF compatible routines can be
1323 retained (in particular wrt extensions) without having to duplicate the
1324 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1327 *) Enhance the general user interface with mechanisms for inner control
1328 and with possibilities to have yes/no kind of prompts.
1331 *) Change all calls to low level digest routines in the library and
1332 applications to use EVP. Add missing calls to HMAC_cleanup() and
1333 don't assume HMAC_CTX can be copied using memcpy().
1334 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1336 *) Add the possibility to control engines through control names but with
1337 arbitrary arguments instead of just a string.
1338 Change the key loaders to take a UI_METHOD instead of a callback
1339 function pointer. NOTE: this breaks binary compatibility with earlier
1340 versions of OpenSSL [engine].
1341 Adapt the nCipher code for these new conditions and add a card insertion
1345 *) Enhance the general user interface with mechanisms to better support
1346 dialog box interfaces, application-defined prompts, the possibility
1347 to use defaults (for example default passwords from somewhere else)
1348 and interrupts/cancellations.
1351 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1352 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1355 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1356 tidy up some unnecessarily weird code in 'sk_new()').
1357 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1359 *) Change the key loading routines for ENGINEs to use the same kind
1360 callback (pem_password_cb) as all other routines that need this
1364 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1365 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1366 than this minimum value is recommended.
1369 *) New random seeder for OpenVMS, using the system process statistics
1370 that are easily reachable.
1373 *) Windows apparently can't transparently handle global
1374 variables defined in DLLs. Initialisations such as:
1376 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1378 wont compile. This is used by the any applications that need to
1379 declare their own ASN1 modules. This was fixed by adding the option
1380 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1381 needed for static libraries under Win32.
1384 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1385 setting of purpose and trust fields. New X509_STORE trust and
1386 purpose functions and tidy up setting in other SSL functions.
1389 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1390 structure. These are inherited by X509_STORE_CTX when it is
1391 initialised. This allows various defaults to be set in the
1392 X509_STORE structure (such as flags for CRL checking and custom
1393 purpose or trust settings) for functions which only use X509_STORE_CTX
1394 internally such as S/MIME.
1396 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1397 trust settings if they are not set in X509_STORE. This allows X509_STORE
1398 purposes and trust (in S/MIME for example) to override any set by default.
1400 Add command line options for CRL checking to smime, s_client and s_server
1404 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1405 are set then the CRL is looked up in the X509_STORE structure and
1406 its validity and signature checked, then if the certificate is found
1407 in the CRL the verify fails with a revoked error.
1409 Various new CRL related callbacks added to X509_STORE_CTX structure.
1411 Command line options added to 'verify' application to support this.
1413 This needs some additional work, such as being able to handle multiple
1414 CRLs with different times, extension based lookup (rather than just
1415 by subject name) and ultimately more complete V2 CRL extension
1419 *) Add a general user interface API (crypto/ui/). This is designed
1420 to replace things like des_read_password and friends (backward
1421 compatibility functions using this new API are provided).
1422 The purpose is to remove prompting functions from the DES code
1423 section as well as provide for prompting through dialog boxes in
1424 a window system and the like.
1427 *) Add "ex_data" support to ENGINE so implementations can add state at a
1428 per-structure level rather than having to store it globally.
1431 *) Make it possible for ENGINE structures to be copied when retrieved by
1432 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1433 This causes the "original" ENGINE structure to act like a template,
1434 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1435 operational state can be localised to each ENGINE structure, despite the
1436 fact they all share the same "methods". New ENGINE structures returned in
1437 this case have no functional references and the return value is the single
1438 structural reference. This matches the single structural reference returned
1439 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1443 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1444 needs to match any other type at all we need to manually clear the
1448 *) Changes to the "openssl engine" utility to include;
1449 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1450 about an ENGINE's available control commands.
1451 - executing control commands from command line arguments using the
1452 '-pre' and '-post' switches. '-post' is only used if '-t' is
1453 specified and the ENGINE is successfully initialised. The syntax for
1454 the individual commands are colon-separated, for example;
1455 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1458 *) New dynamic control command support for ENGINEs. ENGINEs can now
1459 declare their own commands (numbers), names (strings), descriptions,
1460 and input types for run-time discovery by calling applications. A
1461 subset of these commands are implicitly classed as "executable"
1462 depending on their input type, and only these can be invoked through
1463 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1464 can be based on user input, config files, etc). The distinction is
1465 that "executable" commands cannot return anything other than a boolean
1466 result and can only support numeric or string input, whereas some
1467 discoverable commands may only be for direct use through
1468 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1469 pointers, or other custom uses. The "executable" commands are to
1470 support parameterisations of ENGINE behaviour that can be
1471 unambiguously defined by ENGINEs and used consistently across any
1472 OpenSSL-based application. Commands have been added to all the
1473 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1474 control over shared-library paths without source code alterations.
1477 *) Changed all ENGINE implementations to dynamically allocate their
1478 ENGINEs rather than declaring them statically. Apart from this being
1479 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1480 this also allows the implementations to compile without using the
1481 internal engine_int.h header.
1484 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1485 'const' value. Any code that should be able to modify a RAND_METHOD
1486 should already have non-const pointers to it (ie. they should only
1487 modify their own ones).
1490 *) Made a variety of little tweaks to the ENGINE code.
1491 - "atalla" and "ubsec" string definitions were moved from header files
1492 to C code. "nuron" string definitions were placed in variables
1493 rather than hard-coded - allowing parameterisation of these values
1494 later on via ctrl() commands.
1495 - Removed unused "#if 0"'d code.
1496 - Fixed engine list iteration code so it uses ENGINE_free() to release
1497 structural references.
1498 - Constified the RAND_METHOD element of ENGINE structures.
1499 - Constified various get/set functions as appropriate and added
1500 missing functions (including a catch-all ENGINE_cpy that duplicates
1501 all ENGINE values onto a new ENGINE except reference counts/state).
1502 - Removed NULL parameter checks in get/set functions. Setting a method
1503 or function to NULL is a way of cancelling out a previously set
1504 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1505 and doesn't justify the extra error symbols and code.
1506 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1507 flags from engine_int.h to engine.h.
1508 - Changed prototypes for ENGINE handler functions (init(), finish(),
1509 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1512 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1513 to the algorithm using long division. The binary algorithm can be
1514 used only if the modulus is odd. On 32-bit systems, it is faster
1515 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1516 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1517 up to 450 bits. In 64-bit environments, the binary algorithm
1518 appears to be advantageous for much longer moduli; here we use it
1519 for moduli up to 2048 bits.
1522 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1523 could not support the combine flag in choice fields.
1526 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1527 extensions from a certificate request to the certificate.
1530 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1531 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1532 file: this allows the display of the certificate about to be
1533 signed to be customised, to allow certain fields to be included
1534 or excluded and extension details. The old system didn't display
1535 multicharacter strings properly, omitted fields not in the policy
1536 and couldn't display additional details such as extensions.
1539 *) Function EC_POINTs_mul for multiple scalar multiplication
1540 of an arbitrary number of elliptic curve points
1541 \sum scalars[i]*points[i],
1542 optionally including the generator defined for the EC_GROUP:
1543 scalar*generator + \sum scalars[i]*points[i].
1545 EC_POINT_mul is a simple wrapper function for the typical case
1546 that the point list has just one item (besides the optional
1550 *) First EC_METHODs for curves over GF(p):
1552 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1553 operations and provides various method functions that can also
1554 operate with faster implementations of modular arithmetic.
1556 EC_GFp_mont_method() reuses most functions that are part of
1557 EC_GFp_simple_method, but uses Montgomery arithmetic.
1559 [Bodo Moeller; point addition and point doubling
1560 implementation directly derived from source code provided by
1561 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1563 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1564 crypto/ec/ec_lib.c):
1566 Curves are EC_GROUP objects (with an optional group generator)
1567 based on EC_METHODs that are built into the library.
1569 Points are EC_POINT objects based on EC_GROUP objects.
1571 Most of the framework would be able to handle curves over arbitrary
1572 finite fields, but as there are no obvious types for fields other
1573 than GF(p), some functions are limited to that for now.
1576 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1577 that the file contains a complete HTTP response.
1580 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1581 change the def and num file printf format specifier from "%-40sXXX"
1582 to "%-39s XXX". The latter will always guarantee a space after the
1583 field while the former will cause them to run together if the field
1584 is 40 of more characters long.
1587 *) Constify the cipher and digest 'method' functions and structures
1588 and modify related functions to take constant EVP_MD and EVP_CIPHER
1592 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1593 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1596 *) Modify EVP_Digest*() routines so they now return values. Although the
1597 internal software routines can never fail additional hardware versions
1601 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1603 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1604 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1607 ERR_R_NESTED_ASN1_ERROR
1609 ERR_R_MISSING_ASN1_EOS
1610 were 4 .. 9, conflicting with
1611 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1613 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1614 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1616 Add new error code 'ERR_R_INTERNAL_ERROR'.
1619 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1623 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1624 sets the subject name for a new request or supersedes the
1625 subject name in a given request. Formats that can be parsed are
1626 'CN=Some Name, OU=myOU, C=IT'
1628 'CN=Some Name/OU=myOU/C=IT'.
1630 Add options '-batch' and '-verbose' to 'openssl req'.
1631 [Massimiliano Pala <madwolf@hackmasters.net>]
1633 *) Introduce the possibility to access global variables through
1634 functions on platform were that's the best way to handle exporting
1635 global variables in shared libraries. To enable this functionality,
1636 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1637 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1638 is normally done by Configure or something similar).
1640 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1641 in the source file (foo.c) like this:
1643 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1644 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1646 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1647 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1649 OPENSSL_DECLARE_GLOBAL(int,foo);
1650 #define foo OPENSSL_GLOBAL_REF(foo)
1651 OPENSSL_DECLARE_GLOBAL(double,bar);
1652 #define bar OPENSSL_GLOBAL_REF(bar)
1654 The #defines are very important, and therefore so is including the
1655 header file everywhere where the defined globals are used.
1657 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1658 of ASN.1 items, but that structure is a bit different.
1660 The largest change is in util/mkdef.pl which has been enhanced with
1661 better and easier to understand logic to choose which symbols should
1662 go into the Windows .def files as well as a number of fixes and code
1663 cleanup (among others, algorithm keywords are now sorted
1664 lexicographically to avoid constant rewrites).
1667 *) In BN_div() keep a copy of the sign of 'num' before writing the
1668 result to 'rm' because if rm==num the value will be overwritten
1669 and produce the wrong result if 'num' is negative: this caused
1670 problems with BN_mod() and BN_nnmod().
1673 *) Function OCSP_request_verify(). This checks the signature on an
1674 OCSP request and verifies the signer certificate. The signer
1675 certificate is just checked for a generic purpose and OCSP request
1679 *) Add OCSP_check_validity() function to check the validity of OCSP
1680 responses. OCSP responses are prepared in real time and may only
1681 be a few seconds old. Simply checking that the current time lies
1682 between thisUpdate and nextUpdate max reject otherwise valid responses
1683 caused by either OCSP responder or client clock inaccuracy. Instead
1684 we allow thisUpdate and nextUpdate to fall within a certain period of
1685 the current time. The age of the response can also optionally be
1686 checked. Two new options -validity_period and -status_age added to
1690 *) If signature or public key algorithm is unrecognized print out its
1691 OID rather that just UNKNOWN.
1694 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1695 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1696 ID to be generated from the issuer certificate alone which can then be
1697 passed to OCSP_id_issuer_cmp().
1700 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1701 ASN1 modules to export functions returning ASN1_ITEM pointers
1702 instead of the ASN1_ITEM structures themselves. This adds several
1703 new macros which allow the underlying ASN1 function/structure to
1704 be accessed transparently. As a result code should not use ASN1_ITEM
1705 references directly (such as &X509_it) but instead use the relevant
1706 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1707 use of the new ASN1 code on platforms where exporting structures
1708 is problematical (for example in shared libraries) but exporting
1709 functions returning pointers to structures is not.
1712 *) Add support for overriding the generation of SSL/TLS session IDs.
1713 These callbacks can be registered either in an SSL_CTX or per SSL.
1714 The purpose of this is to allow applications to control, if they wish,
1715 the arbitrary values chosen for use as session IDs, particularly as it
1716 can be useful for session caching in multiple-server environments. A
1717 command-line switch for testing this (and any client code that wishes
1718 to use such a feature) has been added to "s_server".
1719 [Geoff Thorpe, Lutz Jaenicke]
1721 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1722 of the form '#if defined(...) || defined(...) || ...' and
1723 '#if !defined(...) && !defined(...) && ...'. This also avoids
1724 the growing number of special cases it was previously handling.
1727 *) Make all configuration macros available for application by making
1728 sure they are available in opensslconf.h, by giving them names starting
1729 with "OPENSSL_" to avoid conflicts with other packages and by making
1730 sure e_os2.h will cover all platform-specific cases together with
1732 Additionally, it is now possible to define configuration/platform-
1733 specific names (called "system identities"). In the C code, these
1734 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1735 macro with the name beginning with "OPENSSL_SYS_", which is determined
1736 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1740 *) New option -set_serial to 'req' and 'x509' this allows the serial
1741 number to use to be specified on the command line. Previously self
1742 signed certificates were hard coded with serial number 0 and the
1743 CA options of 'x509' had to use a serial number in a file which was
1747 *) New options to 'ca' utility to support V2 CRL entry extensions.
1748 Currently CRL reason, invalidity date and hold instruction are
1749 supported. Add new CRL extensions to V3 code and some new objects.
1752 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1753 disable standard block padding (aka PKCS#5 padding) in the EVP
1754 API, which was previously mandatory. This means that the data is
1755 not padded in any way and so the total length much be a multiple
1756 of the block size, otherwise an error occurs.
1759 *) Initial (incomplete) OCSP SSL support.
1762 *) New function OCSP_parse_url(). This splits up a URL into its host,
1763 port and path components: primarily to parse OCSP URLs. New -url
1764 option to ocsp utility.
1767 *) New nonce behavior. The return value of OCSP_check_nonce() now
1768 reflects the various checks performed. Applications can decide
1769 whether to tolerate certain situations such as an absent nonce
1770 in a response when one was present in a request: the ocsp application
1771 just prints out a warning. New function OCSP_add1_basic_nonce()
1772 this is to allow responders to include a nonce in a response even if
1773 the request is nonce-less.
1776 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1777 skipped when using openssl x509 multiple times on a single input file,
1778 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1781 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1782 set string type: to handle setting ASN1_TIME structures. Fix ca
1783 utility to correctly initialize revocation date of CRLs.
1786 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1787 the clients preferred ciphersuites and rather use its own preferences.
1788 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1789 Internet Explorer by ensuring unchanged hash method during stepup.
1790 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1793 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1794 to aes and add a new 'exist' option to print out symbols that don't
1798 *) Additional options to ocsp utility to allow flags to be set and
1799 additional certificates supplied.
1802 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1803 OCSP client a number of certificate to only verify the response
1807 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1808 handle the new API. Currently only ECB, CBC modes supported. Add new
1811 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1812 Encryption Standard (AES) Ciphersuites for Transport Layer
1813 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1814 not enabled by default and were not part of the "ALL" ciphersuite
1815 alias because they were not yet official; they could be
1816 explicitly requested by specifying the "AESdraft" ciphersuite
1817 group alias. In the final release of OpenSSL 0.9.7, the group
1818 alias is called "AES" and is part of "ALL".)
1819 [Ben Laurie, Steve Henson, Bodo Moeller]
1821 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1822 request to response.
1825 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1826 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1827 extract information from a certificate request. OCSP_response_create()
1828 creates a response and optionally adds a basic response structure.
1829 OCSP_basic_add1_status() adds a complete single response to a basic
1830 response and returns the OCSP_SINGLERESP structure just added (to allow
1831 extensions to be included for example). OCSP_basic_add1_cert() adds a
1832 certificate to a basic response and OCSP_basic_sign() signs a basic
1833 response with various flags. New helper functions ASN1_TIME_check()
1834 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1835 (converts ASN1_TIME to GeneralizedTime).
1838 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1839 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1840 structure from a certificate. X509_pubkey_digest() digests the public_key
1841 contents: this is used in various key identifiers.
1844 *) Make sk_sort() tolerate a NULL argument.
1845 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1847 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1848 passed by the function are trusted implicitly. If any of them signed the
1849 response then it is assumed to be valid and is not verified.
1852 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1853 to data. This was previously part of the PKCS7 ASN1 code. This
1854 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1855 [Steve Henson, reported by Kenneth R. Robinette
1856 <support@securenetterm.com>]
1858 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1859 routines: without these tracing memory leaks is very painful.
1860 Fix leaks in PKCS12 and PKCS7 routines.
1863 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1864 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1865 effectively meant GeneralizedTime would never be used. Now it
1866 is initialised to -1 but X509_time_adj() now has to check the value
1867 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1868 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1869 [Steve Henson, reported by Kenneth R. Robinette
1870 <support@securenetterm.com>]
1872 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1873 result in a zero length in the ASN1_INTEGER structure which was
1874 not consistent with the structure when d2i_ASN1_INTEGER() was used
1875 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1876 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1877 where it did not print out a minus for negative ASN1_INTEGER.
1880 *) Add summary printout to ocsp utility. The various functions which
1881 convert status values to strings have been renamed to:
1882 OCSP_response_status_str(), OCSP_cert_status_str() and
1883 OCSP_crl_reason_str() and are no longer static. New options
1884 to verify nonce values and to disable verification. OCSP response
1885 printout format cleaned up.
1888 *) Add additional OCSP certificate checks. These are those specified
1889 in RFC2560. This consists of two separate checks: the CA of the
1890 certificate being checked must either be the OCSP signer certificate
1891 or the issuer of the OCSP signer certificate. In the latter case the
1892 OCSP signer certificate must contain the OCSP signing extended key
1893 usage. This check is performed by attempting to match the OCSP
1894 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1895 in the OCSP_CERTID structures of the response.
1898 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1899 and related routines. This uses the standard OpenSSL certificate
1900 verify routines to perform initial checks (just CA validity) and
1901 to obtain the certificate chain. Then additional checks will be
1902 performed on the chain. Currently the root CA is checked to see
1903 if it is explicitly trusted for OCSP signing. This is used to set
1904 a root CA as a global signing root: that is any certificate that
1905 chains to that CA is an acceptable OCSP signing certificate.
1908 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1909 extensions from a separate configuration file.
1910 As when reading extensions from the main configuration file,
1911 the '-extensions ...' option may be used for specifying the
1913 [Massimiliano Pala <madwolf@comune.modena.it>]
1915 *) New OCSP utility. Allows OCSP requests to be generated or
1916 read. The request can be sent to a responder and the output
1917 parsed, outputed or printed in text form. Not complete yet:
1918 still needs to check the OCSP response validity.
1921 *) New subcommands for 'openssl ca':
1922 'openssl ca -status <serial>' prints the status of the cert with
1923 the given serial number (according to the index file).
1924 'openssl ca -updatedb' updates the expiry status of certificates
1926 [Massimiliano Pala <madwolf@comune.modena.it>]
1928 *) New '-newreq-nodes' command option to CA.pl. This is like
1929 '-newreq', but calls 'openssl req' with the '-nodes' option
1930 so that the resulting key is not encrypted.
1931 [Damien Miller <djm@mindrot.org>]
1933 *) New configuration for the GNU Hurd.
1934 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1936 *) Initial code to implement OCSP basic response verify. This
1937 is currently incomplete. Currently just finds the signer's
1938 certificate and verifies the signature on the response.
1941 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1942 value of OPENSSLDIR. This is available via the new '-d' option
1943 to 'openssl version', and is also included in 'openssl version -a'.
1946 *) Allowing defining memory allocation callbacks that will be given
1947 file name and line number information in additional arguments
1948 (a const char* and an int). The basic functionality remains, as
1949 well as the original possibility to just replace malloc(),
1950 realloc() and free() by functions that do not know about these
1951 additional arguments. To register and find out the current
1952 settings for extended allocation functions, the following
1953 functions are provided:
1955 CRYPTO_set_mem_ex_functions
1956 CRYPTO_set_locked_mem_ex_functions
1957 CRYPTO_get_mem_ex_functions
1958 CRYPTO_get_locked_mem_ex_functions
1960 These work the same way as CRYPTO_set_mem_functions and friends.
1961 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1962 extended allocation function is enabled.
1963 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1964 a conventional allocation function is enabled.
1965 [Richard Levitte, Bodo Moeller]
1967 *) Finish off removing the remaining LHASH function pointer casts.
1968 There should no longer be any prototype-casting required when using
1969 the LHASH abstraction, and any casts that remain are "bugs". See
1970 the callback types and macros at the head of lhash.h for details
1971 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1974 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1975 If /dev/[u]random devices are not available or do not return enough
1976 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1978 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1979 /etc/entropy will be queried once each in this sequence, quering stops
1980 when enough entropy was collected without querying more sockets.
1983 *) Change the Unix RAND_poll() variant to be able to poll several
1984 random devices, as specified by DEVRANDOM, until a sufficient amount
1985 of data has been collected. We spend at most 10 ms on each file
1986 (select timeout) and read in non-blocking mode. DEVRANDOM now
1987 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1988 (previously it was just the string "/dev/urandom"), so on typical
1989 platforms the 10 ms delay will never occur.
1990 Also separate out the Unix variant to its own file, rand_unix.c.
1991 For VMS, there's a currently-empty rand_vms.c.
1994 *) Move OCSP client related routines to ocsp_cl.c. These
1995 provide utility functions which an application needing
1996 to issue a request to an OCSP responder and analyse the
1997 response will typically need: as opposed to those which an
1998 OCSP responder itself would need which will be added later.
2000 OCSP_request_sign() signs an OCSP request with an API similar
2001 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2002 response. OCSP_response_get1_basic() extracts basic response
2003 from response. OCSP_resp_find_status(): finds and extracts status
2004 information from an OCSP_CERTID structure (which will be created
2005 when the request structure is built). These are built from lower
2006 level functions which work on OCSP_SINGLERESP structures but
2007 wont normally be used unless the application wishes to examine
2008 extensions in the OCSP response for example.
2010 Replace nonce routines with a pair of functions.
2011 OCSP_request_add1_nonce() adds a nonce value and optionally
2012 generates a random value. OCSP_check_nonce() checks the
2013 validity of the nonce in an OCSP response.
2016 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2017 This doesn't copy the supplied OCSP_CERTID and avoids the
2018 need to free up the newly created id. Change return type
2019 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2020 This can then be used to add extensions to the request.
2021 Deleted OCSP_request_new(), since most of its functionality
2022 is now in OCSP_REQUEST_new() (and the case insensitive name
2023 clash) apart from the ability to set the request name which
2024 will be added elsewhere.
2027 *) Update OCSP API. Remove obsolete extensions argument from
2028 various functions. Extensions are now handled using the new
2029 OCSP extension code. New simple OCSP HTTP function which
2030 can be used to send requests and parse the response.
2033 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2034 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2035 uses the special reorder version of SET OF to sort the attributes
2036 and reorder them to match the encoded order. This resolves a long
2037 standing problem: a verify on a PKCS7 structure just after signing
2038 it used to fail because the attribute order did not match the
2039 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2040 it uses the received order. This is necessary to tolerate some broken
2041 software that does not order SET OF. This is handled by encoding
2042 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2043 to produce the required SET OF.
2046 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2047 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2048 files to get correct declarations of the ASN.1 item variables.
2051 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2052 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2053 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2054 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2055 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2056 ASN1_ITEM and no wrapper functions.
2059 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2060 replace the old function pointer based I/O routines. Change most of
2061 the *_d2i_bio() and *_d2i_fp() functions to use these.
2064 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2065 lines, recognice more "algorithms" that can be deselected, and make
2066 it complain about algorithm deselection that isn't recognised.
2069 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2070 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2071 to use new functions. Add NO_ASN1_OLD which can be set to remove
2072 some old style ASN1 functions: this can be used to determine if old
2073 code will still work when these eventually go away.
2076 *) New extension functions for OCSP structures, these follow the
2077 same conventions as certificates and CRLs.
2080 *) New function X509V3_add1_i2d(). This automatically encodes and
2081 adds an extension. Its behaviour can be customised with various
2082 flags to append, replace or delete. Various wrappers added for
2083 certifcates and CRLs.
2086 *) Fix to avoid calling the underlying ASN1 print routine when
2087 an extension cannot be parsed. Correct a typo in the
2088 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2091 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2092 entries for variables.
2095 *) Add functionality to apps/openssl.c for detecting locking
2096 problems: As the program is single-threaded, all we have
2097 to do is register a locking callback using an array for
2098 storing which locks are currently held by the program.
2101 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2102 SSL_get_ex_data_X509_STORE_idx(), which is used in
2103 ssl_verify_cert_chain() and thus can be called at any time
2104 during TLS/SSL handshakes so that thread-safety is essential.
2105 Unfortunately, the ex_data design is not at all suited
2106 for multi-threaded use, so it probably should be abolished.
2109 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2110 [Broadcom, tweaked and integrated by Geoff Thorpe]
2112 *) Move common extension printing code to new function
2113 X509V3_print_extensions(). Reorganise OCSP print routines and
2114 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2117 *) New function X509_signature_print() to remove duplication in some
2121 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2122 set (this was treated exactly the same as SET OF previously). This
2123 is used to reorder the STACK representing the structure to match the
2124 encoding. This will be used to get round a problem where a PKCS7
2125 structure which was signed could not be verified because the STACK
2126 order did not reflect the encoded order.
2129 *) Reimplement the OCSP ASN1 module using the new code.
2132 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2133 for its ASN1 operations. The old style function pointers still exist
2134 for now but they will eventually go away.
2137 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2138 completely replaces the old ASN1 functionality with a table driven
2139 encoder and decoder which interprets an ASN1_ITEM structure describing
2140 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2141 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2142 has also been converted to the new form.
2145 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2146 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2147 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2148 for negative moduli.
2151 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2152 of not touching the result's sign bit.
2155 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2159 *) Changed the LHASH code to use prototypes for callbacks, and created
2160 macros to declare and implement thin (optionally static) functions
2161 that provide type-safety and avoid function pointer casting for the
2162 type-specific callbacks.
2165 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2167 [Veers Staats <staatsvr@asc.hpc.mil>,
2168 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2170 *) Reformat the FAQ so the different questions and answers can be divided
2171 in sections depending on the subject.
2174 *) Have the zlib compression code load ZLIB.DLL dynamically under
2178 *) New function BN_mod_sqrt for computing square roots modulo a prime
2179 (using the probabilistic Tonelli-Shanks algorithm unless
2180 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2181 be handled deterministically).
2182 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2184 *) Make BN_mod_inverse faster by explicitly handling small quotients
2185 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2186 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2189 *) New function BN_kronecker.
2192 *) Fix BN_gcd so that it works on negative inputs; the result is
2193 positive unless both parameters are zero.
2194 Previously something reasonably close to an infinite loop was
2195 possible because numbers could be growing instead of shrinking
2196 in the implementation of Euclid's algorithm.
2199 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2200 sign of the number in question.
2202 Fix BN_is_word(a,w) to work correctly for w == 0.
2204 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2205 because its test if the absolute value of 'a' equals 'w'.
2206 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2207 it exists mostly for use in the implementations of BN_is_zero(),
2208 BN_is_one(), and BN_is_word().
2211 *) New function BN_swap.
2214 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2215 the exponentiation functions are more likely to produce reasonable
2216 results on negative inputs.
2219 *) Change BN_mod_mul so that the result is always non-negative.
2220 Previously, it could be negative if one of the factors was negative;
2221 I don't think anyone really wanted that behaviour.
2224 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2225 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2226 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2227 and add new functions:
2236 BN_mod_lshift1_quick
2240 These functions always generate non-negative results.
2242 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2243 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2245 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2246 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2247 be reduced modulo m.
2248 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2251 The following entry accidentily appeared in the CHANGES file
2252 distributed with OpenSSL 0.9.7. The modifications described in
2253 it do *not* apply to OpenSSL 0.9.7.
2255 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2256 was actually never needed) and in BN_mul(). The removal in BN_mul()
2257 required a small change in bn_mul_part_recursive() and the addition
2258 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2259 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2260 bn_sub_words() and bn_add_words() except they take arrays with
2265 *) In 'openssl passwd', verify passwords read from the terminal
2266 unless the '-salt' option is used (which usually means that
2267 verification would just waste user's time since the resulting
2268 hash is going to be compared with some given password hash)
2269 or the new '-noverify' option is used.
2271 This is an incompatible change, but it does not affect
2272 non-interactive use of 'openssl passwd' (passwords on the command
2273 line, '-stdin' option, '-in ...' option) and thus should not
2277 *) Remove all references to RSAref, since there's no more need for it.
2280 *) Make DSO load along a path given through an environment variable
2281 (SHLIB_PATH) with shl_load().
2284 *) Constify the ENGINE code as a result of BIGNUM constification.
2285 Also constify the RSA code and most things related to it. In a
2286 few places, most notable in the depth of the ASN.1 code, ugly
2287 casts back to non-const were required (to be solved at a later
2291 *) Make it so the openssl application has all engines loaded by default.
2294 *) Constify the BIGNUM routines a little more.
2297 *) Add the following functions:
2299 ENGINE_load_cswift()
2301 ENGINE_load_atalla()
2303 ENGINE_load_builtin_engines()
2305 That way, an application can itself choose if external engines that
2306 are built-in in OpenSSL shall ever be used or not. The benefit is
2307 that applications won't have to be linked with libdl or other dso
2308 libraries unless it's really needed.
2310 Changed 'openssl engine' to load all engines on demand.
2311 Changed the engine header files to avoid the duplication of some
2312 declarations (they differed!).
2315 *) 'openssl engine' can now list capabilities.
2318 *) Better error reporting in 'openssl engine'.
2321 *) Never call load_dh_param(NULL) in s_server.
2324 *) Add engine application. It can currently list engines by name and
2325 identity, and test if they are actually available.
2328 *) Improve RPM specification file by forcing symbolic linking and making
2329 sure the installed documentation is also owned by root.root.
2330 [Damien Miller <djm@mindrot.org>]
2332 *) Give the OpenSSL applications more possibilities to make use of
2333 keys (public as well as private) handled by engines.
2336 *) Add OCSP code that comes from CertCo.
2339 *) Add VMS support for the Rijndael code.
2342 *) Added untested support for Nuron crypto accelerator.
2345 *) Add support for external cryptographic devices. This code was
2346 previously distributed separately as the "engine" branch.
2347 [Geoff Thorpe, Richard Levitte]
2349 *) Rework the filename-translation in the DSO code. It is now possible to
2350 have far greater control over how a "name" is turned into a filename
2351 depending on the operating environment and any oddities about the
2352 different shared library filenames on each system.
2355 *) Support threads on FreeBSD-elf in Configure.
2358 *) Fix for SHA1 assembly problem with MASM: it produces
2359 warnings about corrupt line number information when assembling
2360 with debugging information. This is caused by the overlapping
2362 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2365 NCONF_get_number() has no error checking at all. As a replacement,
2366 NCONF_get_number_e() is defined (_e for "error checking") and is
2367 promoted strongly. The old NCONF_get_number is kept around for
2368 binary backward compatibility.
2369 Make it possible for methods to load from something other than a BIO,
2370 by providing a function pointer that is given a name instead of a BIO.
2371 For example, this could be used to load configuration data from an
2375 *) Fix for non blocking accept BIOs. Added new I/O special reason
2376 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2377 with non blocking I/O was not possible because no retry code was
2378 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2382 *) Added the beginnings of Rijndael support.
2385 *) Fix for bug in DirectoryString mask setting. Add support for
2386 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2387 to allow certificate printing to more controllable, additional
2388 'certopt' option to 'x509' to allow new printing options to be
2392 *) Clean old EAY MD5 hack from e_os.h.
2395 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2397 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2398 via timing by performing a MAC computation even if incorrrect
2399 block cipher padding has been found. This is a countermeasure
2400 against active attacks where the attacker has to distinguish
2401 between bad padding and a MAC verification error. (CAN-2003-0078)
2403 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2404 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2405 Martin Vuagnoux (EPFL, Ilion)]
2407 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2409 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2410 memory from it's contents. This is done with a counter that will
2411 place alternating values in each byte. This can be used to solve
2412 two issues: 1) the removal of calls to memset() by highly optimizing
2413 compilers, and 2) cleansing with other values than 0, since those can
2414 be read through on certain media, for example a swap space on disk.
2417 *) Bugfix: client side session caching did not work with external caching,
2418 because the session->cipher setting was not restored when reloading
2419 from the external cache. This problem was masked, when
2420 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2421 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2424 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2425 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2426 [Zeev Lieber <zeev-l@yahoo.com>]
2428 *) Undo an undocumented change introduced in 0.9.6e which caused
2429 repeated calls to OpenSSL_add_all_ciphers() and
2430 OpenSSL_add_all_digests() to be ignored, even after calling
2434 *) Change the default configuration reader to deal with last line not
2435 being properly terminated.
2438 *) Change X509_NAME_cmp() so it applies the special rules on handling
2439 DN values that are of type PrintableString, as well as RDNs of type
2440 emailAddress where the value has the type ia5String.
2441 [stefank@valicert.com via Richard Levitte]
2443 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2444 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2445 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2446 the bitwise-OR of the two for use by the majority of applications
2447 wanting this behaviour, and update the docs. The documented
2448 behaviour and actual behaviour were inconsistent and had been
2449 changing anyway, so this is more a bug-fix than a behavioural
2451 [Geoff Thorpe, diagnosed by Nadav Har'El]
2453 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2454 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2457 *) Fix initialization code race conditions in
2458 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2459 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2460 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2461 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2462 ssl2_get_cipher_by_char(),
2463 ssl3_get_cipher_by_char().
2464 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2466 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2467 the cached sessions are flushed, as the remove_cb() might use ex_data
2468 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2469 (see [openssl.org #212]).
2470 [Geoff Thorpe, Lutz Jaenicke]
2472 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2473 length, instead of the encoding length to d2i_ASN1_OBJECT.
2476 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2478 *) [In 0.9.6g-engine release:]
2479 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2480 [Lynn Gazis <lgazis@rainbow.com>]
2482 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2484 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2485 and get fix the header length calculation.
2486 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2487 Alon Kantor <alonk@checkpoint.com> (and others),
2490 *) Use proper error handling instead of 'assertions' in buffer
2491 overflow checks added in 0.9.6e. This prevents DoS (the
2492 assertions could call abort()).
2493 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2495 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2497 *) Add various sanity checks to asn1_get_length() to reject
2498 the ASN1 length bytes if they exceed sizeof(long), will appear
2499 negative or the content length exceeds the length of the
2501 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2503 *) Fix cipher selection routines: ciphers without encryption had no flags
2504 for the cipher strength set and where therefore not handled correctly
2505 by the selection routines (PR #130).
2508 *) Fix EVP_dsa_sha macro.
2512 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2513 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2514 that was added in OpenSSL 0.9.6d.
2516 As the countermeasure turned out to be incompatible with some
2517 broken SSL implementations, the new option is part of SSL_OP_ALL.
2518 SSL_OP_ALL is usually employed when compatibility with weird SSL
2519 implementations is desired (e.g. '-bugs' option to 's_client' and
2520 's_server'), so the new option is automatically set in many
2524 *) Changes in security patch:
2526 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2527 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2528 Air Force Materiel Command, USAF, under agreement number
2531 *) Add various sanity checks to asn1_get_length() to reject
2532 the ASN1 length bytes if they exceed sizeof(long), will appear
2533 negative or the content length exceeds the length of the
2534 supplied buffer. (CAN-2002-0659)
2535 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2537 *) Assertions for various potential buffer overflows, not known to
2539 [Ben Laurie (CHATS)]
2541 *) Various temporary buffers to hold ASCII versions of integers were
2542 too small for 64 bit platforms. (CAN-2002-0655)
2543 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2545 *) Remote buffer overflow in SSL3 protocol - an attacker could
2546 supply an oversized session ID to a client. (CAN-2002-0656)
2547 [Ben Laurie (CHATS)]
2549 *) Remote buffer overflow in SSL2 protocol - an attacker could
2550 supply an oversized client master key. (CAN-2002-0656)
2551 [Ben Laurie (CHATS)]
2553 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2555 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2556 encoded as NULL) with id-dsa-with-sha1.
2557 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2559 *) Check various X509_...() return values in apps/req.c.
2560 [Nils Larsch <nla@trustcenter.de>]
2562 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2563 an end-of-file condition would erronously be flagged, when the CRLF
2564 was just at the end of a processed block. The bug was discovered when
2565 processing data through a buffering memory BIO handing the data to a
2566 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2567 <ptsekov@syntrex.com> and Nedelcho Stanev.
2570 *) Implement a countermeasure against a vulnerability recently found
2571 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2572 before application data chunks to avoid the use of known IVs
2573 with data potentially chosen by the attacker.
2576 *) Fix length checks in ssl3_get_client_hello().
2579 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2580 to prevent ssl3_read_internal() from incorrectly assuming that
2581 ssl3_read_bytes() found application data while handshake
2582 processing was enabled when in fact s->s3->in_read_app_data was
2583 merely automatically cleared during the initial handshake.
2584 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2586 *) Fix object definitions for Private and Enterprise: they were not
2587 recognized in their shortname (=lowercase) representation. Extend
2588 obj_dat.pl to issue an error when using undefined keywords instead
2589 of silently ignoring the problem (Svenning Sorensen
2590 <sss@sss.dnsalias.net>).
2593 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2594 generators, i.e. generators other than 2 and 5. (Previously, the
2595 code did not properly initialise the 'add' and 'rem' values to
2596 BN_generate_prime().)
2598 In the new general case, we do not insist that 'generator' is
2599 actually a primitive root: This requirement is rather pointless;
2600 a generator of the order-q subgroup is just as good, if not
2604 *) Map new X509 verification errors to alerts. Discovered and submitted by
2605 Tom Wu <tom@arcot.com>.
2608 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2609 returning non-zero before the data has been completely received
2610 when using non-blocking I/O.
2611 [Bodo Moeller; problem pointed out by John Hughes]
2613 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2614 [Ben Laurie, Lutz Jaenicke]
2616 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2617 Yoram Zahavi <YoramZ@gilian.com>).
2620 *) Add information about CygWin 1.3 and on, and preserve proper
2621 configuration for the versions before that.
2622 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2624 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2625 check whether we deal with a copy of a session and do not delete from
2626 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2627 <izhar@checkpoint.com>.
2630 *) Do not store session data into the internal session cache, if it
2631 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2632 flag is set). Proposed by Aslam <aslam@funk.com>.
2635 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2639 *) [In 0.9.6d-engine release:]
2640 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2641 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2643 *) Add the configuration target linux-s390x.
2644 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2646 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2647 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2648 variable as an indication that a ClientHello message has been
2649 received. As the flag value will be lost between multiple
2650 invocations of ssl3_accept when using non-blocking I/O, the
2651 function may not be aware that a handshake has actually taken
2652 place, thus preventing a new session from being added to the
2655 To avoid this problem, we now set s->new_session to 2 instead of
2656 using a local variable.
2657 [Lutz Jaenicke, Bodo Moeller]
2659 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2660 if the SSL_R_LENGTH_MISMATCH error is detected.
2661 [Geoff Thorpe, Bodo Moeller]
2663 *) New 'shared_ldflag' column in Configure platform table.
2666 *) Fix EVP_CIPHER_mode macro.
2667 ["Dan S. Camper" <dan@bti.net>]
2669 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2670 type, we must throw them away by setting rr->length to 0.
2671 [D P Chang <dpc@qualys.com>]
2673 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2675 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2676 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2677 worked incorrectly for those cases where range = 10..._2 and
2678 3*range is two bits longer than range.)
2681 *) Only add signing time to PKCS7 structures if it is not already
2685 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2686 OBJ_ld_ce should be OBJ_id_ce.
2687 Also some ip-pda OIDs in crypto/objects/objects.txt were
2688 incorrect (cf. RFC 3039).
2689 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2691 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2692 returns early because it has nothing to do.
2693 [Andy Schneider <andy.schneider@bjss.co.uk>]
2695 *) [In 0.9.6c-engine release:]
2696 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2697 [Andy Schneider <andy.schneider@bjss.co.uk>]
2699 *) [In 0.9.6c-engine release:]
2700 Add support for Cryptographic Appliance's keyserver technology.
2701 (Use engine 'keyclient')
2702 [Cryptographic Appliances and Geoff Thorpe]
2704 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2705 is called via tools/c89.sh because arguments have to be
2706 rearranged (all '-L' options must appear before the first object
2708 [Richard Shapiro <rshapiro@abinitio.com>]
2710 *) [In 0.9.6c-engine release:]
2711 Add support for Broadcom crypto accelerator cards, backported
2713 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2715 *) [In 0.9.6c-engine release:]
2716 Add support for SureWare crypto accelerator cards from
2717 Baltimore Technologies. (Use engine 'sureware')
2718 [Baltimore Technologies and Mark Cox]
2720 *) [In 0.9.6c-engine release:]
2721 Add support for crypto accelerator cards from Accelerated
2722 Encryption Processing, www.aep.ie. (Use engine 'aep')
2723 [AEP Inc. and Mark Cox]
2725 *) Add a configuration entry for gcc on UnixWare.
2726 [Gary Benson <gbenson@redhat.com>]
2728 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2729 messages are stored in a single piece (fixed-length part and
2730 variable-length part combined) and fix various bugs found on the way.
2733 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2734 instead. BIO_gethostbyname() does not know what timeouts are
2735 appropriate, so entries would stay in cache even when they have
2737 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2739 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2740 faced with a pathologically small ClientHello fragment that does
2741 not contain client_version: Instead of aborting with an error,
2742 simply choose the highest available protocol version (i.e.,
2743 TLS 1.0 unless it is disabled). In practice, ClientHello
2744 messages are never sent like this, but this change gives us
2745 strictly correct behaviour at least for TLS.
2748 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2749 never resets s->method to s->ctx->method when called from within
2750 one of the SSL handshake functions.
2751 [Bodo Moeller; problem pointed out by Niko Baric]
2753 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2754 (sent using the client's version number) if client_version is
2755 smaller than the protocol version in use. Also change
2756 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2757 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2758 the client will at least see that alert.
2761 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2765 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2766 client receives HelloRequest while in a handshake.
2767 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2769 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2770 should end in 'break', not 'goto end' which circuments various
2771 cleanups done in state SSL_ST_OK. But session related stuff
2772 must be disabled for SSL_ST_OK in the case that we just sent a
2775 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2776 before just sending a HelloRequest.
2777 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2779 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2780 reveal whether illegal block cipher padding was found or a MAC
2781 verification error occured. (Neither SSLerr() codes nor alerts
2782 are directly visible to potential attackers, but the information
2783 may leak via logfiles.)
2785 Similar changes are not required for the SSL 2.0 implementation
2786 because the number of padding bytes is sent in clear for SSL 2.0,
2787 and the extra bytes are just ignored. However ssl/s2_pkt.c
2788 failed to verify that the purported number of padding bytes is in
2792 *) Add OpenUNIX-8 support including shared libraries
2793 (Boyd Lynn Gerber <gerberb@zenez.com>).
2796 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2797 'wristwatch attack' using huge encoding parameters (cf.
2798 James H. Manger's CRYPTO 2001 paper). Note that the
2799 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2800 encoding parameters and hence was not vulnerable.
2803 *) BN_sqr() bug fix.
2804 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2806 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2807 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2808 followed by modular reduction.
2809 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2811 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2812 equivalent based on BN_pseudo_rand() instead of BN_rand().
2815 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2816 This function was broken, as the check for a new client hello message
2817 to handle SGC did not allow these large messages.
2818 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2821 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2824 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2825 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2828 *) Rework the configuration and shared library support for Tru64 Unix.
2829 The configuration part makes use of modern compiler features and
2830 still retains old compiler behavior for those that run older versions
2831 of the OS. The shared library support part includes a variant that
2832 uses the RPATH feature, and is available through the special
2833 configuration target "alpha-cc-rpath", which will never be selected
2835 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2837 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2838 with the same message size as in ssl3_get_certificate_request().
2839 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2840 messages might inadvertently be reject as too long.
2841 [Petr Lampa <lampa@fee.vutbr.cz>]
2843 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2846 *) Modified SSL library such that the verify_callback that has been set
2847 specificly for an SSL object with SSL_set_verify() is actually being
2848 used. Before the change, a verify_callback set with this function was
2849 ignored and the verify_callback() set in the SSL_CTX at the time of
2850 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2851 to allow the necessary settings.
2854 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2855 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2856 done automatically (in contradiction to the requirements of the C
2857 standard). This made problems when used from OpenSSH.
2860 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2861 dh->length and always used
2863 BN_rand_range(priv_key, dh->p).
2865 BN_rand_range() is not necessary for Diffie-Hellman, and this
2866 specific range makes Diffie-Hellman unnecessarily inefficient if
2867 dh->length (recommended exponent length) is much smaller than the
2868 length of dh->p. We could use BN_rand_range() if the order of
2869 the subgroup was stored in the DH structure, but we only have
2874 BN_rand(priv_key, l, ...)
2876 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2882 RSA_eay_public_encrypt
2883 RSA_eay_private_decrypt
2884 RSA_eay_private_encrypt (signing)
2885 RSA_eay_public_decrypt (signature verification)
2887 (default implementations for RSA_public_encrypt,
2888 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2889 always reject numbers >= n.
2892 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2893 to synchronize access to 'locking_thread'. This is necessary on
2894 systems where access to 'locking_thread' (an 'unsigned long'
2895 variable) is not atomic.
2898 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2899 *before* setting the 'crypto_lock_rand' flag. The previous code had
2900 a race condition if 0 is a valid thread ID.
2901 [Travis Vitek <vitek@roguewave.com>]
2903 *) Add support for shared libraries under Irix.
2904 [Albert Chin-A-Young <china@thewrittenword.com>]
2906 *) Add configuration option to build on Linux on both big-endian and
2908 [Ralf Baechle <ralf@uni-koblenz.de>]
2910 *) Add the possibility to create shared libraries on HP-UX.
2913 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2915 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2916 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2917 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2918 PRNG state recovery was possible based on the output of
2919 one PRNG request appropriately sized to gain knowledge on
2920 'md' followed by enough consecutive 1-byte PRNG requests
2921 to traverse all of 'state'.
2923 1. When updating 'md_local' (the current thread's copy of 'md')
2924 during PRNG output generation, hash all of the previous
2925 'md_local' value, not just the half used for PRNG output.
2927 2. Make the number of bytes from 'state' included into the hash
2928 independent from the number of PRNG bytes requested.
2930 The first measure alone would be sufficient to avoid
2931 Markku-Juhani's attack. (Actually it had never occurred
2932 to me that the half of 'md_local' used for chaining was the
2933 half from which PRNG output bytes were taken -- I had always
2934 assumed that the secret half would be used.) The second
2935 measure makes sure that additional data from 'state' is never
2936 mixed into 'md_local' in small portions; this heuristically
2937 further strengthens the PRNG.
2940 *) Fix crypto/bn/asm/mips3.s.
2943 *) When only the key is given to "enc", the IV is undefined. Print out
2944 an error message in this case.
2947 *) Handle special case when X509_NAME is empty in X509 printing routines.
2950 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2951 positive and less than q.
2954 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2955 used: it isn't thread safe and the add_lock_callback should handle
2957 [Paul Rose <Paul.Rose@bridge.com>]
2959 *) Verify that incoming data obeys the block size in
2960 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2964 [Ulf Möller, Bodo Möller]
2966 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2967 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2968 when fixing the server behaviour for backwards-compatible 'client
2969 hello' messages. (Note that the attack is impractical against
2970 SSL 3.0 and TLS 1.0 anyway because length and version checking
2971 means that the probability of guessing a valid ciphertext is
2972 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2975 Before 0.9.5, the countermeasure (hide the error by generating a
2976 random 'decryption result') did not work properly because
2977 ERR_clear_error() was missing, meaning that SSL_get_error() would
2978 detect the supposedly ignored error.
2980 Both problems are now fixed.
2983 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2984 (previously it was 1024).
2987 *) Fix for compatibility mode trust settings: ignore trust settings
2988 unless some valid trust or reject settings are present.
2991 *) Fix for blowfish EVP: its a variable length cipher.
2994 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2995 parameters in DSA public key structures and return an error in the
2996 DSA routines if parameters are absent.
2999 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3000 in the current directory if neither $RANDFILE nor $HOME was set.
3001 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3002 caused some confusion to Windows users who haven't defined $HOME.
3003 Thus RAND_file_name() is changed again: e_os.h can define a
3004 DEFAULT_HOME, which will be used if $HOME is not set.
3005 For Windows, we use "C:"; on other platforms, we still require
3006 environment variables.
3008 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3009 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3010 having multiple threads call RAND_poll() concurrently.
3013 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3014 combination of a flag and a thread ID variable.
3015 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3016 flag), *other* threads can enter ssleay_add_bytes without obeying
3017 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3018 that they do not hold after the first thread unsets add_do_not_lock).
3021 *) Change bctest again: '-x' expressions are not available in all
3025 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3027 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3028 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3030 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3031 the default extension for executables, if any. Also, make the perl
3032 scripts that use symlink() to test if it really exists and use "cp"
3033 if it doesn't. All this made OpenSSL compilable and installable in
3037 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3038 If SEQUENCE is length is indefinite just set c->slen to the total
3039 amount of data available.
3040 [Steve Henson, reported by shige@FreeBSD.org]
3041 [This change does not apply to 0.9.7.]
3043 *) Change bctest to avoid here-documents inside command substitution
3044 (workaround for FreeBSD /bin/sh bug).
3045 For compatibility with Ultrix, avoid shell functions (introduced
3046 in the bctest version that searches along $PATH).
3049 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3050 with des_encrypt() defined on some operating systems, like Solaris
3054 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3055 On the Importance of Eliminating Errors in Cryptographic
3056 Computations, J. Cryptology 14 (2001) 2, 101-119,
3057 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3060 *) MIPS assembler BIGNUM division bug fix.
3063 *) Disabled incorrect Alpha assembler code.
3066 *) Fix PKCS#7 decode routines so they correctly update the length
3067 after reading an EOC for the EXPLICIT tag.
3069 [This change does not apply to 0.9.7.]
3071 *) Fix bug in PKCS#12 key generation routines. This was triggered
3072 if a 3DES key was generated with a 0 initial byte. Include
3073 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3074 (but broken) behaviour.
3077 *) Enhance bctest to search for a working bc along $PATH and print
3079 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3081 *) Fix memory leaks in err.c: free err_data string if necessary;
3082 don't write to the wrong index in ERR_set_error_data.
3085 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3089 *) Replace rdtsc with _emit statements for VC++ version 5.
3090 [Jeremy Cooper <jeremy@baymoo.org>]
3092 *) Make it possible to reuse SSLv2 sessions.
3095 *) In copy_email() check for >= 0 as a return value for
3096 X509_NAME_get_index_by_NID() since 0 is a valid index.
3097 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3099 *) Avoid coredump with unsupported or invalid public keys by checking if
3100 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3101 PKCS7_verify() fails with non detached data.
3104 *) Don't use getenv in library functions when run as setuid/setgid.
3105 New function OPENSSL_issetugid().
3108 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3109 due to incorrect handling of multi-threading:
3111 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3113 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3115 3. Count how many times MemCheck_off() has been called so that
3116 nested use can be treated correctly. This also avoids
3117 inband-signalling in the previous code (which relied on the
3118 assumption that thread ID 0 is impossible).
3121 *) Add "-rand" option also to s_client and s_server.
3124 *) Fix CPU detection on Irix 6.x.
3125 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3126 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3128 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3131 [This change does not apply to 0.9.7.]
3133 *) Use the cached encoding of an X509_NAME structure rather than
3134 copying it. This is apparently the reason for the libsafe "errors"
3135 but the code is actually correct.
3138 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3139 Bleichenbacher's DSA attack.
3140 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3141 to be set and top=0 forces the highest bit to be set; top=-1 is new
3142 and leaves the highest bit random.
3143 [Ulf Moeller, Bodo Moeller]
3145 *) In the NCONF_...-based implementations for CONF_... queries
3146 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3147 a temporary CONF structure with the data component set to NULL
3148 (which gives segmentation faults in lh_retrieve).
3149 Instead, use NULL for the CONF pointer in CONF_get_string and
3150 CONF_get_number (which may use environment variables) and directly
3151 return NULL from CONF_get_section.
3154 *) Fix potential buffer overrun for EBCDIC.
3157 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3158 keyUsage if basicConstraints absent for a CA.
3161 *) Make SMIME_write_PKCS7() write mail header values with a format that
3162 is more generally accepted (no spaces before the semicolon), since
3163 some programs can't parse those values properly otherwise. Also make
3164 sure BIO's that break lines after each write do not create invalid
3168 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3169 macros previously used would not encode an empty SEQUENCE OF
3170 and break the signature.
3172 [This change does not apply to 0.9.7.]
3174 *) Zero the premaster secret after deriving the master secret in
3178 *) Add some EVP_add_digest_alias registrations (as found in
3179 OpenSSL_add_all_digests()) to SSL_library_init()
3180 aka OpenSSL_add_ssl_algorithms(). This provides improved
3181 compatibility with peers using X.509 certificates
3182 with unconventional AlgorithmIdentifier OIDs.
3185 *) Fix for Irix with NO_ASM.
3186 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3188 *) ./config script fixes.
3189 [Ulf Moeller, Richard Levitte]
3191 *) Fix 'openssl passwd -1'.
3194 *) Change PKCS12_key_gen_asc() so it can cope with non null
3195 terminated strings whose length is passed in the passlen
3196 parameter, for example from PEM callbacks. This was done
3197 by adding an extra length parameter to asc2uni().
3198 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3200 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3201 call failed, free the DSA structure.
3204 *) Fix to uni2asc() to cope with zero length Unicode strings.
3205 These are present in some PKCS#12 files.
3208 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3209 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3210 when writing a 32767 byte record.
3211 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3213 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3214 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3216 (RSA objects have a reference count access to which is protected
3217 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3218 so they are meant to be shared between threads.)
3219 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3220 "Reddie, Steven" <Steven.Reddie@ca.com>]
3222 *) Fix a deadlock in CRYPTO_mem_leaks().
3225 *) Use better test patterns in bntest.
3228 *) rand_win.c fix for Borland C.
3231 *) BN_rshift bugfix for n == 0.
3234 *) Add a 'bctest' script that checks for some known 'bc' bugs
3235 so that 'make test' does not abort just because 'bc' is broken.
3238 *) Store verify_result within SSL_SESSION also for client side to
3239 avoid potential security hole. (Re-used sessions on the client side
3240 always resulted in verify_result==X509_V_OK, not using the original
3241 result of the server certificate verification.)
3244 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3245 SSL3_RT_APPLICATION_DATA, return 0.
3246 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3250 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3251 releases, have been re-implemented by renaming the previous
3252 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3253 and ssl3_read_internal, respectively, and adding 'peek' parameters
3254 to them. The new ssl[23]_{read,peek} functions are calls to
3255 ssl[23]_read_internal with the 'peek' flag set appropriately.
3256 A 'peek' parameter has also been added to ssl3_read_bytes, which
3257 does the actual work for ssl3_read_internal.
3260 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3261 the method-specific "init()" handler. Also clean up ex_data after
3262 calling the method-specific "finish()" handler. Previously, this was
3263 happening the other way round.
3266 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3267 The previous value, 12, was not always sufficient for BN_mod_exp().
3270 *) Make sure that shared libraries get the internal name engine with
3271 the full version number and not just 0. This should mark the
3272 shared libraries as not backward compatible. Of course, this should
3273 be changed again when we can guarantee backward binary compatibility.
3276 *) Fix typo in get_cert_by_subject() in by_dir.c
3277 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3279 *) Rework the system to generate shared libraries:
3281 - Make note of the expected extension for the shared libraries and
3282 if there is a need for symbolic links from for example libcrypto.so.0
3283 to libcrypto.so.0.9.7. There is extended info in Configure for
3286 - Make as few rebuilds of the shared libraries as possible.
3288 - Still avoid linking the OpenSSL programs with the shared libraries.
3290 - When installing, install the shared libraries separately from the
3294 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3296 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3297 and not in SSL_clear because the latter is also used by the
3298 accept/connect functions; previously, the settings made by
3299 SSL_set_read_ahead would be lost during the handshake.
3300 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3302 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3303 Previously, it would create entries for disableed algorithms no
3307 *) Added several new manual pages for SSL_* function.
3310 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3312 *) In ssl23_get_client_hello, generate an error message when faced
3313 with an initial SSL 3.0/TLS record that is too small to contain the
3314 first two bytes of the ClientHello message, i.e. client_version.
3315 (Note that this is a pathologic case that probably has never happened
3316 in real life.) The previous approach was to use the version number
3317 from the record header as a substitute; but our protocol choice
3318 should not depend on that one because it is not authenticated
3319 by the Finished messages.
3322 *) More robust randomness gathering functions for Windows.
3323 [Jeffrey Altman <jaltman@columbia.edu>]
3325 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3326 not set then we don't setup the error code for issuer check errors
3327 to avoid possibly overwriting other errors which the callback does
3328 handle. If an application does set the flag then we assume it knows
3329 what it is doing and can handle the new informational codes
3333 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3334 a general "ANY" type, as such it should be able to decode anything
3335 including tagged types. However it didn't check the class so it would
3336 wrongly interpret tagged types in the same way as their universal
3337 counterpart and unknown types were just rejected. Changed so that the
3338 tagged and unknown types are handled in the same way as a SEQUENCE:
3339 that is the encoding is stored intact. There is also a new type
3340 "V_ASN1_OTHER" which is used when the class is not universal, in this
3341 case we have no idea what the actual type is so we just lump them all
3345 *) On VMS, stdout may very well lead to a file that is written to
3346 in a record-oriented fashion. That means that every write() will
3347 write a separate record, which will be read separately by the
3348 programs trying to read from it. This can be very confusing.
3350 The solution is to put a BIO filter in the way that will buffer
3351 text until a linefeed is reached, and then write everything a
3352 line at a time, so every record written will be an actual line,
3353 not chunks of lines and not (usually doesn't happen, but I've
3354 seen it once) several lines in one record. BIO_f_linebuffer() is
3357 Currently, it's a VMS-only method, because that's where it has
3358 been tested well enough.
3361 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3362 it can return incorrect results.
3363 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3364 but it was in 0.9.6-beta[12].)
3367 *) Disable the check for content being present when verifying detached
3368 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3369 include zero length content when signing messages.
3372 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3373 BIO_ctrl (for BIO pairs).
3376 *) Add DSO method for VMS.
3379 *) Bug fix: Montgomery multiplication could produce results with the
3383 *) Add RPM specification openssl.spec and modify it to build three
3384 packages. The default package contains applications, application
3385 documentation and run-time libraries. The devel package contains
3386 include files, static libraries and function documentation. The
3387 doc package contains the contents of the doc directory. The original
3388 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3391 *) Add a large number of documentation files for many SSL routines.
3392 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3394 *) Add a configuration entry for Sony News 4.
3395 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3397 *) Don't set the two most significant bits to one when generating a
3398 random number < q in the DSA library.
3401 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3402 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3403 the underlying transport is blocking) if a handshake took place.
3404 (The default behaviour is needed by applications such as s_client
3405 and s_server that use select() to determine when to use SSL_read;
3406 but for applications that know in advance when to expect data, it
3407 just makes things more complicated.)
3410 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3414 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3415 work better on such systems.
3416 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3418 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3419 Update PKCS12_parse() so it copies the friendlyName and the
3420 keyid to the certificates aux info.
3423 *) Fix bug in PKCS7_verify() which caused an infinite loop
3424 if there was more than one signature.
3425 [Sven Uszpelkat <su@celocom.de>]
3427 *) Major change in util/mkdef.pl to include extra information
3428 about each symbol, as well as presentig variables as well
3429 as functions. This change means that there's n more need
3430 to rebuild the .num files when some algorithms are excluded.
3433 *) Allow the verify time to be set by an application,
3434 rather than always using the current time.
3437 *) Phase 2 verify code reorganisation. The certificate
3438 verify code now looks up an issuer certificate by a
3439 number of criteria: subject name, authority key id
3440 and key usage. It also verifies self signed certificates
3441 by the same criteria. The main comparison function is
3442 X509_check_issued() which performs these checks.
3444 Lot of changes were necessary in order to support this
3445 without completely rewriting the lookup code.
3447 Authority and subject key identifier are now cached.
3449 The LHASH 'certs' is X509_STORE has now been replaced
3450 by a STACK_OF(X509_OBJECT). This is mainly because an
3451 LHASH can't store or retrieve multiple objects with
3452 the same hash value.
3454 As a result various functions (which were all internal
3455 use only) have changed to handle the new X509_STORE
3456 structure. This will break anything that messed round
3457 with X509_STORE internally.
3459 The functions X509_STORE_add_cert() now checks for an
3460 exact match, rather than just subject name.
3462 The X509_STORE API doesn't directly support the retrieval
3463 of multiple certificates matching a given criteria, however
3464 this can be worked round by performing a lookup first
3465 (which will fill the cache with candidate certificates)
3466 and then examining the cache for matches. This is probably
3467 the best we can do without throwing out X509_LOOKUP
3468 entirely (maybe later...).
3470 The X509_VERIFY_CTX structure has been enhanced considerably.
3472 All certificate lookup operations now go via a get_issuer()
3473 callback. Although this currently uses an X509_STORE it
3474 can be replaced by custom lookups. This is a simple way
3475 to bypass the X509_STORE hackery necessary to make this
3476 work and makes it possible to use more efficient techniques
3477 in future. A very simple version which uses a simple
3478 STACK for its trusted certificate store is also provided
3479 using X509_STORE_CTX_trusted_stack().
3481 The verify_cb() and verify() callbacks now have equivalents
3482 in the X509_STORE_CTX structure.
3484 X509_STORE_CTX also has a 'flags' field which can be used
3485 to customise the verify behaviour.
3488 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3489 excludes S/MIME capabilities.
3492 *) When a certificate request is read in keep a copy of the
3493 original encoding of the signed data and use it when outputing
3494 again. Signatures then use the original encoding rather than
3495 a decoded, encoded version which may cause problems if the
3496 request is improperly encoded.
3499 *) For consistency with other BIO_puts implementations, call
3500 buffer_write(b, ...) directly in buffer_puts instead of calling
3503 In BIO_puts, increment b->num_write as in BIO_write.
3504 [Peter.Sylvester@EdelWeb.fr]
3506 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3507 BN_zero, we may not return a BIGNUM with an array consisting of
3511 *) Avoid calling abort() from within the library when problems are
3512 detected, except if preprocessor symbols have been defined
3513 (such as REF_CHECK, BN_DEBUG etc.).
3516 *) New openssl application 'rsautl'. This utility can be
3517 used for low level RSA operations. DER public key
3518 BIO/fp routines also added.
3521 *) New Configure entry and patches for compiling on QNX 4.
3522 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3524 *) A demo state-machine implementation was sponsored by
3525 Nuron (http://www.nuron.com/) and is now available in
3526 demos/state_machine.
3529 *) New options added to the 'dgst' utility for signature
3530 generation and verification.
3533 *) Unrecognized PKCS#7 content types are now handled via a
3534 catch all ASN1_TYPE structure. This allows unsupported
3535 types to be stored as a "blob" and an application can
3536 encode and decode it manually.
3539 *) Fix various signed/unsigned issues to make a_strex.c
3541 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3543 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3544 length if passed a buffer. ASN1_INTEGER_to_BN failed
3545 if passed a NULL BN and its argument was negative.
3546 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3548 *) Modification to PKCS#7 encoding routines to output definite
3549 length encoding. Since currently the whole structures are in
3550 memory there's not real point in using indefinite length
3551 constructed encoding. However if OpenSSL is compiled with
3552 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3555 *) Added BIO_vprintf() and BIO_vsnprintf().
3558 *) Added more prefixes to parse for in the the strings written
3559 through a logging bio, to cover all the levels that are available
3560 through syslog. The prefixes are now:
3562 PANIC, EMERG, EMR => LOG_EMERG
3563 ALERT, ALR => LOG_ALERT
3564 CRIT, CRI => LOG_CRIT
3565 ERROR, ERR => LOG_ERR
3566 WARNING, WARN, WAR => LOG_WARNING
3567 NOTICE, NOTE, NOT => LOG_NOTICE
3568 INFO, INF => LOG_INFO
3569 DEBUG, DBG => LOG_DEBUG
3571 and as before, if none of those prefixes are present at the
3572 beginning of the string, LOG_ERR is chosen.
3574 On Win32, the LOG_* levels are mapped according to this:
3576 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3577 LOG_WARNING => EVENTLOG_WARNING_TYPE
3578 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3582 *) Made it possible to reconfigure with just the configuration
3583 argument "reconf" or "reconfigure". The command line arguments
3584 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3585 and are retrieved from there when reconfiguring.
3589 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3591 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3594 *) The obj_dat.pl script was messing up the sorting of object
3595 names. The reason was that it compared the quoted version
3596 of strings as a result "OCSP" > "OCSP Signing" because
3597 " > SPACE. Changed script to store unquoted versions of
3598 names and add quotes on output. It was also omitting some
3599 names from the lookup table if they were given a default
3600 value (that is if SN is missing it is given the same
3601 value as LN and vice versa), these are now added on the
3602 grounds that if an object has a name we should be able to
3603 look it up. Finally added warning output when duplicate
3604 short or long names are found.
3607 *) Changes needed for Tandem NSK.
3608 [Scott Uroff <scott@xypro.com>]
3610 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3611 RSA_padding_check_SSLv23(), special padding was never detected
3612 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3613 version rollback attacks was not effective.
3615 In s23_clnt.c, don't use special rollback-attack detection padding
3616 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3617 client; similarly, in s23_srvr.c, don't do the rollback check if
3618 SSL 2.0 is the only protocol enabled in the server.
3621 *) Make it possible to get hexdumps of unprintable data with 'openssl
3622 asn1parse'. By implication, the functions ASN1_parse_dump() and
3623 BIO_dump_indent() are added.
3626 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3627 these print out strings and name structures based on various
3628 flags including RFC2253 support and proper handling of
3629 multibyte characters. Added options to the 'x509' utility
3630 to allow the various flags to be set.
3633 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3634 Also change the functions X509_cmp_current_time() and
3635 X509_gmtime_adj() work with an ASN1_TIME structure,
3636 this will enable certificates using GeneralizedTime in validity
3637 dates to be checked.
3640 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3641 negative public key encodings) on by default,
3642 NO_NEG_PUBKEY_BUG can be set to disable it.
3645 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3646 content octets. An i2c_ASN1_OBJECT is unnecessary because
3647 the encoding can be trivially obtained from the structure.
3650 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3651 not read locks (CRYPTO_r_[un]lock).
3654 *) A first attempt at creating official support for shared
3655 libraries through configuration. I've kept it so the
3656 default is static libraries only, and the OpenSSL programs
3657 are always statically linked for now, but there are
3658 preparations for dynamic linking in place.
3659 This has been tested on Linux and Tru64.
3662 *) Randomness polling function for Win9x, as described in:
3663 Peter Gutmann, Software Generation of Practically Strong
3667 *) Fix so PRNG is seeded in req if using an already existing
3671 *) New options to smime application. -inform and -outform
3672 allow alternative formats for the S/MIME message including
3673 PEM and DER. The -content option allows the content to be
3674 specified separately. This should allow things like Netscape
3675 form signing output easier to verify.
3678 *) Fix the ASN1 encoding of tags using the 'long form'.
3681 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3682 STRING types. These convert content octets to and from the
3683 underlying type. The actual tag and length octets are
3684 already assumed to have been read in and checked. These
3685 are needed because all other string types have virtually
3686 identical handling apart from the tag. By having versions
3687 of the ASN1 functions that just operate on content octets
3688 IMPLICIT tagging can be handled properly. It also allows
3689 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3690 and ASN1_INTEGER are identical apart from the tag.
3693 *) Change the handling of OID objects as follows:
3695 - New object identifiers are inserted in objects.txt, following
3696 the syntax given in objects.README.
3697 - objects.pl is used to process obj_mac.num and create a new
3699 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3702 This is currently kind of a hack, and the perl code in objects.pl
3703 isn't very elegant, but it works as I intended. The simplest way
3704 to check that it worked correctly is to look in obj_dat.h and
3705 check the array nid_objs and make sure the objects haven't moved
3706 around (this is important!). Additions are OK, as well as
3707 consistent name changes.
3710 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3713 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3714 The given file adds to whatever has already been seeded into the
3715 random pool through the RANDFILE configuration file option or
3716 environment variable, or the default random state file.
3719 *) mkstack.pl now sorts each macro group into lexical order.
3720 Previously the output order depended on the order the files
3721 appeared in the directory, resulting in needless rewriting
3725 *) Patches to make OpenSSL compile under Win32 again. Mostly
3726 work arounds for the VC++ problem that it treats func() as
3727 func(void). Also stripped out the parts of mkdef.pl that
3728 added extra typesafe functions: these no longer exist.
3731 *) Reorganisation of the stack code. The macros are now all
3732 collected in safestack.h . Each macro is defined in terms of
3733 a "stack macro" of the form SKM_<name>(type, a, b). The
3734 DEBUG_SAFESTACK is now handled in terms of function casts,
3735 this has the advantage of retaining type safety without the
3736 use of additional functions. If DEBUG_SAFESTACK is not defined
3737 then the non typesafe macros are used instead. Also modified the
3738 mkstack.pl script to handle the new form. Needs testing to see
3739 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3740 the default if no major problems. Similar behaviour for ASN1_SET_OF
3741 and PKCS12_STACK_OF.
3744 *) When some versions of IIS use the 'NET' form of private key the
3745 key derivation algorithm is different. Normally MD5(password) is
3746 used as a 128 bit RC4 key. In the modified case
3747 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3748 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3749 as the old Netscape_RSA functions except they have an additional
3750 'sgckey' parameter which uses the modified algorithm. Also added
3751 an -sgckey command line option to the rsa utility. Thanks to
3752 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3753 algorithm to openssl-dev.
3756 *) The evp_local.h macros were using 'c.##kname' which resulted in
3757 invalid expansion on some systems (SCO 5.0.5 for example).
3758 Corrected to 'c.kname'.
3759 [Phillip Porch <root@theporch.com>]
3761 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3762 a STACK of email addresses from a certificate or request, these look
3763 in the subject name and the subject alternative name extensions and
3764 omit any duplicate addresses.
3767 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3768 This makes DSA verification about 2 % faster.
3771 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3772 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3773 plus overhead for 1024 bit moduli).
3774 This makes exponentiations about 0.5 % faster for 1024 bit
3775 exponents (as measured by "openssl speed rsa2048").
3778 *) Rename memory handling macros to avoid conflicts with other
3780 Malloc => OPENSSL_malloc
3781 Malloc_locked => OPENSSL_malloc_locked
3782 Realloc => OPENSSL_realloc
3783 Free => OPENSSL_free
3786 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3787 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3790 *) CygWin32 support.
3791 [John Jarvie <jjarvie@newsguy.com>]
3793 *) The type-safe stack code has been rejigged. It is now only compiled
3794 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3795 by default all type-specific stack functions are "#define"d back to
3796 standard stack functions. This results in more streamlined output
3797 but retains the type-safety checking possibilities of the original
3801 *) The STACK code has been cleaned up, and certain type declarations
3802 that didn't make a lot of sense have been brought in line. This has
3803 also involved a cleanup of sorts in safestack.h to more correctly
3804 map type-safe stack functions onto their plain stack counterparts.
3805 This work has also resulted in a variety of "const"ifications of
3806 lots of the code, especially "_cmp" operations which should normally
3807 be prototyped with "const" parameters anyway.
3810 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3811 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3812 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3813 where all of 'md' is used each time the PRNG is used, but 'state'
3814 is used only indexed by a cyclic counter. As entropy may not be
3815 well distributed from the beginning, 'md' is important as a
3816 chaining variable. However, the output function chains only half
3817 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3818 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3819 in all of 'state' being rewritten, with the new values depending
3820 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3823 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3824 the handshake is continued after ssl_verify_cert_chain();
3825 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3826 can lead to 'unexplainable' connection aborts later.
3827 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3829 *) Major EVP API cipher revision.
3830 Add hooks for extra EVP features. This allows various cipher
3831 parameters to be set in the EVP interface. Support added for variable
3832 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3833 setting of RC2 and RC5 parameters.
3835 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3838 Remove lots of duplicated code from the EVP library. For example *every*
3839 cipher init() function handles the 'iv' in the same way according to the
3840 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3841 for CFB and OFB modes they zero ctx->num.
3843 New functionality allows removal of S/MIME code RC2 hack.
3845 Most of the routines have the same form and so can be declared in terms
3848 By shifting this to the top level EVP_CipherInit() it can be removed from
3849 all individual ciphers. If the cipher wants to handle IVs or keys
3850 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3853 Change lots of functions like EVP_EncryptUpdate() to now return a
3854 value: although software versions of the algorithms cannot fail
3855 any installed hardware versions can.
3858 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3859 this option is set, tolerate broken clients that send the negotiated
3860 protocol version number instead of the requested protocol version
3864 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3865 i.e. non-zero for export ciphersuites, zero otherwise.
3866 Previous versions had this flag inverted, inconsistent with
3867 rsa_tmp_cb (..._TMP_RSA_CB).
3868 [Bodo Moeller; problem reported by Amit Chopra]
3870 *) Add missing DSA library text string. Work around for some IIS
3871 key files with invalid SEQUENCE encoding.
3874 *) Add a document (doc/standards.txt) that list all kinds of standards
3875 and so on that are implemented in OpenSSL.
3878 *) Enhance c_rehash script. Old version would mishandle certificates
3879 with the same subject name hash and wouldn't handle CRLs at all.
3880 Added -fingerprint option to crl utility, to support new c_rehash
3884 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3887 *) Fix for SSL server purpose checking. Server checking was
3888 rejecting certificates which had extended key usage present
3889 but no ssl client purpose.
3890 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3892 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3893 is a little unclear about how a blank password is handled.
3894 Since the password in encoded as a BMPString with terminating
3895 double NULL a zero length password would end up as just the
3896 double NULL. However no password at all is different and is
3897 handled differently in the PKCS#12 key generation code. NS
3898 treats a blank password as zero length. MSIE treats it as no
3899 password on export: but it will try both on import. We now do
3900 the same: PKCS12_parse() tries zero length and no password if
3901 the password is set to "" or NULL (NULL is now a valid password:
3902 it wasn't before) as does the pkcs12 application.
3905 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3906 perror when PEM_read_bio_X509_REQ fails, the error message must
3907 be obtained from the error queue.
3910 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3911 it in ERR_remove_state if appropriate, and change ERR_get_state
3912 accordingly to avoid race conditions (this is necessary because
3913 thread_hash is no longer constant once set).
3916 *) Bugfix for linux-elf makefile.one.
3919 *) RSA_get_default_method() will now cause a default
3920 RSA_METHOD to be chosen if one doesn't exist already.
3921 Previously this was only set during a call to RSA_new()
3922 or RSA_new_method(NULL) meaning it was possible for
3923 RSA_get_default_method() to return NULL.
3926 *) Added native name translation to the existing DSO code
3927 that will convert (if the flag to do so is set) filenames
3928 that are sufficiently small and have no path information
3929 into a canonical native form. Eg. "blah" converted to
3930 "libblah.so" or "blah.dll" etc.
3933 *) New function ERR_error_string_n(e, buf, len) which is like
3934 ERR_error_string(e, buf), but writes at most 'len' bytes
3935 including the 0 terminator. For ERR_error_string_n, 'buf'
3937 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3939 *) CONF library reworked to become more general. A new CONF
3940 configuration file reader "class" is implemented as well as a
3941 new functions (NCONF_*, for "New CONF") to handle it. The now
3942 old CONF_* functions are still there, but are reimplemented to
3943 work in terms of the new functions. Also, a set of functions
3944 to handle the internal storage of the configuration data is
3945 provided to make it easier to write new configuration file
3946 reader "classes" (I can definitely see something reading a
3947 configuration file in XML format, for example), called _CONF_*,
3948 or "the configuration storage API"...
3950 The new configuration file reading functions are:
3952 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3953 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3955 NCONF_default, NCONF_WIN32
3957 NCONF_dump_fp, NCONF_dump_bio
3959 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3960 NCONF_new creates a new CONF object. This works in the same way
3961 as other interfaces in OpenSSL, like the BIO interface.
3962 NCONF_dump_* dump the internal storage of the configuration file,
3963 which is useful for debugging. All other functions take the same
3964 arguments as the old CONF_* functions wth the exception of the
3965 first that must be a `CONF *' instead of a `LHASH *'.
3967 To make it easer to use the new classes with the old CONF_* functions,
3968 the function CONF_set_default_method is provided.
3971 *) Add '-tls1' option to 'openssl ciphers', which was already
3972 mentioned in the documentation but had not been implemented.
3973 (This option is not yet really useful because even the additional
3974 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3977 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3978 OpenSSL-based applications) load shared libraries and bind to
3979 them in a portable way.
3980 [Geoff Thorpe, with contributions from Richard Levitte]
3982 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3984 *) Make sure _lrotl and _lrotr are only used with MSVC.
3986 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3987 (the default implementation of RAND_status).
3989 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3990 to '-clrext' (= clear extensions), as intended and documented.
3991 [Bodo Moeller; inconsistency pointed out by Michael Attili
3994 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3995 was larger than the MD block size.
3996 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3998 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3999 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4000 using the passed key: if the passed key was a private key the result
4001 of X509_print(), for example, would be to print out all the private key
4005 *) des_quad_cksum() byte order bug fix.
4006 [Ulf Möller, using the problem description in krb4-0.9.7, where
4007 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4009 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4011 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4013 *) For easily testing in shell scripts whether some command
4014 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4015 returns with exit code 0 iff no command of the given name is available.
4016 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4017 the output goes to stdout and nothing is printed to stderr.
4018 Additional arguments are always ignored.
4020 Since for each cipher there is a command of the same name,
4021 the 'no-cipher' compilation switches can be tested this way.
4023 ('openssl no-XXX' is not able to detect pseudo-commands such
4024 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4027 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4030 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4031 is set; it will be thrown away anyway because each handshake creates
4033 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4034 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4035 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4036 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4039 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4040 'Q' and 'R' lose their special meanings (quit/renegotiate).
4041 This is part of what -quiet does; unlike -quiet, -ign_eof
4042 does not suppress any output.
4045 *) Add compatibility options to the purpose and trust code. The
4046 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4047 accepts a certificate or CA, this was the previous behaviour,
4048 with all the associated security issues.
4050 X509_TRUST_COMPAT is the old trust behaviour: only and
4051 automatically trust self signed roots in certificate store. A
4052 new trust setting X509_TRUST_DEFAULT is used to specify that
4053 a purpose has no associated trust setting and it should instead
4054 use the value in the default purpose.
4057 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4058 and fix a memory leak.
4061 *) In util/mkerr.pl (which implements 'make errors'), preserve
4062 reason strings from the previous version of the .c file, as
4063 the default to have only downcase letters (and digits) in
4064 automatically generated reasons codes is not always appropriate.
4067 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4068 using strerror. Previously, ERR_reason_error_string() returned
4069 library names as reason strings for SYSerr; but SYSerr is a special
4070 case where small numbers are errno values, not library numbers.
4073 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4074 converts DSA parameters into DH parameters. (When creating parameters,
4075 DSA_generate_parameters is used.)
4078 *) Include 'length' (recommended exponent length) in C code generated
4079 by 'openssl dhparam -C'.
4082 *) The second argument to set_label in perlasm was already being used
4083 so couldn't be used as a "file scope" flag. Moved to third argument
4087 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4088 instead of RAND_bytes for encryption IVs and salts.
4091 *) Include RAND_status() into RAND_METHOD instead of implementing
4092 it only for md_rand.c Otherwise replacing the PRNG by calling
4093 RAND_set_rand_method would be impossible.
4096 *) Don't let DSA_generate_key() enter an infinite loop if the random
4097 number generation fails.
4100 *) New 'rand' application for creating pseudo-random output.
4103 *) Added configuration support for Linux/IA64
4104 [Rolf Haberrecker <rolf@suse.de>]
4106 *) Assembler module support for Mingw32.
4109 *) Shared library support for HPUX (in shlib/).
4110 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4112 *) Shared library support for Solaris gcc.
4113 [Lutz Behnke <behnke@trustcenter.de>]
4115 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4117 *) PKCS7_encrypt() was adding text MIME headers twice because they
4118 were added manually and by SMIME_crlf_copy().
4121 *) In bntest.c don't call BN_rand with zero bits argument.
4122 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4124 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4125 case was implemented. This caused BN_div_recp() to fail occasionally.
4128 *) Add an optional second argument to the set_label() in the perl
4129 assembly language builder. If this argument exists and is set
4130 to 1 it signals that the assembler should use a symbol whose
4131 scope is the entire file, not just the current function. This
4132 is needed with MASM which uses the format label:: for this scope.
4133 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4135 *) Change the ASN1 types so they are typedefs by default. Before
4136 almost all types were #define'd to ASN1_STRING which was causing
4137 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4141 *) Change names of new functions to the new get1/get0 naming
4142 convention: After 'get1', the caller owns a reference count
4143 and has to call ..._free; 'get0' returns a pointer to some
4144 data structure without incrementing reference counters.
4145 (Some of the existing 'get' functions increment a reference
4146 counter, some don't.)
4147 Similarly, 'set1' and 'add1' functions increase reference
4148 counters or duplicate objects.
4151 *) Allow for the possibility of temp RSA key generation failure:
4152 the code used to assume it always worked and crashed on failure.
4155 *) Fix potential buffer overrun problem in BIO_printf().
4156 [Ulf Möller, using public domain code by Patrick Powell; problem
4157 pointed out by David Sacerdote <das33@cornell.edu>]
4159 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4160 RAND_egd() and RAND_status(). In the command line application,
4161 the EGD socket can be specified like a seed file using RANDFILE
4165 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4166 Some CAs (e.g. Verisign) distribute certificates in this form.
4169 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4170 list to exclude them. This means that no special compilation option
4171 is needed to use anonymous DH: it just needs to be included in the
4175 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4176 EVP_MD_type. The old functionality is available in a new macro called
4177 EVP_MD_md(). Change code that uses it and update docs.
4180 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4181 where the 'void *' argument is replaced by a function pointer argument.
4182 Previously 'void *' was abused to point to functions, which works on
4183 many platforms, but is not correct. As these functions are usually
4184 called by macros defined in OpenSSL header files, most source code
4185 should work without changes.
4188 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4189 sections with information on -D... compiler switches used for
4190 compiling the library so that applications can see them. To enable
4191 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4192 must be defined. E.g.,
4193 #define OPENSSL_ALGORITHM_DEFINES
4194 #include <openssl/opensslconf.h>
4195 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4196 [Richard Levitte, Ulf and Bodo Möller]
4198 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4202 *) Change the 'other' type in certificate aux info to a STACK_OF
4203 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4204 the required ASN1 format: arbitrary types determined by an OID.
4207 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4208 argument to 'req'. This is not because the function is newer or
4209 better than others it just uses the work 'NEW' in the certificate
4210 request header lines. Some software needs this.
4213 *) Reorganise password command line arguments: now passwords can be
4214 obtained from various sources. Delete the PEM_cb function and make
4215 it the default behaviour: i.e. if the callback is NULL and the
4216 usrdata argument is not NULL interpret it as a null terminated pass
4217 phrase. If usrdata and the callback are NULL then the pass phrase
4218 is prompted for as usual.
4221 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4222 the support is automatically enabled. The resulting binaries will
4223 autodetect the card and use it if present.
4224 [Ben Laurie and Compaq Inc.]
4226 *) Work around for Netscape hang bug. This sends certificate request
4227 and server done in one record. Since this is perfectly legal in the
4228 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4229 the bugs/SSLv3 entry for more info.
4232 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4235 *) Add -rand argument to smime and pkcs12 applications and read/write
4239 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4242 *) Add command line password options to the remaining applications.
4245 *) Bug fix for BN_div_recp() for numerators with an even number of
4249 *) More tests in bntest.c, and changed test_bn output.
4252 *) ./config recognizes MacOS X now.
4255 *) Bug fix for BN_div() when the first words of num and divsor are
4256 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4259 *) Add support for various broken PKCS#8 formats, and command line
4260 options to produce them.
4263 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4264 get temporary BIGNUMs from a BN_CTX.
4267 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4271 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4272 include a #define from the old name to the new. The original intent
4273 was that statically linked binaries could for example just call
4274 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4275 link with digests. This never worked becayse SSLeay_add_all_digests()
4276 and SSLeay_add_all_ciphers() were in the same source file so calling
4277 one would link with the other. They are now in separate source files.
4280 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4283 *) Use a less unusual form of the Miller-Rabin primality test (it used
4284 a binary algorithm for exponentiation integrated into the Miller-Rabin
4285 loop, our standard modexp algorithms are faster).
4288 *) Support for the EBCDIC character set completed.
4289 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4291 *) Source code cleanups: use const where appropriate, eliminate casts,
4292 use void * instead of char * in lhash.
4295 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4296 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4297 this the server could overwrite ephemeral keys that the client
4301 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4302 using 50 iterations of the Rabin-Miller test.
4304 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4305 iterations of the Rabin-Miller test as required by the appendix
4306 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4307 As BN_is_prime_fasttest includes trial division, DSA parameter
4308 generation becomes much faster.
4310 This implies a change for the callback functions in DSA_is_prime
4311 and DSA_generate_parameters: The callback function is called once
4312 for each positive witness in the Rabin-Miller test, not just
4313 occasionally in the inner loop; and the parameters to the
4314 callback function now provide an iteration count for the outer
4315 loop rather than for the current invocation of the inner loop.
4316 DSA_generate_parameters additionally can call the callback
4317 function with an 'iteration count' of -1, meaning that a
4318 candidate has passed the trial division test (when q is generated
4319 from an application-provided seed, trial division is skipped).
4322 *) New function BN_is_prime_fasttest that optionally does trial
4323 division before starting the Rabin-Miller test and has
4324 an additional BN_CTX * argument (whereas BN_is_prime always
4325 has to allocate at least one BN_CTX).
4326 'callback(1, -1, cb_arg)' is called when a number has passed the
4327 trial division stage.
4330 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4334 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4337 *) New function BN_pseudo_rand().
4340 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4341 bignum version of BN_from_montgomery() with the working code from
4342 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4346 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4347 made it impossible to use the same SSL_SESSION data structure in
4348 SSL2 clients in multiple threads.
4351 *) The return value of RAND_load_file() no longer counts bytes obtained
4352 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4353 to seed the PRNG (previously an explicit byte count was required).
4354 [Ulf Möller, Bodo Möller]
4356 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4357 used (char *) instead of (void *) and had casts all over the place.
4360 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4363 *) Retain source code compatibility for BN_prime_checks macro:
4364 BN_is_prime(..., BN_prime_checks, ...) now uses
4365 BN_prime_checks_for_size to determine the appropriate number of
4366 Rabin-Miller iterations.
4369 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4370 DH_CHECK_P_NOT_SAFE_PRIME.
4371 (Check if this is true? OpenPGP calls them "strong".)
4374 *) Merge the functionality of "dh" and "gendh" programs into a new program
4375 "dhparam". The old programs are retained for now but will handle DH keys
4376 (instead of parameters) in future.
4379 *) Make the ciphers, s_server and s_client programs check the return values
4380 when a new cipher list is set.
4383 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4384 ciphers. Before when the 56bit ciphers were enabled the sorting was
4387 The syntax for the cipher sorting has been extended to support sorting by
4388 cipher-strength (using the strength_bits hard coded in the tables).
4389 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4391 Fix a bug in the cipher-command parser: when supplying a cipher command
4392 string with an "undefined" symbol (neither command nor alphanumeric
4393 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4394 an error is flagged.
4396 Due to the strength-sorting extension, the code of the
4397 ssl_create_cipher_list() function was completely rearranged. I hope that
4398 the readability was also increased :-)
4399 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4401 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4402 for the first serial number and places 2 in the serial number file. This
4403 avoids problems when the root CA is created with serial number zero and
4404 the first user certificate has the same issuer name and serial number
4408 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4409 the new code. Add documentation for this stuff.
4412 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4413 X509_*() to X509at_*() on the grounds that they don't handle X509
4414 structures and behave in an analagous way to the X509v3 functions:
4415 they shouldn't be called directly but wrapper functions should be used
4418 So we also now have some wrapper functions that call the X509at functions
4419 when passed certificate requests. (TO DO: similar things can be done with
4420 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4421 things. Some of these need some d2i or i2d and print functionality
4422 because they handle more complex structures.)
4425 *) Add missing #ifndefs that caused missing symbols when building libssl
4426 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4427 NO_RSA in ssl/s2*.c.
4428 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4430 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4431 has a return value which indicates the quality of the random data
4432 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4433 error queue. New function RAND_pseudo_bytes() generates output that is
4434 guaranteed to be unique but not unpredictable. RAND_add is like
4435 RAND_seed, but takes an extra argument for an entropy estimate
4436 (RAND_seed always assumes full entropy).
4439 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4440 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4441 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4442 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4443 false-positive rate of at most 2^-80 for random input.
4446 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4449 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4450 in the 0.9.5 release), this returns the chain
4451 from an X509_CTX structure with a dup of the stack and all
4452 the X509 reference counts upped: so the stack will exist
4453 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4456 Also make SSL_SESSION_print() print out the verify return
4460 *) Add manpage for the pkcs12 command. Also change the default
4461 behaviour so MAC iteration counts are used unless the new
4462 -nomaciter option is used. This improves file security and
4463 only older versions of MSIE (4.0 for example) need it.
4466 *) Honor the no-xxx Configure options when creating .DEF files.
4469 *) Add PKCS#10 attributes to field table: challengePassword,
4470 unstructuredName and unstructuredAddress. These are taken from
4471 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4472 international characters are used.
4474 More changes to X509_ATTRIBUTE code: allow the setting of types
4475 based on strings. Remove the 'loc' parameter when adding
4476 attributes because these will be a SET OF encoding which is sorted
4480 *) Initial changes to the 'req' utility to allow request generation
4481 automation. This will allow an application to just generate a template
4482 file containing all the field values and have req construct the
4485 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4486 used all over the place including certificate requests and PKCS#7
4487 structures. They are currently handled manually where necessary with
4488 some primitive wrappers for PKCS#7. The new functions behave in a
4489 manner analogous to the X509 extension functions: they allow
4490 attributes to be looked up by NID and added.
4492 Later something similar to the X509V3 code would be desirable to
4493 automatically handle the encoding, decoding and printing of the
4494 more complex types. The string types like challengePassword can
4495 be handled by the string table functions.
4497 Also modified the multi byte string table handling. Now there is
4498 a 'global mask' which masks out certain types. The table itself
4499 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4500 is useful when for example there is only one permissible type
4501 (as in countryName) and using the mask might result in no valid
4505 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4506 SSL_get_peer_finished to allow applications to obtain the latest
4507 Finished messages sent to the peer or expected from the peer,
4508 respectively. (SSL_get_peer_finished is usually the Finished message
4509 actually received from the peer, otherwise the protocol will be aborted.)
4511 As the Finished message are message digests of the complete handshake
4512 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4513 be used for external authentication procedures when the authentication
4514 provided by SSL/TLS is not desired or is not enough.
4517 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4518 the host supports BWX extension and if Compaq C is present on the
4519 $PATH. Just exploiting of the BWX extension results in 20-30%
4520 performance kick for some algorithms, e.g. DES and RC4 to mention
4521 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4525 *) Add support for MS "fast SGC". This is arguably a violation of the
4526 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4527 weak crypto and after checking the certificate is SGC a second one
4528 with strong crypto. MS SGC stops the first handshake after receiving
4529 the server certificate message and sends a second client hello. Since
4530 a server will typically do all the time consuming operations before
4531 expecting any further messages from the client (server key exchange
4532 is the most expensive) there is little difference between the two.
4534 To get OpenSSL to support MS SGC we have to permit a second client
4535 hello message after we have sent server done. In addition we have to
4536 reset the MAC if we do get this second client hello.
4539 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4540 if a DER encoded private key is RSA or DSA traditional format. Changed
4541 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4542 format DER encoded private key. Newer code should use PKCS#8 format which
4543 has the key type encoded in the ASN1 structure. Added DER private key
4544 support to pkcs8 application.
4547 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4548 ciphersuites has been selected (as required by the SSL 3/TLS 1
4549 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4550 is set, we interpret this as a request to violate the specification
4551 (the worst that can happen is a handshake failure, and 'correct'
4552 behaviour would result in a handshake failure anyway).
4555 *) In SSL_CTX_add_session, take into account that there might be multiple
4556 SSL_SESSION structures with the same session ID (e.g. when two threads
4557 concurrently obtain them from an external cache).
4558 The internal cache can handle only one SSL_SESSION with a given ID,
4559 so if there's a conflict, we now throw out the old one to achieve
4563 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4564 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4565 some routines that use cipher OIDs: some ciphers do not have OIDs
4566 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4570 *) Simplify the trust setting structure and code. Now we just have
4571 two sequences of OIDs for trusted and rejected settings. These will
4572 typically have values the same as the extended key usage extension
4573 and any application specific purposes.
4575 The trust checking code now has a default behaviour: it will just
4576 check for an object with the same NID as the passed id. Functions can
4577 be provided to override either the default behaviour or the behaviour
4578 for a given id. SSL client, server and email already have functions
4579 in place for compatibility: they check the NID and also return "trusted"
4580 if the certificate is self signed.
4583 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4584 traditional format into an EVP_PKEY structure.
4587 *) Add a password callback function PEM_cb() which either prompts for
4588 a password if usr_data is NULL or otherwise assumes it is a null
4589 terminated password. Allow passwords to be passed on command line
4590 environment or config files in a few more utilities.
4593 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4594 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4595 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4596 Update documentation.
4599 *) Support for ASN1 "NULL" type. This could be handled before by using
4600 ASN1_TYPE but there wasn't any function that would try to read a NULL
4601 and produce an error if it couldn't. For compatibility we also have
4602 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4603 don't allocate anything because they don't need to.
4606 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4608 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4610 *) Rebuild of the memory allocation routines used by OpenSSL code and
4611 possibly others as well. The purpose is to make an interface that
4612 provide hooks so anyone can build a separate set of allocation and
4613 deallocation routines to be used by OpenSSL, for example memory
4614 pool implementations, or something else, which was previously hard
4615 since Malloc(), Realloc() and Free() were defined as macros having
4616 the values malloc, realloc and free, respectively (except for Win32
4617 compilations). The same is provided for memory debugging code.
4618 OpenSSL already comes with functionality to find memory leaks, but
4619 this gives people a chance to debug other memory problems.
4621 With these changes, a new set of functions and macros have appeared:
4623 CRYPTO_set_mem_debug_functions() [F]
4624 CRYPTO_get_mem_debug_functions() [F]
4625 CRYPTO_dbg_set_options() [F]
4626 CRYPTO_dbg_get_options() [F]
4627 CRYPTO_malloc_debug_init() [M]
4629 The memory debug functions are NULL by default, unless the library
4630 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4631 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4632 gives the standard debugging functions that come with OpenSSL) or
4633 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4634 provided by the library user) must be used. When the standard
4635 debugging functions are used, CRYPTO_dbg_set_options can be used to
4636 request additional information:
4637 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4638 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4640 Also, things like CRYPTO_set_mem_functions will always give the
4641 expected result (the new set of functions is used for allocation
4642 and deallocation) at all times, regardless of platform and compiler
4645 To finish it up, some functions that were never use in any other
4646 way than through macros have a new API and new semantic:
4649 CRYPTO_dbg_realloc()
4652 All macros of value have retained their old syntax.
4653 [Richard Levitte and Bodo Moeller]
4655 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4656 ordering of SMIMECapabilities wasn't in "strength order" and there
4657 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4661 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4662 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4663 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4665 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4666 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4667 functionality to handle multipart/signed properly) and a utility
4668 called 'smime' to call all this stuff. This is based on code I
4669 originally wrote for Celo who have kindly allowed it to be
4670 included in OpenSSL.
4673 *) Add variants des_set_key_checked and des_set_key_unchecked of
4674 des_set_key (aka des_key_sched). Global variable des_check_key
4675 decides which of these is called by des_set_key; this way
4676 des_check_key behaves as it always did, but applications and
4677 the library itself, which was buggy for des_check_key == 1,
4678 have a cleaner way to pick the version they need.
4681 *) New function PKCS12_newpass() which changes the password of a
4685 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4686 dynamic mix. In both cases the ids can be used as an index into the
4687 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4688 functions so they accept a list of the field values and the
4689 application doesn't need to directly manipulate the X509_TRUST
4693 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4697 *) Modify the way the V3 extension code looks up extensions. This now
4698 works in a similar way to the object code: we have some "standard"
4699 extensions in a static table which is searched with OBJ_bsearch()
4700 and the application can add dynamic ones if needed. The file
4701 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4702 updated whenever a new extension is added to the core code and kept
4703 in ext_nid order. There is a simple program 'tabtest.c' which checks
4704 this. New extensions are not added too often so this file can readily
4705 be maintained manually.
4707 There are two big advantages in doing things this way. The extensions
4708 can be looked up immediately and no longer need to be "added" using
4709 X509V3_add_standard_extensions(): this function now does nothing.
4710 [Side note: I get *lots* of email saying the extension code doesn't
4711 work because people forget to call this function]
4712 Also no dynamic allocation is done unless new extensions are added:
4713 so if we don't add custom extensions there is no need to call
4714 X509V3_EXT_cleanup().
4717 *) Modify enc utility's salting as follows: make salting the default. Add a
4718 magic header, so unsalted files fail gracefully instead of just decrypting
4719 to garbage. This is because not salting is a big security hole, so people
4720 should be discouraged from doing it.
4723 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4724 digest to be passed on the command line but it only used this
4725 parameter when signing a certificate. Modified so all relevant
4726 operations are affected by the digest parameter including the
4727 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4728 DSA key was used because it didn't fix the digest.
4731 *) Initial certificate chain verify code. Currently tests the untrusted
4732 certificates for consistency with the verify purpose (which is set
4733 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4735 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4736 this is because it will reject chains with invalid extensions whereas
4737 every previous version of OpenSSL and SSLeay made no checks at all.
4739 Trust code: checks the root CA for the relevant trust settings. Trust
4740 settings have an initial value consistent with the verify purpose: e.g.
4741 if the verify purpose is for SSL client use it expects the CA to be
4742 trusted for SSL client use. However the default value can be changed to
4743 permit custom trust settings: one example of this would be to only trust
4744 certificates from a specific "secure" set of CAs.
4746 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4747 which should be used for version portability: especially since the
4748 verify structure is likely to change more often now.
4750 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4751 to set them. If not set then assume SSL clients will verify SSL servers
4754 Two new options to the verify program: -untrusted allows a set of
4755 untrusted certificates to be passed in and -purpose which sets the
4756 intended purpose of the certificate. If a purpose is set then the
4757 new chain verify code is used to check extension consistency.
4760 *) Support for the authority information access extension.
4763 *) Modify RSA and DSA PEM read routines to transparently handle
4764 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4765 public keys in a format compatible with certificate
4766 SubjectPublicKeyInfo structures. Unfortunately there were already
4767 functions called *_PublicKey_* which used various odd formats so
4768 these are retained for compatibility: however the DSA variants were
4769 never in a public release so they have been deleted. Changed dsa/rsa
4770 utilities to handle the new format: note no releases ever handled public
4771 keys so we should be OK.
4773 The primary motivation for this change is to avoid the same fiasco
4774 that dogs private keys: there are several incompatible private key
4775 formats some of which are standard and some OpenSSL specific and
4776 require various evil hacks to allow partial transparent handling and
4777 even then it doesn't work with DER formats. Given the option anything
4778 other than PKCS#8 should be dumped: but the other formats have to
4779 stay in the name of compatibility.
4781 With public keys and the benefit of hindsight one standard format
4782 is used which works with EVP_PKEY, RSA or DSA structures: though
4783 it clearly returns an error if you try to read the wrong kind of key.
4785 Added a -pubkey option to the 'x509' utility to output the public key.
4786 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4787 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4788 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4789 that do the same as the EVP_PKEY_assign_*() except they up the
4790 reference count of the added key (they don't "swallow" the
4794 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4795 CRLs would fail if the file contained no certificates or no CRLs:
4796 added a new function to read in both types and return the number
4797 read: this means that if none are read it will be an error. The
4798 DER versions of the certificate and CRL reader would always fail
4799 because it isn't possible to mix certificates and CRLs in DER format
4800 without choking one or the other routine. Changed this to just read
4801 a certificate: this is the best we can do. Also modified the code
4802 in apps/verify.c to take notice of return codes: it was previously
4803 attempting to read in certificates from NULL pointers and ignoring
4804 any errors: this is one reason why the cert and CRL reader seemed
4805 to work. It doesn't check return codes from the default certificate
4806 routines: these may well fail if the certificates aren't installed.
4809 *) Code to support otherName option in GeneralName.
4812 *) First update to verify code. Change the verify utility
4813 so it warns if it is passed a self signed certificate:
4814 for consistency with the normal behaviour. X509_verify
4815 has been modified to it will now verify a self signed
4816 certificate if *exactly* the same certificate appears
4817 in the store: it was previously impossible to trust a
4818 single self signed certificate. This means that:
4819 openssl verify ss.pem
4820 now gives a warning about a self signed certificate but
4821 openssl verify -CAfile ss.pem ss.pem
4825 *) For servers, store verify_result in SSL_SESSION data structure
4826 (and add it to external session representation).
4827 This is needed when client certificate verifications fails,
4828 but an application-provided verification callback (set by
4829 SSL_CTX_set_cert_verify_callback) allows accepting the session
4830 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4831 but returns 1): When the session is reused, we have to set
4832 ssl->verify_result to the appropriate error code to avoid
4834 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4836 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4837 case in PKCS7_dataInit() where the signed PKCS7 structure
4838 didn't contain any existing data because it was being created.
4839 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4841 *) Add a salt to the key derivation routines in enc.c. This
4842 forms the first 8 bytes of the encrypted file. Also add a
4843 -S option to allow a salt to be input on the command line.
4846 *) New function X509_cmp(). Oddly enough there wasn't a function
4847 to compare two certificates. We do this by working out the SHA1
4848 hash and comparing that. X509_cmp() will be needed by the trust
4852 *) SSL_get1_session() is like SSL_get_session(), but increments
4853 the reference count in the SSL_SESSION returned.
4854 [Geoff Thorpe <geoff@eu.c2.net>]
4856 *) Fix for 'req': it was adding a null to request attributes.
4857 Also change the X509_LOOKUP and X509_INFO code to handle
4858 certificate auxiliary information.
4861 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4865 *) Add the possibility to add extra information to the memory leak
4866 detecting output, to form tracebacks, showing from where each
4867 allocation was originated: CRYPTO_push_info("constant string") adds
4868 the string plus current file name and line number to a per-thread
4869 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4870 is like calling CYRPTO_pop_info() until the stack is empty.
4871 Also updated memory leak detection code to be multi-thread-safe.
4874 *) Add options -text and -noout to pkcs7 utility and delete the
4875 encryption options which never did anything. Update docs.
4878 *) Add options to some of the utilities to allow the pass phrase
4879 to be included on either the command line (not recommended on
4880 OSes like Unix) or read from the environment. Update the
4881 manpages and fix a few bugs.
4884 *) Add a few manpages for some of the openssl commands.
4887 *) Fix the -revoke option in ca. It was freeing up memory twice,
4888 leaking and not finding already revoked certificates.
4891 *) Extensive changes to support certificate auxiliary information.
4892 This involves the use of X509_CERT_AUX structure and X509_AUX
4893 functions. An X509_AUX function such as PEM_read_X509_AUX()
4894 can still read in a certificate file in the usual way but it
4895 will also read in any additional "auxiliary information". By
4896 doing things this way a fair degree of compatibility can be
4897 retained: existing certificates can have this information added
4898 using the new 'x509' options.
4900 Current auxiliary information includes an "alias" and some trust
4901 settings. The trust settings will ultimately be used in enhanced
4902 certificate chain verification routines: currently a certificate
4903 can only be trusted if it is self signed and then it is trusted
4907 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4908 The problem was that one of the replacement routines had not been working
4909 since SSLeay releases. For now the offending routine has been replaced
4910 with non-optimised assembler. Even so, this now gives around 95%
4911 performance improvement for 1024 bit RSA signs.
4914 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4915 handling. Most clients have the effective key size in bits equal to
4916 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4917 A few however don't do this and instead use the size of the decrypted key
4918 to determine the RC2 key length and the AlgorithmIdentifier to determine
4919 the effective key length. In this case the effective key length can still
4920 be 40 bits but the key length can be 168 bits for example. This is fixed
4921 by manually forcing an RC2 key into the EVP_PKEY structure because the
4922 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4923 the key length and effective key length are equal.
4926 *) Add a bunch of functions that should simplify the creation of
4927 X509_NAME structures. Now you should be able to do:
4928 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4929 and have it automatically work out the correct field type and fill in
4930 the structures. The more adventurous can try:
4931 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4932 and it will (hopefully) work out the correct multibyte encoding.
4935 *) Change the 'req' utility to use the new field handling and multibyte
4936 copy routines. Before the DN field creation was handled in an ad hoc
4937 way in req, ca, and x509 which was rather broken and didn't support
4938 BMPStrings or UTF8Strings. Since some software doesn't implement
4939 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4940 using the dirstring_type option. See the new comment in the default
4941 openssl.cnf for more info.
4944 *) Make crypto/rand/md_rand.c more robust:
4945 - Assure unique random numbers after fork().
4946 - Make sure that concurrent threads access the global counter and
4947 md serializably so that we never lose entropy in them
4948 or use exactly the same state in multiple threads.
4949 Access to the large state is not always serializable because
4950 the additional locking could be a performance killer, and
4951 md should be large enough anyway.
4954 *) New file apps/app_rand.c with commonly needed functionality
4955 for handling the random seed file.
4957 Use the random seed file in some applications that previously did not:
4959 dsaparam -genkey (which also ignored its '-rand' option),
4962 x509 (when signing).
4963 Except on systems with /dev/urandom, it is crucial to have a random
4964 seed file at least for key creation, DSA signing, and for DH exchanges;
4965 for RSA signatures we could do without one.
4967 gendh and gendsa (unlike genrsa) used to read only the first byte
4968 of each file listed in the '-rand' option. The function as previously
4969 found in genrsa is now in app_rand.c and is used by all programs
4970 that support '-rand'.
4973 *) In RAND_write_file, use mode 0600 for creating files;
4974 don't just chmod when it may be too late.
4977 *) Report an error from X509_STORE_load_locations
4978 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4981 *) New function ASN1_mbstring_copy() this copies a string in either
4982 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4983 into an ASN1_STRING type. A mask of permissible types is passed
4984 and it chooses the "minimal" type to use or an error if not type
4988 *) Add function equivalents to the various macros in asn1.h. The old
4989 macros are retained with an M_ prefix. Code inside the library can
4990 use the M_ macros. External code (including the openssl utility)
4991 should *NOT* in order to be "shared library friendly".
4994 *) Add various functions that can check a certificate's extensions
4995 to see if it usable for various purposes such as SSL client,
4996 server or S/MIME and CAs of these types. This is currently
4997 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4998 verification. Also added a -purpose flag to x509 utility to
4999 print out all the purposes.
5002 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5006 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5007 for, obtain and decode and extension and obtain its critical flag.
5008 This allows all the necessary extension code to be handled in a
5009 single function call.
5012 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5013 platforms. See crypto/rc4/rc4_enc.c for further details.
5016 *) New -noout option to asn1parse. This causes no output to be produced
5017 its main use is when combined with -strparse and -out to extract data
5018 from a file (which may not be in ASN.1 format).
5021 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5022 when producing the local key id.
5023 [Richard Levitte <levitte@stacken.kth.se>]
5025 *) New option -dhparam in s_server. This allows a DH parameter file to be
5026 stated explicitly. If it is not stated then it tries the first server
5027 certificate file. The previous behaviour hard coded the filename
5031 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5032 a public key to be input or output. For example:
5033 openssl rsa -in key.pem -pubout -out pubkey.pem
5034 Also added necessary DSA public key functions to handle this.
5037 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5038 in the message. This was handled by allowing
5039 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5040 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5042 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5043 to the end of the strings whereas this didn't. This would cause problems
5044 if strings read with d2i_ASN1_bytes() were later modified.
5045 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5047 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5048 data and it contains EOF it will end up returning an error. This is
5049 caused by input 46 bytes long. The cause is due to the way base64
5050 BIOs find the start of base64 encoded data. They do this by trying a
5051 trial decode on each line until they find one that works. When they
5052 do a flag is set and it starts again knowing it can pass all the
5053 data directly through the decoder. Unfortunately it doesn't reset
5054 the context it uses. This means that if EOF is reached an attempt
5055 is made to pass two EOFs through the context and this causes the
5056 resulting error. This can also cause other problems as well. As is
5057 usual with these problems it takes *ages* to find and the fix is
5058 trivial: move one line.
5059 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5061 *) Ugly workaround to get s_client and s_server working under Windows. The
5062 old code wouldn't work because it needed to select() on sockets and the
5063 tty (for keypresses and to see if data could be written). Win32 only
5064 supports select() on sockets so we select() with a 1s timeout on the
5065 sockets and then see if any characters are waiting to be read, if none
5066 are present then we retry, we also assume we can always write data to
5067 the tty. This isn't nice because the code then blocks until we've
5068 received a complete line of data and it is effectively polling the
5069 keyboard at 1s intervals: however it's quite a bit better than not
5070 working at all :-) A dedicated Windows application might handle this
5071 with an event loop for example.
5074 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5075 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5076 will be called when RSA_sign() and RSA_verify() are used. This is useful
5077 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5078 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5079 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5080 This necessitated the support of an extra signature type NID_md5_sha1
5081 for SSL signatures and modifications to the SSL library to use it instead
5082 of calling RSA_public_decrypt() and RSA_private_encrypt().
5085 *) Add new -verify -CAfile and -CApath options to the crl program, these
5086 will lookup a CRL issuers certificate and verify the signature in a
5087 similar way to the verify program. Tidy up the crl program so it
5088 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5089 less strict. It will now permit CRL extensions even if it is not
5090 a V2 CRL: this will allow it to tolerate some broken CRLs.
5093 *) Initialize all non-automatic variables each time one of the openssl
5094 sub-programs is started (this is necessary as they may be started
5095 multiple times from the "OpenSSL>" prompt).
5096 [Lennart Bang, Bodo Moeller]
5098 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5099 removing all other RSA functionality (this is what NO_RSA does). This
5100 is so (for example) those in the US can disable those operations covered
5101 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5105 *) Non-copying interface to BIO pairs.
5106 (still largely untested)
5109 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5110 ASCII string. This was handled independently in various places before.
5113 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5114 UTF8 strings a character at a time.
5117 *) Use client_version from client hello to select the protocol
5118 (s23_srvr.c) and for RSA client key exchange verification
5119 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5122 *) Add various utility functions to handle SPKACs, these were previously
5123 handled by poking round in the structure internals. Added new function
5124 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5125 print, verify and generate SPKACs. Based on an original idea from
5126 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5129 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5132 *) Allow the config file extension section to be overwritten on the
5133 command line. Based on an original idea from Massimiliano Pala
5134 <madwolf@comune.modena.it>. The new option is called -extensions
5135 and can be applied to ca, req and x509. Also -reqexts to override
5136 the request extensions in req and -crlexts to override the crl extensions
5140 *) Add new feature to the SPKAC handling in ca. Now you can include
5141 the same field multiple times by preceding it by "XXXX." for example:
5144 this is the same syntax as used in the req config file.
5147 *) Allow certificate extensions to be added to certificate requests. These
5148 are specified in a 'req_extensions' option of the req section of the
5149 config file. They can be printed out with the -text option to req but
5150 are otherwise ignored at present.
5153 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5154 data read consists of only the final block it would not decrypted because
5155 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5156 A misplaced 'break' also meant the decrypted final block might not be
5157 copied until the next read.
5160 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5161 a few extra parameters to the DH structure: these will be useful if
5162 for example we want the value of 'q' or implement X9.42 DH.
5165 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5166 provides hooks that allow the default DSA functions or functions on a
5167 "per key" basis to be replaced. This allows hardware acceleration and
5168 hardware key storage to be handled without major modification to the
5169 library. Also added low level modexp hooks and CRYPTO_EX structure and
5170 associated functions.
5173 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5174 as "read only": it can't be written to and the buffer it points to will
5175 not be freed. Reading from a read only BIO is much more efficient than
5176 a normal memory BIO. This was added because there are several times when
5177 an area of memory needs to be read from a BIO. The previous method was
5178 to create a memory BIO and write the data to it, this results in two
5179 copies of the data and an O(n^2) reading algorithm. There is a new
5180 function BIO_new_mem_buf() which creates a read only memory BIO from
5181 an area of memory. Also modified the PKCS#7 routines to use read only
5185 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5186 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5187 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5188 but a retry condition occured while trying to read the rest.
5191 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5192 NID_pkcs7_encrypted by default: this was wrong since this should almost
5193 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5194 the encrypted data type: this is a more sensible place to put it and it
5195 allows the PKCS#12 code to be tidied up that duplicated this
5199 *) Changed obj_dat.pl script so it takes its input and output files on
5200 the command line. This should avoid shell escape redirection problems
5204 *) Initial support for certificate extension requests, these are included
5205 in things like Xenroll certificate requests. Included functions to allow
5206 extensions to be obtained and added.
5209 *) -crlf option to s_client and s_server for sending newlines as
5210 CRLF (as required by many protocols).
5213 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5215 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5216 [Ralf S. Engelschall]
5218 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5219 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5221 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5225 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5226 DH parameters/keys (q is lost during that conversion, but the resulting
5227 DH parameters contain its length).
5229 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5230 much faster than DH_generate_parameters (which creates parameters
5231 where p = 2*q + 1), and also the smaller q makes DH computations
5232 much more efficient (160-bit exponentiation instead of 1024-bit
5233 exponentiation); so this provides a convenient way to support DHE
5234 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5235 utter importance to use
5236 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5238 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5239 when such DH parameters are used, because otherwise small subgroup
5240 attacks may become possible!
5243 *) Avoid memory leak in i2d_DHparams.
5246 *) Allow the -k option to be used more than once in the enc program:
5247 this allows the same encrypted message to be read by multiple recipients.
5250 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5251 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5252 it will always use the numerical form of the OID, even if it has a short
5256 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5257 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5258 otherwise bn_mod_exp was called. In the case of hardware keys for example
5259 no private key components need be present and it might store extra data
5260 in the RSA structure, which cannot be accessed from bn_mod_exp.
5261 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5262 private key operations.
5265 *) Added support for SPARC Linux.
5268 *) pem_password_cb function type incompatibly changed from
5269 typedef int pem_password_cb(char *buf, int size, int rwflag);
5271 ....(char *buf, int size, int rwflag, void *userdata);
5272 so that applications can pass data to their callbacks:
5273 The PEM[_ASN1]_{read,write}... functions and macros now take an
5274 additional void * argument, which is just handed through whenever
5275 the password callback is called.
5276 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5278 New function SSL_CTX_set_default_passwd_cb_userdata.
5280 Compatibility note: As many C implementations push function arguments
5281 onto the stack in reverse order, the new library version is likely to
5282 interoperate with programs that have been compiled with the old
5283 pem_password_cb definition (PEM_whatever takes some data that
5284 happens to be on the stack as its last argument, and the callback
5285 just ignores this garbage); but there is no guarantee whatsoever that
5288 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5289 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5290 problems not only on Windows, but also on some Unix platforms.
5291 To avoid problematic command lines, these definitions are now in an
5292 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5293 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5296 *) MIPS III/IV assembler module is reimplemented.
5299 *) More DES library cleanups: remove references to srand/rand and
5300 delete an unused file.
5303 *) Add support for the the free Netwide assembler (NASM) under Win32,
5304 since not many people have MASM (ml) and it can be hard to obtain.
5305 This is currently experimental but it seems to work OK and pass all
5306 the tests. Check out INSTALL.W32 for info.
5309 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5310 without temporary keys kept an extra copy of the server key,
5311 and connections with temporary keys did not free everything in case
5315 *) New function RSA_check_key and new openssl rsa option -check
5316 for verifying the consistency of RSA keys.
5317 [Ulf Moeller, Bodo Moeller]
5319 *) Various changes to make Win32 compile work:
5320 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5321 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5322 comparison" warnings.
5323 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5326 *) Add a debugging option to PKCS#5 v2 key generation function: when
5327 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5328 derived keys are printed to stderr.
5331 *) Copy the flags in ASN1_STRING_dup().
5332 [Roman E. Pavlov <pre@mo.msk.ru>]
5334 *) The x509 application mishandled signing requests containing DSA
5335 keys when the signing key was also DSA and the parameters didn't match.
5337 It was supposed to omit the parameters when they matched the signing key:
5338 the verifying software was then supposed to automatically use the CA's
5339 parameters if they were absent from the end user certificate.
5341 Omitting parameters is no longer recommended. The test was also
5342 the wrong way round! This was probably due to unusual behaviour in
5343 EVP_cmp_parameters() which returns 1 if the parameters match.
5344 This meant that parameters were omitted when they *didn't* match and
5345 the certificate was useless. Certificates signed with 'ca' didn't have
5347 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5349 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5350 The interface is as follows:
5351 Applications can use
5352 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5353 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5354 "off" is now the default.
5355 The library internally uses
5356 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5357 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5358 to disable memory-checking temporarily.
5360 Some inconsistent states that previously were possible (and were
5361 even the default) are now avoided.
5363 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5364 with each memory chunk allocated; this is occasionally more helpful
5365 than just having a counter.
5367 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5369 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5373 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5374 which largely parallels "options", but is for changing API behaviour,
5375 whereas "options" are about protocol behaviour.
5376 Initial "mode" flags are:
5378 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5379 a single record has been written.
5380 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5381 retries use the same buffer location.
5382 (But all of the contents must be
5386 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5389 *) Fix problems with no-hmac etc.
5390 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5392 *) New functions RSA_get_default_method(), RSA_set_method() and
5393 RSA_get_method(). These allows replacement of RSA_METHODs without having
5394 to mess around with the internals of an RSA structure.
5397 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5398 Also really enable memory leak checks in openssl.c and in some
5400 [Chad C. Mulligan, Bodo Moeller]
5402 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5403 up the length of negative integers. This has now been simplified to just
5404 store the length when it is first determined and use it later, rather
5405 than trying to keep track of where data is copied and updating it to
5407 [Steve Henson, reported by Brien Wheeler
5408 <bwheeler@authentica-security.com>]
5410 *) Add a new function PKCS7_signatureVerify. This allows the verification
5411 of a PKCS#7 signature but with the signing certificate passed to the
5412 function itself. This contrasts with PKCS7_dataVerify which assumes the
5413 certificate is present in the PKCS#7 structure. This isn't always the
5414 case: certificates can be omitted from a PKCS#7 structure and be
5415 distributed by "out of band" means (such as a certificate database).
5418 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5419 function prototypes in pem.h, also change util/mkdef.pl to add the
5420 necessary function names.
5423 *) mk1mf.pl (used by Windows builds) did not properly read the
5424 options set by Configure in the top level Makefile, and Configure
5425 was not even able to write more than one option correctly.
5426 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5429 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5430 file to be loaded from a BIO or FILE pointer. The BIO version will
5431 for example allow memory BIOs to contain config info.
5434 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5435 Whoever hopes to achieve shared-library compatibility across versions
5436 must use this, not the compile-time macro.
5437 (Exercise 0.9.4: Which is the minimum library version required by
5439 Note: All this applies only to multi-threaded programs, others don't
5443 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5444 through a BIO pair triggered the default case, i.e.
5445 SSLerr(...,SSL_R_UNKNOWN_STATE).
5448 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5449 can use the SSL library even if none of the specific BIOs is
5453 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5454 for the encoded length.
5455 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5457 *) Add initial documentation of the X509V3 functions.
5460 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5461 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5462 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5463 secure PKCS#8 private key format with a high iteration count.
5466 *) Fix determination of Perl interpreter: A perl or perl5
5467 _directory_ in $PATH was also accepted as the interpreter.
5468 [Ralf S. Engelschall]
5470 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5471 wrong with it but it was very old and did things like calling
5472 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5476 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5477 to use the new extension code.
5480 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5481 with macros. This should make it easier to change their form, add extra
5482 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5486 *) Add to configuration table a new entry that can specify an alternative
5487 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5488 according to Mark Crispin <MRC@Panda.COM>.
5492 *) DES CBC did not update the IV. Weird.
5495 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5496 Changing the behaviour of the former might break existing programs --
5497 where IV updating is needed, des_ncbc_encrypt can be used.
5500 *) When bntest is run from "make test" it drives bc to check its
5501 calculations, as well as internally checking them. If an internal check
5502 fails, it needs to cause bc to give a non-zero result or make test carries
5503 on without noticing the failure. Fixed.
5506 *) DES library cleanups.
5509 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5510 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5511 ciphers. NOTE: although the key derivation function has been verified
5512 against some published test vectors it has not been extensively tested
5513 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5517 *) Instead of "mkdir -p", which is not fully portable, use new
5518 Perl script "util/mkdir-p.pl".
5521 *) Rewrite the way password based encryption (PBE) is handled. It used to
5522 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5523 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5524 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5525 the 'parameter' field of the AlgorithmIdentifier is passed to the
5526 underlying key generation function so it must do its own ASN1 parsing.
5527 This has also changed the EVP_PBE_CipherInit() function which now has a
5528 'parameter' argument instead of literal salt and iteration count values
5529 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5532 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5533 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5534 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5535 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5536 value was just used as a "magic string" and not used directly its
5537 value doesn't matter.
5540 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5544 *) "linux-sparc64" configuration (ultrapenguin).
5545 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5546 "linux-sparc" configuration.
5547 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5549 *) config now generates no-xxx options for missing ciphers.
5552 *) Support the EBCDIC character set (work in progress).
5553 File ebcdic.c not yet included because it has a different license.
5554 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5556 *) Support BS2000/OSD-POSIX.
5557 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5559 *) Make callbacks for key generation use void * instead of char *.
5562 *) Make S/MIME samples compile (not yet tested).
5565 *) Additional typesafe stacks.
5568 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5572 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5574 *) New configuration variant "sco5-gcc".
5576 *) Updated some demos.
5577 [Sean O Riordain, Wade Scholine]
5579 *) Add missing BIO_free at exit of pkcs12 application.
5582 *) Fix memory leak in conf.c.
5585 *) Updates for Win32 to assembler version of MD5.
5588 *) Set #! path to perl in apps/der_chop to where we found it
5589 instead of using a fixed path.
5592 *) SHA library changes for irix64-mips4-cc.
5595 *) Improvements for VMS support.
5599 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5601 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5602 This also avoids the problems with SC4.2 and unpatched SC5.
5603 [Andy Polyakov <appro@fy.chalmers.se>]
5605 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5606 These are required because of the typesafe stack would otherwise break
5607 existing code. If old code used a structure member which used to be STACK
5608 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5609 sk_num or sk_value it would produce an error because the num, data members
5610 are not present in STACK_OF. Now it just produces a warning. sk_set
5611 replaces the old method of assigning a value to sk_value
5612 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5613 that does this will no longer work (and should use sk_set instead) but
5614 this could be regarded as a "questionable" behaviour anyway.
5617 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5618 correctly handle encrypted S/MIME data.
5621 *) Change type of various DES function arguments from des_cblock
5622 (which means, in function argument declarations, pointer to char)
5623 to des_cblock * (meaning pointer to array with 8 char elements),
5624 which allows the compiler to do more typechecking; it was like
5625 that back in SSLeay, but with lots of ugly casts.
5627 Introduce new type const_des_cblock.
5630 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5631 problems: find RecipientInfo structure that matches recipient certificate
5632 and initialise the ASN1 structures properly based on passed cipher.
5635 *) Belatedly make the BN tests actually check the results.
5638 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5639 to and from BNs: it was completely broken. New compilation option
5640 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5641 key elements as negative integers.
5644 *) Reorganize and speed up MD5.
5645 [Andy Polyakov <appro@fy.chalmers.se>]
5648 [Richard Levitte <richard@levitte.org>]
5650 *) New option -out to asn1parse to allow the parsed structure to be
5651 output to a file. This is most useful when combined with the -strparse
5652 option to examine the output of things like OCTET STRINGS.
5655 *) Make SSL library a little more fool-proof by not requiring any longer
5656 that SSL_set_{accept,connect}_state be called before
5657 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5658 in many applications because usually everything *appeared* to work as
5659 intended anyway -- now it really works as intended).
5662 *) Move openssl.cnf out of lib/.
5665 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5666 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5667 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5668 [Ralf S. Engelschall]
5670 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5671 handle PKCS#7 enveloped data properly.
5672 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5674 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5675 copying pointers. The cert_st handling is changed by this in
5676 various ways (and thus what used to be known as ctx->default_cert
5677 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5678 any longer when s->cert does not give us what we need).
5679 ssl_cert_instantiate becomes obsolete by this change.
5680 As soon as we've got the new code right (possibly it already is?),
5681 we have solved a couple of bugs of the earlier code where s->cert
5682 was used as if it could not have been shared with other SSL structures.
5684 Note that using the SSL API in certain dirty ways now will result
5685 in different behaviour than observed with earlier library versions:
5686 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5687 does not influence s as it used to.
5689 In order to clean up things more thoroughly, inside SSL_SESSION
5690 we don't use CERT any longer, but a new structure SESS_CERT
5691 that holds per-session data (if available); currently, this is
5692 the peer's certificate chain and, for clients, the server's certificate
5693 and temporary key. CERT holds only those values that can have
5694 meaningful defaults in an SSL_CTX.
5697 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5698 from the internal representation. Various PKCS#7 fixes: remove some
5699 evil casts and set the enc_dig_alg field properly based on the signing
5703 *) Allow PKCS#12 password to be set from the command line or the
5704 environment. Let 'ca' get its config file name from the environment
5705 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5709 *) Allow certificate policies extension to use an IA5STRING for the
5710 organization field. This is contrary to the PKIX definition but
5711 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5715 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5716 without disallowing inline assembler and the like for non-pedantic builds.
5719 *) Support Borland C++ builder.
5720 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5725 *) SHA-1 cleanups and performance enhancements.
5726 [Andy Polyakov <appro@fy.chalmers.se>]
5728 *) Sparc v8plus assembler for the bignum library.
5729 [Andy Polyakov <appro@fy.chalmers.se>]
5731 *) Accept any -xxx and +xxx compiler options in Configure.
5734 *) Update HPUX configuration.
5737 *) Add missing sk_<type>_unshift() function to safestack.h
5738 [Ralf S. Engelschall]
5740 *) New function SSL_CTX_use_certificate_chain_file that sets the
5741 "extra_cert"s in addition to the certificate. (This makes sense
5742 only for "PEM" format files, as chains as a whole are not
5746 *) Support verify_depth from the SSL API.
5747 x509_vfy.c had what can be considered an off-by-one-error:
5748 Its depth (which was not part of the external interface)
5749 was actually counting the number of certificates in a chain;
5750 now it really counts the depth.
5753 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5754 instead of X509err, which often resulted in confusing error
5755 messages since the error codes are not globally unique
5756 (e.g. an alleged error in ssl3_accept when a certificate
5757 didn't match the private key).
5759 *) New function SSL_CTX_set_session_id_context that allows to set a default
5760 value (so that you don't need SSL_set_session_id_context for each
5761 connection using the SSL_CTX).
5764 *) OAEP decoding bug fix.
5767 *) Support INSTALL_PREFIX for package builders, as proposed by
5771 *) New Configure options "threads" and "no-threads". For systems
5772 where the proper compiler options are known (currently Solaris
5773 and Linux), "threads" is the default.
5776 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5779 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5780 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5781 such as /usr/local/bin.
5784 *) "make linux-shared" to build shared libraries.
5785 [Niels Poppe <niels@netbox.org>]
5787 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5790 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5791 extension adding in x509 utility.
5794 *) Remove NOPROTO sections and error code comments.
5797 *) Partial rewrite of the DEF file generator to now parse the ANSI
5801 *) New Configure options --prefix=DIR and --openssldir=DIR.
5804 *) Complete rewrite of the error code script(s). It is all now handled
5805 by one script at the top level which handles error code gathering,
5806 header rewriting and C source file generation. It should be much better
5807 than the old method: it now uses a modified version of Ulf's parser to
5808 read the ANSI prototypes in all header files (thus the old K&R definitions
5809 aren't needed for error creation any more) and do a better job of
5810 translating function codes into names. The old 'ASN1 error code imbedded
5811 in a comment' is no longer necessary and it doesn't use .err files which
5812 have now been deleted. Also the error code call doesn't have to appear all
5813 on one line (which resulted in some large lines...).
5816 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5819 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5820 0 (which usually indicates a closed connection), but continue reading.
5823 *) Fix some race conditions.
5826 *) Add support for CRL distribution points extension. Add Certificate
5827 Policies and CRL distribution points documentation.
5830 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5833 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5834 8 of keying material. Merlin has also confirmed interop with this fix
5835 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5836 [Merlin Hughes <merlin@baltimore.ie>]
5838 *) Fix lots of warnings.
5839 [Richard Levitte <levitte@stacken.kth.se>]
5841 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5842 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5843 [Richard Levitte <levitte@stacken.kth.se>]
5845 *) Fix problems with sizeof(long) == 8.
5846 [Andy Polyakov <appro@fy.chalmers.se>]
5848 *) Change functions to ANSI C.
5851 *) Fix typos in error codes.
5852 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5854 *) Remove defunct assembler files from Configure.
5857 *) SPARC v8 assembler BIGNUM implementation.
5858 [Andy Polyakov <appro@fy.chalmers.se>]
5860 *) Support for Certificate Policies extension: both print and set.
5861 Various additions to support the r2i method this uses.
5864 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5865 return a const string when you are expecting an allocated buffer.
5868 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5869 types DirectoryString and DisplayText.
5872 *) Add code to allow r2i extensions to access the configuration database,
5873 add an LHASH database driver and add several ctx helper functions.
5876 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5877 fail when they extended the size of a BIGNUM.
5880 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5881 support typesafe stack.
5884 *) Fix typo in SSL_[gs]et_options().
5885 [Nils Frostberg <nils@medcom.se>]
5887 *) Delete various functions and files that belonged to the (now obsolete)
5888 old X509V3 handling code.
5891 *) New Configure option "rsaref".
5894 *) Don't auto-generate pem.h.
5897 *) Introduce type-safe ASN.1 SETs.
5900 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5901 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5903 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5904 that links with OpenSSL (well at least cause lots of warnings), but fear
5905 not: the conversion is trivial, and it eliminates loads of evil casts. A
5906 few STACKed things have been converted already. Feel free to convert more.
5907 In the fullness of time, I'll do away with the STACK type altogether.
5910 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5911 specified in <certfile> by updating the entry in the index.txt file.
5912 This way one no longer has to edit the index.txt file manually for
5913 revoking a certificate. The -revoke option does the gory details now.
5914 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5916 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5917 `-text' option at all and this way the `-noout -text' combination was
5918 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5919 [Ralf S. Engelschall]
5921 *) Make sure a corresponding plain text error message exists for the
5922 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5923 verify callback function determined that a certificate was revoked.
5924 [Ralf S. Engelschall]
5926 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5927 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5928 all available cipers including rc5, which was forgotten until now.
5929 In order to let the testing shell script know which algorithms
5930 are available, a new (up to now undocumented) command
5931 "openssl list-cipher-commands" is used.
5934 *) Bugfix: s_client occasionally would sleep in select() when
5935 it should have checked SSL_pending() first.
5938 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5939 the raw DSA values prior to ASN.1 encoding.
5942 *) Tweaks to Configure
5943 [Niels Poppe <niels@netbox.org>]
5945 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5949 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5952 *) New config option to avoid instructions that are illegal on the 80386.
5953 The default code is faster, but requires at least a 486.
5956 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5957 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5958 same as SSL2_VERSION anyway.
5961 *) New "-showcerts" option for s_client.
5964 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5965 application. Various cleanups and fixes.
5968 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5969 modify error routines to work internally. Add error codes and PBE init
5970 to library startup routines.
5973 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5974 packing functions to asn1 and evp. Changed function names and error
5975 codes along the way.
5978 *) PKCS12 integration: and so it begins... First of several patches to
5979 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5980 objects to objects.h
5983 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5984 and display support for Thawte strong extranet extension.
5987 *) Add LinuxPPC support.
5988 [Jeff Dubrule <igor@pobox.org>]
5990 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5991 bn_div_words in alpha.s.
5992 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5994 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5995 OAEP isn't supported when OpenSSL is built with RSAref.
5996 [Ulf Moeller <ulf@fitug.de>]
5998 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5999 so they no longer are missing under -DNOPROTO.
6000 [Soren S. Jorvang <soren@t.dk>]
6003 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6005 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6006 doesn't work when the session is reused. Coming soon!
6009 *) Fix a security hole, that allows sessions to be reused in the wrong
6010 context thus bypassing client cert protection! All software that uses
6011 client certs and session caches in multiple contexts NEEDS PATCHING to
6012 allow session reuse! A fuller solution is in the works.
6013 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6015 *) Some more source tree cleanups (removed obsolete files
6016 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6017 permission on "config" script to be executable) and a fix for the INSTALL
6019 [Ulf Moeller <ulf@fitug.de>]
6021 *) Remove some legacy and erroneous uses of malloc, free instead of
6023 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6025 *) Make rsa_oaep_test return non-zero on error.
6026 [Ulf Moeller <ulf@fitug.de>]
6028 *) Add support for native Solaris shared libraries. Configure
6029 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6030 if someone would make that last step automatic.
6031 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6033 *) ctx_size was not built with the right compiler during "make links". Fixed.
6036 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6037 except NULL ciphers". This means the default cipher list will no longer
6038 enable NULL ciphers. They need to be specifically enabled e.g. with
6039 the string "DEFAULT:eNULL".
6042 *) Fix to RSA private encryption routines: if p < q then it would
6043 occasionally produce an invalid result. This will only happen with
6044 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6047 *) Be less restrictive and allow also `perl util/perlpath.pl
6048 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6049 because this way one can also use an interpreter named `perl5' (which is
6050 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6051 installed as `perl').
6052 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6054 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6055 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6057 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6058 advapi32.lib to Win32 build and change the pem test comparision
6059 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6060 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6061 and crypto/des/ede_cbcm_enc.c.
6064 *) DES quad checksum was broken on big-endian architectures. Fixed.
6067 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6068 Win32 test batch file so it (might) work again. The Win32 test batch file
6069 is horrible: I feel ill....
6072 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6073 in e_os.h. Audit of header files to check ANSI and non ANSI
6074 sections: 10 functions were absent from non ANSI section and not exported
6075 from Windows DLLs. Fixed up libeay.num for new functions.
6078 *) Make `openssl version' output lines consistent.
6079 [Ralf S. Engelschall]
6081 *) Fix Win32 symbol export lists for BIO functions: Added
6082 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6083 to ms/libeay{16,32}.def.
6084 [Ralf S. Engelschall]
6086 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6087 fine under Unix and passes some trivial tests I've now added. But the
6088 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6089 added to make sure no one expects that this stuff really works in the
6090 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6091 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6093 [Ralf S. Engelschall]
6095 *) Fix the generation of two part addresses in perl.
6096 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6098 *) Add config entry for Linux on MIPS.
6099 [John Tobey <jtobey@channel1.com>]
6101 *) Make links whenever Configure is run, unless we are on Windoze.
6104 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6105 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6109 *) Add a useful kludge to allow package maintainers to specify compiler and
6110 other platforms details on the command line without having to patch the
6111 Configure script everytime: One now can use ``perl Configure
6112 <id>:<details>'', i.e. platform ids are allowed to have details appended
6113 to them (seperated by colons). This is treated as there would be a static
6114 pre-configured entry in Configure's %table under key <id> with value
6115 <details> and ``perl Configure <id>'' is called. So, when you want to
6116 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6117 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6118 now, which overrides the FreeBSD-elf entry on-the-fly.
6119 [Ralf S. Engelschall]
6121 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6124 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6125 on the `perl Configure ...' command line. This way one can compile
6126 OpenSSL libraries with Position Independent Code (PIC) which is needed
6127 for linking it into DSOs.
6128 [Ralf S. Engelschall]
6130 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6134 *) Cleaned up the LICENSE document: The official contact for any license
6135 questions now is the OpenSSL core team under openssl-core@openssl.org.
6136 And add a paragraph about the dual-license situation to make sure people
6137 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6138 to the OpenSSL toolkit.
6139 [Ralf S. Engelschall]
6141 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6142 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6143 Additonally cleaned up the `make links' target: Remove unnecessary
6144 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6145 to speed processing and no longer clutter the display with confusing
6146 stuff. Instead only the actually done links are displayed.
6147 [Ralf S. Engelschall]
6149 *) Permit null encryption ciphersuites, used for authentication only. It used
6150 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6151 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6155 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6156 signed attributes when verifying signatures (this would break them),
6157 the detached data encoding was wrong and public keys obtained using
6158 X509_get_pubkey() weren't freed.
6161 *) Add text documentation for the BUFFER functions. Also added a work around
6162 to a Win95 console bug. This was triggered by the password read stuff: the
6163 last character typed gets carried over to the next fread(). If you were
6164 generating a new cert request using 'req' for example then the last
6165 character of the passphrase would be CR which would then enter the first
6169 *) Added the new `Includes OpenSSL Cryptography Software' button as
6170 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6171 button and can be used by applications based on OpenSSL to show the
6172 relationship to the OpenSSL project.
6173 [Ralf S. Engelschall]
6175 *) Remove confusing variables in function signatures in files
6176 ssl/ssl_lib.c and ssl/ssl.h.
6177 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6179 *) Don't install bss_file.c under PREFIX/include/
6180 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6182 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6183 functions that return function pointers and has support for NT specific
6184 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6185 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6186 unsigned to signed types: this was killing the Win32 compile.
6189 *) Add new certificate file to stack functions,
6190 SSL_add_dir_cert_subjects_to_stack() and
6191 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6192 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6193 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6194 This means that Apache-SSL and similar packages don't have to mess around
6195 to add as many CAs as they want to the preferred list.
6198 *) Experiment with doxygen documentation. Currently only partially applied to
6200 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6201 openssl.doxy as the configuration file.
6204 *) Get rid of remaining C++-style comments which strict C compilers hate.
6205 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6207 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6208 compiled in by default: it has problems with large keys.
6211 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6212 DH private keys and/or callback functions which directly correspond to
6213 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6214 is needed for applications which have to configure certificates on a
6215 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6217 For the RSA certificate situation is makes no difference, but
6218 for the DSA certificate situation this fixes the "no shared cipher"
6219 problem where the OpenSSL cipher selection procedure failed because the
6220 temporary keys were not overtaken from the context and the API provided
6221 no way to reconfigure them.
6222 The new functions now let applications reconfigure the stuff and they
6223 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6224 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6225 non-public-API function ssl_cert_instantiate() is used as a helper
6226 function and also to reduce code redundancy inside ssl_rsa.c.
6227 [Ralf S. Engelschall]
6229 *) Move s_server -dcert and -dkey options out of the undocumented feature
6230 area because they are useful for the DSA situation and should be
6231 recognized by the users.
6232 [Ralf S. Engelschall]
6234 *) Fix the cipher decision scheme for export ciphers: the export bits are
6235 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6236 SSL_EXP_MASK. So, the original variable has to be used instead of the
6237 already masked variable.
6238 [Richard Levitte <levitte@stacken.kth.se>]
6240 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6241 [Richard Levitte <levitte@stacken.kth.se>]
6243 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6244 from `int' to `unsigned int' because it's a length and initialized by
6245 EVP_DigestFinal() which expects an `unsigned int *'.
6246 [Richard Levitte <levitte@stacken.kth.se>]
6248 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6249 script. Instead use the usual Shell->Perl transition trick.
6250 [Ralf S. Engelschall]
6252 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6253 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6254 -noout -modulus' as it's already the case for `openssl rsa -noout
6255 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6256 currently the public key is printed (a decision which was already done by
6257 `openssl dsa -modulus' in the past) which serves a similar purpose.
6258 Additionally the NO_RSA no longer completely removes the whole -modulus
6259 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6261 [Ralf S. Engelschall]
6263 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6264 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6265 [Arne Ansper <arne@ats.cyber.ee>]
6267 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6268 to be added. Now both 'req' and 'ca' can use new objects defined in the
6272 *) Add cool BIO that does syslog (or event log on NT).
6273 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6275 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6276 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6277 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6278 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6281 *) Add preliminary config info for new extension code.
6284 *) Make RSA_NO_PADDING really use no padding.
6285 [Ulf Moeller <ulf@fitug.de>]
6287 *) Generate errors when private/public key check is done.
6290 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6291 for some CRL extensions and new objects added.
6294 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6295 key usage extension and fuller support for authority key id.
6298 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6299 padding method for RSA, which is recommended for new applications in PKCS
6300 #1 v2.0 (RFC 2437, October 1998).
6301 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6302 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6303 against Bleichbacher's attack on RSA.
6304 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6307 *) Updates to the new SSL compression code
6308 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6310 *) Fix so that the version number in the master secret, when passed
6311 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6312 (because the server will not accept higher), that the version number
6313 is 0x03,0x01, not 0x03,0x00
6314 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6316 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6317 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6318 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6321 *) Support for RAW extensions where an arbitrary extension can be
6322 created by including its DER encoding. See apps/openssl.cnf for
6326 *) Make sure latest Perl versions don't interpret some generated C array
6327 code as Perl array code in the crypto/err/err_genc.pl script.
6328 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6330 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6331 not many people have the assembler. Various Win32 compilation fixes and
6332 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6336 *) Modify configure script 'Configure' to automatically create crypto/date.h
6337 file under Win32 and also build pem.h from pem.org. New script
6338 util/mkfiles.pl to create the MINFO file on environments that can't do a
6339 'make files': perl util/mkfiles.pl >MINFO should work.
6342 *) Major rework of DES function declarations, in the pursuit of correctness
6343 and purity. As a result, many evil casts evaporated, and some weirdness,
6344 too. You may find this causes warnings in your code. Zapping your evil
6345 casts will probably fix them. Mostly.
6348 *) Fix for a typo in asn1.h. Bug fix to object creation script
6349 obj_dat.pl. It considered a zero in an object definition to mean
6350 "end of object": none of the objects in objects.h have any zeros
6351 so it wasn't spotted.
6352 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6354 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6355 Masking (CBCM). In the absence of test vectors, the best I have been able
6356 to do is check that the decrypt undoes the encrypt, so far. Send me test
6357 vectors if you have them.
6360 *) Correct calculation of key length for export ciphers (too much space was
6361 allocated for null ciphers). This has not been tested!
6364 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6365 message is now correct (it understands "crypto" and "ssl" on its
6366 command line). There is also now an "update" option. This will update
6367 the util/ssleay.num and util/libeay.num files with any new functions.
6369 perl util/mkdef.pl crypto ssl update
6370 it will update them.
6373 *) Overhauled the Perl interface (perl/*):
6374 - ported BN stuff to OpenSSL's different BN library
6375 - made the perl/ source tree CVS-aware
6376 - renamed the package from SSLeay to OpenSSL (the files still contain
6377 their history because I've copied them in the repository)
6378 - removed obsolete files (the test scripts will be replaced
6379 by better Test::Harness variants in the future)
6380 [Ralf S. Engelschall]
6382 *) First cut for a very conservative source tree cleanup:
6383 1. merge various obsolete readme texts into doc/ssleay.txt
6384 where we collect the old documents and readme texts.
6385 2. remove the first part of files where I'm already sure that we no
6386 longer need them because of three reasons: either they are just temporary
6387 files which were left by Eric or they are preserved original files where
6388 I've verified that the diff is also available in the CVS via "cvs diff
6389 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6390 the crypto/md/ stuff).
6391 [Ralf S. Engelschall]
6393 *) More extension code. Incomplete support for subject and issuer alt
6394 name, issuer and authority key id. Change the i2v function parameters
6395 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6396 what that's for :-) Fix to ASN1 macro which messed up
6397 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6400 *) Preliminary support for ENUMERATED type. This is largely copied from the
6404 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6405 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6407 *) Make sure `make rehash' target really finds the `openssl' program.
6408 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6410 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6411 like to hear about it if this slows down other processors.
6414 *) Add CygWin32 platform information to Configure script.
6415 [Alan Batie <batie@aahz.jf.intel.com>]
6417 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6418 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6420 *) New program nseq to manipulate netscape certificate sequences
6423 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6427 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6428 but the BN code had some problems that would cause failures when
6429 doing certificate verification and some other functions.
6430 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6432 *) Add ASN1 and PEM code to support netscape certificate sequences.
6435 *) Add ASN1 and PEM code to support netscape certificate sequences.
6438 *) Add several PKIX and private extended key usage OIDs.
6441 *) Modify the 'ca' program to handle the new extension code. Modify
6442 openssl.cnf for new extension format, add comments.
6445 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6446 and add a sample to openssl.cnf so req -x509 now adds appropriate
6450 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6451 error code, add initial support to X509_print() and x509 application.
6454 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6455 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6456 stuff is currently isolated and isn't even compiled yet.
6459 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6460 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6461 Removed the versions check from X509 routines when loading extensions:
6462 this allows certain broken certificates that don't set the version
6463 properly to be processed.
6466 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6467 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6468 can still be regenerated with "make depend".
6471 *) Spelling mistake in C version of CAST-128.
6472 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6474 *) Changes to the error generation code. The perl script err-code.pl
6475 now reads in the old error codes and retains the old numbers, only
6476 adding new ones if necessary. It also only changes the .err files if new
6477 codes are added. The makefiles have been modified to only insert errors
6478 when needed (to avoid needlessly modifying header files). This is done
6479 by only inserting errors if the .err file is newer than the auto generated
6480 C file. To rebuild all the error codes from scratch (the old behaviour)
6481 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6482 or delete all the .err files.
6485 *) CAST-128 was incorrectly implemented for short keys. The C version has
6486 been fixed, but is untested. The assembler versions are also fixed, but
6487 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6488 to regenerate it if needed.
6489 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6490 Hagino <itojun@kame.net>]
6492 *) File was opened incorrectly in randfile.c.
6493 [Ulf Möller <ulf@fitug.de>]
6495 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6496 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6497 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6498 al: it's just almost always a UTCTime. Note this patch adds new error
6499 codes so do a "make errors" if there are problems.
6502 *) Correct Linux 1 recognition in config.
6503 [Ulf Möller <ulf@fitug.de>]
6505 *) Remove pointless MD5 hash when using DSA keys in ca.
6506 [Anonymous <nobody@replay.com>]
6508 *) Generate an error if given an empty string as a cert directory. Also
6509 generate an error if handed NULL (previously returned 0 to indicate an
6510 error, but didn't set one).
6511 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6513 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6516 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6517 parameters. This was causing a warning which killed off the Win32 compile.
6520 *) Remove C++ style comments from crypto/bn/bn_local.h.
6521 [Neil Costigan <neil.costigan@celocom.com>]
6523 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6524 based on a text string, looking up short and long names and finally
6525 "dot" format. The "dot" format stuff didn't work. Added new function
6526 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6527 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6528 OID is not part of the table.
6531 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6532 X509_LOOKUP_by_alias().
6535 *) Sort openssl functions by name.
6538 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6539 encryption from sample DSA keys (in case anyone is interested the password
6543 *) Make _all_ *_free functions accept a NULL pointer.
6544 [Frans Heymans <fheymans@isaserver.be>]
6546 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6548 [Anonymous <nobody@replay.com>]
6550 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6551 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6553 *) Don't blow it for numeric -newkey arguments to apps/req.
6554 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6556 *) Temp key "for export" tests were wrong in s3_srvr.c.
6557 [Anonymous <nobody@replay.com>]
6559 *) Add prototype for temp key callback functions
6560 SSL_CTX_set_tmp_{rsa,dh}_callback().
6563 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6564 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6567 *) X509_name_add_entry() freed the wrong thing after an error.
6568 [Arne Ansper <arne@ats.cyber.ee>]
6570 *) rsa_eay.c would attempt to free a NULL context.
6571 [Arne Ansper <arne@ats.cyber.ee>]
6573 *) BIO_s_socket() had a broken should_retry() on Windoze.
6574 [Arne Ansper <arne@ats.cyber.ee>]
6576 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6577 [Arne Ansper <arne@ats.cyber.ee>]
6579 *) Make sure the already existing X509_STORE->depth variable is initialized
6580 in X509_STORE_new(), but document the fact that this variable is still
6581 unused in the certificate verification process.
6582 [Ralf S. Engelschall]
6584 *) Fix the various library and apps files to free up pkeys obtained from
6585 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6588 *) Fix reference counting in X509_PUBKEY_get(). This makes
6589 demos/maurice/example2.c work, amongst others, probably.
6590 [Steve Henson and Ben Laurie]
6592 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6593 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6594 are no longer created. This way we have a single and consistent command
6595 line interface `openssl <command>', similar to `cvs <command>'.
6596 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6598 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6599 BIT STRING wrapper always have zero unused bits.
6602 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6605 *) Make the top-level INSTALL documentation easier to understand.
6608 *) Makefiles updated to exit if an error occurs in a sub-directory
6609 make (including if user presses ^C) [Paul Sutton]
6611 *) Make Montgomery context stuff explicit in RSA data structure.
6614 *) Fix build order of pem and err to allow for generated pem.h.
6617 *) Fix renumbering bug in X509_NAME_delete_entry().
6620 *) Enhanced the err-ins.pl script so it makes the error library number
6621 global and can add a library name. This is needed for external ASN1 and
6622 other error libraries.
6625 *) Fixed sk_insert which never worked properly.
6628 *) Fix ASN1 macros so they can handle indefinite length construted
6629 EXPLICIT tags. Some non standard certificates use these: they can now
6633 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6634 into a single doc/ssleay.txt bundle. This way the information is still
6635 preserved but no longer messes up this directory. Now it's new room for
6636 the new set of documenation files.
6637 [Ralf S. Engelschall]
6639 *) SETs were incorrectly DER encoded. This was a major pain, because they
6640 shared code with SEQUENCEs, which aren't coded the same. This means that
6641 almost everything to do with SETs or SEQUENCEs has either changed name or
6642 number of arguments.
6643 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6645 *) Fix test data to work with the above.
6648 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6649 was already fixed by Eric for 0.9.1 it seems.
6650 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6652 *) Autodetect FreeBSD3.
6655 *) Fix various bugs in Configure. This affects the following platforms:
6659 unixware-2.0-pentium
6663 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6664 before they are needed.
6667 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6671 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6673 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6674 changed SSLeay to OpenSSL in version strings.
6675 [Ralf S. Engelschall]
6677 *) Some fixups to the top-level documents.
6680 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6681 because the symlink to include/ was missing.
6682 [Ralf S. Engelschall]
6684 *) Incorporated the popular no-RSA/DSA-only patches
6685 which allow to compile a RSA-free SSLeay.
6686 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6688 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6689 when "ssleay" is still not found.
6690 [Ralf S. Engelschall]
6692 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6693 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6695 *) Updated the README file.
6696 [Ralf S. Engelschall]
6698 *) Added various .cvsignore files in the CVS repository subdirs
6699 to make a "cvs update" really silent.
6700 [Ralf S. Engelschall]
6702 *) Recompiled the error-definition header files and added
6703 missing symbols to the Win32 linker tables.
6704 [Ralf S. Engelschall]
6706 *) Cleaned up the top-level documents;
6707 o new files: CHANGES and LICENSE
6708 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6709 o merged COPYRIGHT into LICENSE
6710 o removed obsolete TODO file
6711 o renamed MICROSOFT to INSTALL.W32
6712 [Ralf S. Engelschall]
6714 *) Removed dummy files from the 0.9.1b source tree:
6715 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6716 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6717 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6718 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6719 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6720 [Ralf S. Engelschall]
6722 *) Added various platform portability fixes.
6725 *) The Genesis of the OpenSSL rpject:
6726 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6727 Young and Tim J. Hudson created while they were working for C2Net until
6729 [The OpenSSL Project]
6732 Changes between 0.9.0b and 0.9.1b [not released]
6734 *) Updated a few CA certificates under certs/
6737 *) Changed some BIGNUM api stuff.
6740 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6741 DGUX x86, Linux Alpha, etc.
6744 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6745 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6749 *) Add -strparse option to asn1pars program which parses nested
6751 [Dr Stephen Henson <shenson@bigfoot.com>]
6753 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6756 *) DSA fix for "ca" program.
6759 *) Added "-genkey" option to "dsaparam" program.
6762 *) Added RIPE MD160 (rmd160) message digest.
6765 *) Added -a (all) option to "ssleay version" command.
6768 *) Added PLATFORM define which is the id given to Configure.
6771 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6774 *) Extended the ASN.1 parser routines.
6777 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6780 *) Added a BN_CTX to the BN library.
6783 *) Fixed the weak key values in DES library
6786 *) Changed API in EVP library for cipher aliases.
6789 *) Added support for RC2/64bit cipher.
6792 *) Converted the lhash library to the crypto/mem.c functions.
6795 *) Added more recognized ASN.1 object ids.
6798 *) Added more RSA padding checks for SSL/TLS.
6801 *) Added BIO proxy/filter functionality.
6804 *) Added extra_certs to SSL_CTX which can be used
6805 send extra CA certificates to the client in the CA cert chain sending
6806 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6809 *) Now Fortezza is denied in the authentication phase because
6810 this is key exchange mechanism is not supported by SSLeay at all.
6813 *) Additional PKCS1 checks.
6816 *) Support the string "TLSv1" for all TLS v1 ciphers.
6819 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6820 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6823 *) Fixed a few memory leaks.
6826 *) Fixed various code and comment typos.
6829 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6830 bytes sent in the client random.
6831 [Edward Bishop <ebishop@spyglass.com>]