]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
We stupidly had a separate LIBKRB5 variable for KRB5 library dependencies,
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
6
7 *) Change the "progress" mechanism used in key-generation and
8 primality testing to functions that take a new BN_GENCB pointer in
9 place of callback/argument pairs. The new API functions have "_ex"
10 postfixes and the older functions are reimplemented as wrappers for
11 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12 declarations of the old functions to help (graceful) attempts to
13 migrate to the new functions. Also, the new key-generation API
14 functions operate on a caller-supplied key-structure and return
15 success/failure rather than returning a key or NULL - this is to
16 help make "keygen" another member function of RSA_METHOD etc.
17 [Geoff Thorpe]
18
19 *) Change the ZLIB compression method to be stateful, and make it
20 available to TLS with the number defined in
21 draft-ietf-tls-compression-04.txt.
22 [Richard Levitte]
23
24 *) Add the ASN.1 structures and functions for CertificatePair, which
25 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
26
27 CertificatePair ::= SEQUENCE {
28 forward [0] Certificate OPTIONAL,
29 reverse [1] Certificate OPTIONAL,
30 -- at least one of the pair shall be present -- }
31
32 Also implement the PEM functions to read and write certificate
33 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
34
35 This needed to be defined, mostly for the sake of the LDAP
36 attribute crossCertificatePair, but may prove useful elsewhere as
37 well.
38 [Richard Levitte]
39
40 *) Make it possible to inhibit symlinking of shared libraries in
41 Makefile.shared, for Cygwin's sake.
42 [Richard Levitte]
43
44 *) Extend the BIGNUM API by creating new macros that behave like
45 functions
46
47 void BN_set_sign(BIGNUM *a, int neg);
48 int BN_get_sign(const BIGNUM *a);
49
50 and avoid the need to access 'a->neg' directly in applications.
51 [Nils Larsch <nla@trustcenter.de>]
52
53 *) Implement fast modular reduction for pseudo-Mersenne primes
54 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
55 EC_GROUP_new_curve_GFp() will now automatically use this
56 if applicable.
57 [Nils Larsch <nla@trustcenter.de>]
58
59 *) Add new lock type (CRYPTO_LOCK_BN).
60 [Bodo Moeller]
61
62 *) Change the ENGINE framework to automatically load engines
63 dynamically from specific directories unless they could be
64 found to already be built in or loaded. Move all the
65 current engines except for the cryptodev one to a new
66 directory engines/.
67 The engines in engines/ are built as shared libraries if
68 the "shared" options was given to ./Configure or ./config.
69 Otherwise, they are inserted in libcrypto.a.
70 /usr/local/ssl/engines is the default directory for dynamic
71 engines, but that can be overriden at configure time through
72 the usual use of --prefix and/or --openssldir, and at run
73 time with the environment variable OPENSSL_ENGINES.
74 [Geoff Thorpe and Richard Levitte]
75
76 *) Add Makefile.shared, a helper makefile to build shared
77 libraries. Addapt Makefile.org.
78 [Richard Levitte]
79
80 *) Add version info to Win32 DLLs.
81 [Peter 'Luna' Runestig" <peter@runestig.com>]
82
83 *) Add new 'medium level' PKCS#12 API. Certificates and keys
84 can be added using this API to created arbitrary PKCS#12
85 files while avoiding the low level API.
86
87 New options to PKCS12_create(), key or cert can be NULL and
88 will then be omitted from the output file. The encryption
89 algorithm NIDs can be set to -1 for no encryption, the mac
90 iteration count can be set to 0 to omit the mac.
91
92 Enhance pkcs12 utility by making the -nokeys and -nocerts
93 options work when creating a PKCS#12 file. New option -nomac
94 to omit the mac, NONE can be set for an encryption algorithm.
95 New code is modified to use the enhanced PKCS12_create()
96 instead of the low level API.
97 [Steve Henson]
98
99 *) Extend ASN1 encoder to support indefinite length constructed
100 encoding. This can output sequences tags and octet strings in
101 this form. Modify pk7_asn1.c to support indefinite length
102 encoding. This is experimental and needs additional code to
103 be useful, such as an ASN1 bio and some enhanced streaming
104 PKCS#7 code.
105
106 Extend template encode functionality so that tagging is passed
107 down to the template encoder.
108 [Steve Henson]
109
110 *) Let 'openssl req' fail if an argument to '-newkey' is not
111 recognized instead of using RSA as a default.
112 [Bodo Moeller]
113
114 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
115 As these are not official, they are not included in "ALL";
116 the "ECCdraft" ciphersuite group alias can be used to select them.
117 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
118
119 *) Add ECDH engine support.
120 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
121
122 *) Add ECDH in new directory crypto/ecdh/.
123 TODO: more general interface (return x coordinate, not its hash)
124 TODO: bug: pad x with leading zeros if necessary
125 [Douglas Stebila (Sun Microsystems Laboratories)]
126
127 *) Let BN_rand_range() abort with an error after 100 iterations
128 without success (which indicates a broken PRNG).
129 [Bodo Moeller]
130
131 *) Change BN_mod_sqrt() so that it verifies that the input value
132 is really the square of the return value. (Previously,
133 BN_mod_sqrt would show GIGO behaviour.)
134 [Bodo Moeller]
135
136 *) Add named elliptic curves over binary fields from X9.62, SECG,
137 and WAP/WTLS; add OIDs that were still missing.
138
139 [Sheueling Chang Shantz and Douglas Stebila
140 (Sun Microsystems Laboratories)]
141
142 *) Extend the EC library for elliptic curves over binary fields
143 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
144 New EC_METHOD:
145
146 EC_GF2m_simple_method
147
148 New API functions:
149
150 EC_GROUP_new_curve_GF2m
151 EC_GROUP_set_curve_GF2m
152 EC_GROUP_get_curve_GF2m
153 EC_POINT_set_affine_coordinates_GF2m
154 EC_POINT_get_affine_coordinates_GF2m
155 EC_POINT_set_compressed_coordinates_GF2m
156
157 Point compression for binary fields is disabled by default for
158 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
159 enable it).
160
161 As binary polynomials are represented as BIGNUMs, various members
162 of the EC_GROUP and EC_POINT data structures can be shared
163 between the implementations for prime fields and binary fields;
164 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
165 are essentially identical to their ..._GFp counterparts.
166 (For simplicity, the '..._GFp' prefix has been dropped from
167 various internal method names.)
168
169 An internal 'field_div' method (similar to 'field_mul' and
170 'field_sqr') has been added; this is used only for binary fields.
171
172 [Sheueling Chang Shantz and Douglas Stebila
173 (Sun Microsystems Laboratories)]
174
175 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
176 through methods ('mul', 'precompute_mult').
177
178 The generic implementations (now internally called 'ec_wNAF_mul'
179 and 'ec_wNAF_precomputed_mult') remain the default if these
180 methods are undefined.
181
182 [Sheueling Chang Shantz and Douglas Stebila
183 (Sun Microsystems Laboratories)]
184
185 *) New function EC_GROUP_get_degree, which is defined through
186 EC_METHOD. For curves over prime fields, this returns the bit
187 length of the modulus.
188
189 [Sheueling Chang Shantz and Douglas Stebila
190 (Sun Microsystems Laboratories)]
191
192 *) New functions EC_GROUP_dup, EC_POINT_dup.
193 (These simply call ..._new and ..._copy).
194
195 [Sheueling Chang Shantz and Douglas Stebila
196 (Sun Microsystems Laboratories)]
197
198 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
199 Polynomials are represented as BIGNUMs (where the sign bit is not
200 used) in the following functions [macros]:
201
202 BN_GF2m_add
203 BN_GF2m_sub [= BN_GF2m_add]
204 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
205 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
206 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
207 BN_GF2m_mod_inv
208 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
209 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
210 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
211 BN_GF2m_cmp [= BN_ucmp]
212
213 (Note that only the 'mod' functions are actually for fields GF(2^m).
214 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
215
216 For some functions, an the irreducible polynomial defining a
217 field can be given as an 'unsigned int[]' with strictly
218 decreasing elements giving the indices of those bits that are set;
219 i.e., p[] represents the polynomial
220 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
221 where
222 p[0] > p[1] > ... > p[k] = 0.
223 This applies to the following functions:
224
225 BN_GF2m_mod_arr
226 BN_GF2m_mod_mul_arr
227 BN_GF2m_mod_sqr_arr
228 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
229 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
230 BN_GF2m_mod_exp_arr
231 BN_GF2m_mod_sqrt_arr
232 BN_GF2m_mod_solve_quad_arr
233 BN_GF2m_poly2arr
234 BN_GF2m_arr2poly
235
236 Conversion can be performed by the following functions:
237
238 BN_GF2m_poly2arr
239 BN_GF2m_arr2poly
240
241 bntest.c has additional tests for binary polynomial arithmetic.
242
243 Two implementations for BN_GF2m_mod_div() are available.
244 The default algorithm simply uses BN_GF2m_mod_inv() and
245 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
246 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
247 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
248
249 [Sheueling Chang Shantz and Douglas Stebila
250 (Sun Microsystems Laboratories)]
251
252 *) Add new error code 'ERR_R_DISABLED' that can be used when some
253 functionality is disabled at compile-time.
254 [Douglas Stebila <douglas.stebila@sun.com>]
255
256 *) Change default behaviour of 'openssl asn1parse' so that more
257 information is visible when viewing, e.g., a certificate:
258
259 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
260 mode the content of non-printable OCTET STRINGs is output in a
261 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
262 avoid the appearance of a printable string.
263 [Nils Larsch <nla@trustcenter.de>]
264
265 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
266 functions
267 EC_GROUP_set_asn1_flag()
268 EC_GROUP_get_asn1_flag()
269 EC_GROUP_set_point_conversion_form()
270 EC_GROUP_get_point_conversion_form()
271 These control ASN1 encoding details:
272 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
273 has been set to OPENSSL_EC_NAMED_CURVE.
274 - Points are encoded in uncompressed form by default; options for
275 asn1_for are as for point2oct, namely
276 POINT_CONVERSION_COMPRESSED
277 POINT_CONVERSION_UNCOMPRESSED
278 POINT_CONVERSION_HYBRID
279
280 Also add 'seed' and 'seed_len' members to EC_GROUP with access
281 functions
282 EC_GROUP_set_seed()
283 EC_GROUP_get0_seed()
284 EC_GROUP_get_seed_len()
285 This is used only for ASN1 purposes (so far).
286 [Nils Larsch <nla@trustcenter.de>]
287
288 *) Add 'field_type' member to EC_METHOD, which holds the NID
289 of the appropriate field type OID. The new function
290 EC_METHOD_get_field_type() returns this value.
291 [Nils Larsch <nla@trustcenter.de>]
292
293 *) Add functions
294 EC_POINT_point2bn()
295 EC_POINT_bn2point()
296 EC_POINT_point2hex()
297 EC_POINT_hex2point()
298 providing useful interfaces to EC_POINT_point2oct() and
299 EC_POINT_oct2point().
300 [Nils Larsch <nla@trustcenter.de>]
301
302 *) Change internals of the EC library so that the functions
303 EC_GROUP_set_generator()
304 EC_GROUP_get_generator()
305 EC_GROUP_get_order()
306 EC_GROUP_get_cofactor()
307 are implemented directly in crypto/ec/ec_lib.c and not dispatched
308 to methods, which would lead to unnecessary code duplication when
309 adding different types of curves.
310 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
311
312 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
313 arithmetic, and such that modified wNAFs are generated
314 (which avoid length expansion in many cases).
315 [Bodo Moeller]
316
317 *) Add a function EC_GROUP_check_discriminant() (defined via
318 EC_METHOD) that verifies that the curve discriminant is non-zero.
319
320 Add a function EC_GROUP_check() that makes some sanity tests
321 on a EC_GROUP, its generator and order. This includes
322 EC_GROUP_check_discriminant().
323 [Nils Larsch <nla@trustcenter.de>]
324
325 *) Add ECDSA in new directory crypto/ecdsa/.
326
327 Add applications 'openssl ecparam' and 'openssl ecdsa'
328 (these are based on 'openssl dsaparam' and 'openssl dsa').
329
330 ECDSA support is also included in various other files across the
331 library. Most notably,
332 - 'openssl req' now has a '-newkey ecdsa:file' option;
333 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
334 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
335 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
336 them suitable for ECDSA where domain parameters must be
337 extracted before the specific public key;
338 - ECDSA engine support has been added.
339 [Nils Larsch <nla@trustcenter.de>]
340
341 *) Include some named elliptic curves, and add OIDs from X9.62,
342 SECG, and WAP/WTLS. Each curve can be obtained from the new
343 function
344 EC_GROUP_new_by_nid(),
345 and the list of available named curves can be obtained with
346 EC_get_builtin_curves().
347 Also add a 'curve_name' member to EC_GROUP objects, which can be
348 accessed via
349 EC_GROUP_set_nid()
350 EC_GROUP_get_nid()
351 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
352
353 Changes between 0.9.6h and 0.9.7 [XX xxx 2002]
354
355 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
356 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
357 removed entirely.
358 [Richard Levitte]
359
360 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
361 seems that in spite of existing for more than a year, many application
362 author have done nothing to provide the necessary callbacks, which
363 means that this particular engine will not work properly anywhere.
364 This is a very unfortunate situation which forces us, in the name
365 of usability, to give the hw_ncipher.c a static lock, which is part
366 of libcrypto.
367 NOTE: This is for the 0.9.7 series ONLY. This hack will never
368 appear in 0.9.8 or later. We EXPECT application authors to have
369 dealt properly with this when 0.9.8 is released (unless we actually
370 make such changes in the libcrypto locking code that changes will
371 have to be made anyway).
372 [Richard Levitte]
373
374 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
375 octets have been read, EOF or an error occurs. Without this change
376 some truncated ASN1 structures will not produce an error.
377 [Steve Henson]
378
379 *) Disable Heimdal support, since it hasn't been fully implemented.
380 Still give the possibility to force the use of Heimdal, but with
381 warnings and a request that patches get sent to openssl-dev.
382 [Richard Levitte]
383
384 *) Add the VC-CE target, introduce the WINCE sysname, and add
385 INSTALL.WCE and appropriate conditionals to make it build.
386 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
387
388 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
389 cygssl-x.y.z.dll, where x, y and z are the major, minor and
390 edit numbers of the version.
391 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
392
393 *) Introduce safe string copy and catenation functions
394 (BUF_strlcpy() and BUF_strlcat()).
395 [Ben Laurie (CHATS) and Richard Levitte]
396
397 *) Avoid using fixed-size buffers for one-line DNs.
398 [Ben Laurie (CHATS)]
399
400 *) Add BUF_MEM_grow_clean() to avoid information leakage when
401 resizing buffers containing secrets, and use where appropriate.
402 [Ben Laurie (CHATS)]
403
404 *) Avoid using fixed size buffers for configuration file location.
405 [Ben Laurie (CHATS)]
406
407 *) Avoid filename truncation for various CA files.
408 [Ben Laurie (CHATS)]
409
410 *) Use sizeof in preference to magic numbers.
411 [Ben Laurie (CHATS)]
412
413 *) Avoid filename truncation in cert requests.
414 [Ben Laurie (CHATS)]
415
416 *) Add assertions to check for (supposedly impossible) buffer
417 overflows.
418 [Ben Laurie (CHATS)]
419
420 *) Don't cache truncated DNS entries in the local cache (this could
421 potentially lead to a spoofing attack).
422 [Ben Laurie (CHATS)]
423
424 *) Fix various buffers to be large enough for hex/decimal
425 representations in a platform independent manner.
426 [Ben Laurie (CHATS)]
427
428 *) Add CRYPTO_realloc_clean() to avoid information leakage when
429 resizing buffers containing secrets, and use where appropriate.
430 [Ben Laurie (CHATS)]
431
432 *) Add BIO_indent() to avoid much slightly worrying code to do
433 indents.
434 [Ben Laurie (CHATS)]
435
436 *) Convert sprintf()/BIO_puts() to BIO_printf().
437 [Ben Laurie (CHATS)]
438
439 *) buffer_gets() could terminate with the buffer only half
440 full. Fixed.
441 [Ben Laurie (CHATS)]
442
443 *) Add assertions to prevent user-supplied crypto functions from
444 overflowing internal buffers by having large block sizes, etc.
445 [Ben Laurie (CHATS)]
446
447 *) New OPENSSL_assert() macro (similar to assert(), but enabled
448 unconditionally).
449 [Ben Laurie (CHATS)]
450
451 *) Eliminate unused copy of key in RC4.
452 [Ben Laurie (CHATS)]
453
454 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
455 [Ben Laurie (CHATS)]
456
457 *) Fix off-by-one error in EGD path.
458 [Ben Laurie (CHATS)]
459
460 *) If RANDFILE path is too long, ignore instead of truncating.
461 [Ben Laurie (CHATS)]
462
463 *) Eliminate unused and incorrectly sized X.509 structure
464 CBCParameter.
465 [Ben Laurie (CHATS)]
466
467 *) Eliminate unused and dangerous function knumber().
468 [Ben Laurie (CHATS)]
469
470 *) Eliminate unused and dangerous structure, KSSL_ERR.
471 [Ben Laurie (CHATS)]
472
473 *) Protect against overlong session ID context length in an encoded
474 session object. Since these are local, this does not appear to be
475 exploitable.
476 [Ben Laurie (CHATS)]
477
478 *) Change from security patch (see 0.9.6e below) that did not affect
479 the 0.9.6 release series:
480
481 Remote buffer overflow in SSL3 protocol - an attacker could
482 supply an oversized master key in Kerberos-enabled versions.
483 (CAN-2002-0657)
484 [Ben Laurie (CHATS)]
485
486 *) Change the SSL kerb5 codes to match RFC 2712.
487 [Richard Levitte]
488
489 *) Make -nameopt work fully for req and add -reqopt switch.
490 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
491
492 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
493 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
494
495 *) Make sure tests can be performed even if the corresponding algorithms
496 have been removed entirely. This was also the last step to make
497 OpenSSL compilable with DJGPP under all reasonable conditions.
498 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
499
500 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
501 to allow version independent disabling of normally unselected ciphers,
502 which may be activated as a side-effect of selecting a single cipher.
503
504 (E.g., cipher list string "RSA" enables ciphersuites that are left
505 out of "ALL" because they do not provide symmetric encryption.
506 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
507 [Lutz Jaenicke, Bodo Moeller]
508
509 *) Add appropriate support for separate platform-dependent build
510 directories. The recommended way to make a platform-dependent
511 build directory is the following (tested on Linux), maybe with
512 some local tweaks:
513
514 # Place yourself outside of the OpenSSL source tree. In
515 # this example, the environment variable OPENSSL_SOURCE
516 # is assumed to contain the absolute OpenSSL source directory.
517 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
518 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
519 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
520 mkdir -p `dirname $F`
521 ln -s $OPENSSL_SOURCE/$F $F
522 done
523
524 To be absolutely sure not to disturb the source tree, a "make clean"
525 is a good thing. If it isn't successfull, don't worry about it,
526 it probably means the source directory is very clean.
527 [Richard Levitte]
528
529 *) Make sure any ENGINE control commands make local copies of string
530 pointers passed to them whenever necessary. Otherwise it is possible
531 the caller may have overwritten (or deallocated) the original string
532 data when a later ENGINE operation tries to use the stored values.
533 [Götz Babin-Ebell <babinebell@trustcenter.de>]
534
535 *) Improve diagnostics in file reading and command-line digests.
536 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
537
538 *) Add AES modes CFB and OFB to the object database. Correct an
539 error in AES-CFB decryption.
540 [Richard Levitte]
541
542 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
543 allows existing EVP_CIPHER_CTX structures to be reused after
544 calling EVP_*Final(). This behaviour is used by encryption
545 BIOs and some applications. This has the side effect that
546 applications must explicitly clean up cipher contexts with
547 EVP_CIPHER_CTX_cleanup() or they will leak memory.
548 [Steve Henson]
549
550 *) Check the values of dna and dnb in bn_mul_recursive before calling
551 bn_mul_comba (a non zero value means the a or b arrays do not contain
552 n2 elements) and fallback to bn_mul_normal if either is not zero.
553 [Steve Henson]
554
555 *) Fix escaping of non-ASCII characters when using the -subj option
556 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
557 [Lutz Jaenicke]
558
559 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
560 form for "surname", serialNumber has no short form.
561 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
562 therefore remove "mail" short name for "internet 7".
563 The OID for unique identifiers in X509 certificates is
564 x500UniqueIdentifier, not uniqueIdentifier.
565 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
566 [Lutz Jaenicke]
567
568 *) Add an "init" command to the ENGINE config module and auto initialize
569 ENGINEs. Without any "init" command the ENGINE will be initialized
570 after all ctrl commands have been executed on it. If init=1 the
571 ENGINE is initailized at that point (ctrls before that point are run
572 on the uninitialized ENGINE and after on the initialized one). If
573 init=0 then the ENGINE will not be iniatialized at all.
574 [Steve Henson]
575
576 *) Fix the 'app_verify_callback' interface so that the user-defined
577 argument is actually passed to the callback: In the
578 SSL_CTX_set_cert_verify_callback() prototype, the callback
579 declaration has been changed from
580 int (*cb)()
581 into
582 int (*cb)(X509_STORE_CTX *,void *);
583 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
584 i=s->ctx->app_verify_callback(&ctx)
585 has been changed into
586 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
587
588 To update applications using SSL_CTX_set_cert_verify_callback(),
589 a dummy argument can be added to their callback functions.
590 [D. K. Smetters <smetters@parc.xerox.com>]
591
592 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
593 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
594
595 *) Add and OPENSSL_LOAD_CONF define which will cause
596 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
597 This allows older applications to transparently support certain
598 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
599 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
600 load the config file and OPENSSL_add_all_algorithms_conf() which will
601 always load it have also been added.
602 [Steve Henson]
603
604 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
605 Adjust NIDs and EVP layer.
606 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
607
608 *) Config modules support in openssl utility.
609
610 Most commands now load modules from the config file,
611 though in a few (such as version) this isn't done
612 because it couldn't be used for anything.
613
614 In the case of ca and req the config file used is
615 the same as the utility itself: that is the -config
616 command line option can be used to specify an
617 alternative file.
618 [Steve Henson]
619
620 *) Move default behaviour from OPENSSL_config(). If appname is NULL
621 use "openssl_conf" if filename is NULL use default openssl config file.
622 [Steve Henson]
623
624 *) Add an argument to OPENSSL_config() to allow the use of an alternative
625 config section name. Add a new flag to tolerate a missing config file
626 and move code to CONF_modules_load_file().
627 [Steve Henson]
628
629 *) Support for crypto accelerator cards from Accelerated Encryption
630 Processing, www.aep.ie. (Use engine 'aep')
631 The support was copied from 0.9.6c [engine] and adapted/corrected
632 to work with the new engine framework.
633 [AEP Inc. and Richard Levitte]
634
635 *) Support for SureWare crypto accelerator cards from Baltimore
636 Technologies. (Use engine 'sureware')
637 The support was copied from 0.9.6c [engine] and adapted
638 to work with the new engine framework.
639 [Richard Levitte]
640
641 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
642 make the newer ENGINE framework commands for the CHIL engine work.
643 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
644
645 *) Make it possible to produce shared libraries on ReliantUNIX.
646 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
647
648 *) Add the configuration target debug-linux-ppro.
649 Make 'openssl rsa' use the general key loading routines
650 implemented in apps.c, and make those routines able to
651 handle the key format FORMAT_NETSCAPE and the variant
652 FORMAT_IISSGC.
653 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
654
655 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
656 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
657
658 *) Add -keyform to rsautl, and document -engine.
659 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
660
661 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
662 BIO_R_NO_SUCH_FILE error code rather than the generic
663 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
664 [Ben Laurie]
665
666 *) Add new functions
667 ERR_peek_last_error
668 ERR_peek_last_error_line
669 ERR_peek_last_error_line_data.
670 These are similar to
671 ERR_peek_error
672 ERR_peek_error_line
673 ERR_peek_error_line_data,
674 but report on the latest error recorded rather than the first one
675 still in the error queue.
676 [Ben Laurie, Bodo Moeller]
677
678 *) default_algorithms option in ENGINE config module. This allows things
679 like:
680 default_algorithms = ALL
681 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
682 [Steve Henson]
683
684 *) Prelminary ENGINE config module.
685 [Steve Henson]
686
687 *) New experimental application configuration code.
688 [Steve Henson]
689
690 *) Change the AES code to follow the same name structure as all other
691 symmetric ciphers, and behave the same way. Move everything to
692 the directory crypto/aes, thereby obsoleting crypto/rijndael.
693 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
694
695 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
696 [Ben Laurie and Theo de Raadt]
697
698 *) Add option to output public keys in req command.
699 [Massimiliano Pala madwolf@openca.org]
700
701 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
702 (up to about 10% better than before for P-192 and P-224).
703 [Bodo Moeller]
704
705 *) New functions/macros
706
707 SSL_CTX_set_msg_callback(ctx, cb)
708 SSL_CTX_set_msg_callback_arg(ctx, arg)
709 SSL_set_msg_callback(ssl, cb)
710 SSL_set_msg_callback_arg(ssl, arg)
711
712 to request calling a callback function
713
714 void cb(int write_p, int version, int content_type,
715 const void *buf, size_t len, SSL *ssl, void *arg)
716
717 whenever a protocol message has been completely received
718 (write_p == 0) or sent (write_p == 1). Here 'version' is the
719 protocol version according to which the SSL library interprets
720 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
721 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
722 the content type as defined in the SSL 3.0/TLS 1.0 protocol
723 specification (change_cipher_spec(20), alert(21), handshake(22)).
724 'buf' and 'len' point to the actual message, 'ssl' to the
725 SSL object, and 'arg' is the application-defined value set by
726 SSL[_CTX]_set_msg_callback_arg().
727
728 'openssl s_client' and 'openssl s_server' have new '-msg' options
729 to enable a callback that displays all protocol messages.
730 [Bodo Moeller]
731
732 *) Change the shared library support so shared libraries are built as
733 soon as the corresponding static library is finished, and thereby get
734 openssl and the test programs linked against the shared library.
735 This still only happens when the keyword "shard" has been given to
736 the configuration scripts.
737
738 NOTE: shared library support is still an experimental thing, and
739 backward binary compatibility is still not guaranteed.
740 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
741
742 *) Add support for Subject Information Access extension.
743 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
744
745 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
746 additional bytes when new memory had to be allocated, not just
747 when reusing an existing buffer.
748 [Bodo Moeller]
749
750 *) New command line and configuration option 'utf8' for the req command.
751 This allows field values to be specified as UTF8 strings.
752 [Steve Henson]
753
754 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
755 runs for the former and machine-readable output for the latter.
756 [Ben Laurie]
757
758 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
759 of the e-mail address in the DN (i.e., it will go into a certificate
760 extension only). The new configuration file option 'email_in_dn = no'
761 has the same effect.
762 [Massimiliano Pala madwolf@openca.org]
763
764 *) Change all functions with names starting with des_ to be starting
765 with DES_ instead. Add wrappers that are compatible with libdes,
766 but are named _ossl_old_des_*. Finally, add macros that map the
767 des_* symbols to the corresponding _ossl_old_des_* if libdes
768 compatibility is desired. If OpenSSL 0.9.6c compatibility is
769 desired, the des_* symbols will be mapped to DES_*, with one
770 exception.
771
772 Since we provide two compatibility mappings, the user needs to
773 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
774 compatibility is desired. The default (i.e., when that macro
775 isn't defined) is OpenSSL 0.9.6c compatibility.
776
777 There are also macros that enable and disable the support of old
778 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
779 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
780 are defined, the default will apply: to support the old des routines.
781
782 In either case, one must include openssl/des.h to get the correct
783 definitions. Do not try to just include openssl/des_old.h, that
784 won't work.
785
786 NOTE: This is a major break of an old API into a new one. Software
787 authors are encouraged to switch to the DES_ style functions. Some
788 time in the future, des_old.h and the libdes compatibility functions
789 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
790 default), and then completely removed.
791 [Richard Levitte]
792
793 *) Test for certificates which contain unsupported critical extensions.
794 If such a certificate is found during a verify operation it is
795 rejected by default: this behaviour can be overridden by either
796 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
797 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
798 X509_supported_extension() has also been added which returns 1 if a
799 particular extension is supported.
800 [Steve Henson]
801
802 *) Modify the behaviour of EVP cipher functions in similar way to digests
803 to retain compatibility with existing code.
804 [Steve Henson]
805
806 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
807 compatibility with existing code. In particular the 'ctx' parameter does
808 not have to be to be initialized before the call to EVP_DigestInit() and
809 it is tidied up after a call to EVP_DigestFinal(). New function
810 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
811 EVP_MD_CTX_copy() changed to not require the destination to be
812 initialized valid and new function EVP_MD_CTX_copy_ex() added which
813 requires the destination to be valid.
814
815 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
816 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
817 [Steve Henson]
818
819 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
820 so that complete 'Handshake' protocol structures are kept in memory
821 instead of overwriting 'msg_type' and 'length' with 'body' data.
822 [Bodo Moeller]
823
824 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
825 [Massimo Santin via Richard Levitte]
826
827 *) Major restructuring to the underlying ENGINE code. This includes
828 reduction of linker bloat, separation of pure "ENGINE" manipulation
829 (initialisation, etc) from functionality dealing with implementations
830 of specific crypto iterfaces. This change also introduces integrated
831 support for symmetric ciphers and digest implementations - so ENGINEs
832 can now accelerate these by providing EVP_CIPHER and EVP_MD
833 implementations of their own. This is detailed in crypto/engine/README
834 as it couldn't be adequately described here. However, there are a few
835 API changes worth noting - some RSA, DSA, DH, and RAND functions that
836 were changed in the original introduction of ENGINE code have now
837 reverted back - the hooking from this code to ENGINE is now a good
838 deal more passive and at run-time, operations deal directly with
839 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
840 dereferencing through an ENGINE pointer any more. Also, the ENGINE
841 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
842 they were not being used by the framework as there is no concept of a
843 BIGNUM_METHOD and they could not be generalised to the new
844 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
845 ENGINE_cpy() has been removed as it cannot be consistently defined in
846 the new code.
847 [Geoff Thorpe]
848
849 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
850 [Steve Henson]
851
852 *) Change mkdef.pl to sort symbols that get the same entry number,
853 and make sure the automatically generated functions ERR_load_*
854 become part of libeay.num as well.
855 [Richard Levitte]
856
857 *) New function SSL_renegotiate_pending(). This returns true once
858 renegotiation has been requested (either SSL_renegotiate() call
859 or HelloRequest/ClientHello receveived from the peer) and becomes
860 false once a handshake has been completed.
861 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
862 sends a HelloRequest, but does not ensure that a handshake takes
863 place. SSL_renegotiate_pending() is useful for checking if the
864 client has followed the request.)
865 [Bodo Moeller]
866
867 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
868 By default, clients may request session resumption even during
869 renegotiation (if session ID contexts permit); with this option,
870 session resumption is possible only in the first handshake.
871
872 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
873 more bits available for options that should not be part of
874 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
875 [Bodo Moeller]
876
877 *) Add some demos for certificate and certificate request creation.
878 [Steve Henson]
879
880 *) Make maximum certificate chain size accepted from the peer application
881 settable (SSL*_get/set_max_cert_list()), as proposed by
882 "Douglas E. Engert" <deengert@anl.gov>.
883 [Lutz Jaenicke]
884
885 *) Add support for shared libraries for Unixware-7
886 (Boyd Lynn Gerber <gerberb@zenez.com>).
887 [Lutz Jaenicke]
888
889 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
890 be done prior to destruction. Use this to unload error strings from
891 ENGINEs that load their own error strings. NB: This adds two new API
892 functions to "get" and "set" this destroy handler in an ENGINE.
893 [Geoff Thorpe]
894
895 *) Alter all existing ENGINE implementations (except "openssl" and
896 "openbsd") to dynamically instantiate their own error strings. This
897 makes them more flexible to be built both as statically-linked ENGINEs
898 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
899 Also, add stub code to each that makes building them as self-contained
900 shared-libraries easier (see README.ENGINE).
901 [Geoff Thorpe]
902
903 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
904 implementations into applications that are completely implemented in
905 self-contained shared-libraries. The "dynamic" ENGINE exposes control
906 commands that can be used to configure what shared-library to load and
907 to control aspects of the way it is handled. Also, made an update to
908 the README.ENGINE file that brings its information up-to-date and
909 provides some information and instructions on the "dynamic" ENGINE
910 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
911 [Geoff Thorpe]
912
913 *) Make it possible to unload ranges of ERR strings with a new
914 "ERR_unload_strings" function.
915 [Geoff Thorpe]
916
917 *) Add a copy() function to EVP_MD.
918 [Ben Laurie]
919
920 *) Make EVP_MD routines take a context pointer instead of just the
921 md_data void pointer.
922 [Ben Laurie]
923
924 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
925 that the digest can only process a single chunk of data
926 (typically because it is provided by a piece of
927 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
928 is only going to provide a single chunk of data, and hence the
929 framework needn't accumulate the data for oneshot drivers.
930 [Ben Laurie]
931
932 *) As with "ERR", make it possible to replace the underlying "ex_data"
933 functions. This change also alters the storage and management of global
934 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
935 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
936 index counters. The API functions that use this state have been changed
937 to take a "class_index" rather than pointers to the class's local STACK
938 and counter, and there is now an API function to dynamically create new
939 classes. This centralisation allows us to (a) plug a lot of the
940 thread-safety problems that existed, and (b) makes it possible to clean
941 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
942 such data would previously have always leaked in application code and
943 workarounds were in place to make the memory debugging turn a blind eye
944 to it. Application code that doesn't use this new function will still
945 leak as before, but their memory debugging output will announce it now
946 rather than letting it slide.
947
948 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
949 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
950 has a return value to indicate success or failure.
951 [Geoff Thorpe]
952
953 *) Make it possible to replace the underlying "ERR" functions such that the
954 global state (2 LHASH tables and 2 locks) is only used by the "default"
955 implementation. This change also adds two functions to "get" and "set"
956 the implementation prior to it being automatically set the first time
957 any other ERR function takes place. Ie. an application can call "get",
958 pass the return value to a module it has just loaded, and that module
959 can call its own "set" function using that value. This means the
960 module's "ERR" operations will use (and modify) the error state in the
961 application and not in its own statically linked copy of OpenSSL code.
962 [Geoff Thorpe]
963
964 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
965 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
966 the operation, and provides a more encapsulated way for external code
967 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
968 to use these functions rather than manually incrementing the counts.
969
970 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
971 [Geoff Thorpe]
972
973 *) Add EVP test program.
974 [Ben Laurie]
975
976 *) Add symmetric cipher support to ENGINE. Expect the API to change!
977 [Ben Laurie]
978
979 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
980 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
981 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
982 These allow a CRL to be built without having to access X509_CRL fields
983 directly. Modify 'ca' application to use new functions.
984 [Steve Henson]
985
986 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
987 bug workarounds. Rollback attack detection is a security feature.
988 The problem will only arise on OpenSSL servers when TLSv1 is not
989 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
990 Software authors not wanting to support TLSv1 will have special reasons
991 for their choice and can explicitly enable this option.
992 [Bodo Moeller, Lutz Jaenicke]
993
994 *) Rationalise EVP so it can be extended: don't include a union of
995 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
996 (similar to those existing for EVP_CIPHER_CTX).
997 Usage example:
998
999 EVP_MD_CTX md;
1000
1001 EVP_MD_CTX_init(&md); /* new function call */
1002 EVP_DigestInit(&md, EVP_sha1());
1003 EVP_DigestUpdate(&md, in, len);
1004 EVP_DigestFinal(&md, out, NULL);
1005 EVP_MD_CTX_cleanup(&md); /* new function call */
1006
1007 [Ben Laurie]
1008
1009 *) Make DES key schedule conform to the usual scheme, as well as
1010 correcting its structure. This means that calls to DES functions
1011 now have to pass a pointer to a des_key_schedule instead of a
1012 plain des_key_schedule (which was actually always a pointer
1013 anyway): E.g.,
1014
1015 des_key_schedule ks;
1016
1017 des_set_key_checked(..., &ks);
1018 des_ncbc_encrypt(..., &ks, ...);
1019
1020 (Note that a later change renames 'des_...' into 'DES_...'.)
1021 [Ben Laurie]
1022
1023 *) Initial reduction of linker bloat: the use of some functions, such as
1024 PEM causes large amounts of unused functions to be linked in due to
1025 poor organisation. For example pem_all.c contains every PEM function
1026 which has a knock on effect of linking in large amounts of (unused)
1027 ASN1 code. Grouping together similar functions and splitting unrelated
1028 functions prevents this.
1029 [Steve Henson]
1030
1031 *) Cleanup of EVP macros.
1032 [Ben Laurie]
1033
1034 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1035 correct _ecb suffix.
1036 [Ben Laurie]
1037
1038 *) Add initial OCSP responder support to ocsp application. The
1039 revocation information is handled using the text based index
1040 use by the ca application. The responder can either handle
1041 requests generated internally, supplied in files (for example
1042 via a CGI script) or using an internal minimal server.
1043 [Steve Henson]
1044
1045 *) Add configuration choices to get zlib compression for TLS.
1046 [Richard Levitte]
1047
1048 *) Changes to Kerberos SSL for RFC 2712 compliance:
1049 1. Implemented real KerberosWrapper, instead of just using
1050 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1051 2. Implemented optional authenticator field of KerberosWrapper.
1052
1053 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1054 and authenticator structs; see crypto/krb5/.
1055
1056 Generalized Kerberos calls to support multiple Kerberos libraries.
1057 [Vern Staats <staatsvr@asc.hpc.mil>,
1058 Jeffrey Altman <jaltman@columbia.edu>
1059 via Richard Levitte]
1060
1061 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1062 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1063 values for each of the key sizes rather than having just
1064 parameters (and 'speed' generating keys each time).
1065 [Geoff Thorpe]
1066
1067 *) Speed up EVP routines.
1068 Before:
1069 encrypt
1070 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1071 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1072 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1073 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1074 decrypt
1075 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1076 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1077 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1078 After:
1079 encrypt
1080 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1081 decrypt
1082 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1083 [Ben Laurie]
1084
1085 *) Added the OS2-EMX target.
1086 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1087
1088 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1089 to support NCONF routines in extension code. New function CONF_set_nconf()
1090 to allow functions which take an NCONF to also handle the old LHASH
1091 structure: this means that the old CONF compatible routines can be
1092 retained (in particular wrt extensions) without having to duplicate the
1093 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1094 [Steve Henson]
1095
1096 *) Enhance the general user interface with mechanisms for inner control
1097 and with possibilities to have yes/no kind of prompts.
1098 [Richard Levitte]
1099
1100 *) Change all calls to low level digest routines in the library and
1101 applications to use EVP. Add missing calls to HMAC_cleanup() and
1102 don't assume HMAC_CTX can be copied using memcpy().
1103 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1104
1105 *) Add the possibility to control engines through control names but with
1106 arbitrary arguments instead of just a string.
1107 Change the key loaders to take a UI_METHOD instead of a callback
1108 function pointer. NOTE: this breaks binary compatibility with earlier
1109 versions of OpenSSL [engine].
1110 Adapt the nCipher code for these new conditions and add a card insertion
1111 callback.
1112 [Richard Levitte]
1113
1114 *) Enhance the general user interface with mechanisms to better support
1115 dialog box interfaces, application-defined prompts, the possibility
1116 to use defaults (for example default passwords from somewhere else)
1117 and interrupts/cancellations.
1118 [Richard Levitte]
1119
1120 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1121 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1122 [Steve Henson]
1123
1124 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1125 tidy up some unnecessarily weird code in 'sk_new()').
1126 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1127
1128 *) Change the key loading routines for ENGINEs to use the same kind
1129 callback (pem_password_cb) as all other routines that need this
1130 kind of callback.
1131 [Richard Levitte]
1132
1133 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1134 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1135 than this minimum value is recommended.
1136 [Lutz Jaenicke]
1137
1138 *) New random seeder for OpenVMS, using the system process statistics
1139 that are easily reachable.
1140 [Richard Levitte]
1141
1142 *) Windows apparently can't transparently handle global
1143 variables defined in DLLs. Initialisations such as:
1144
1145 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1146
1147 wont compile. This is used by the any applications that need to
1148 declare their own ASN1 modules. This was fixed by adding the option
1149 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1150 needed for static libraries under Win32.
1151 [Steve Henson]
1152
1153 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1154 setting of purpose and trust fields. New X509_STORE trust and
1155 purpose functions and tidy up setting in other SSL functions.
1156 [Steve Henson]
1157
1158 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1159 structure. These are inherited by X509_STORE_CTX when it is
1160 initialised. This allows various defaults to be set in the
1161 X509_STORE structure (such as flags for CRL checking and custom
1162 purpose or trust settings) for functions which only use X509_STORE_CTX
1163 internally such as S/MIME.
1164
1165 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1166 trust settings if they are not set in X509_STORE. This allows X509_STORE
1167 purposes and trust (in S/MIME for example) to override any set by default.
1168
1169 Add command line options for CRL checking to smime, s_client and s_server
1170 applications.
1171 [Steve Henson]
1172
1173 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1174 are set then the CRL is looked up in the X509_STORE structure and
1175 its validity and signature checked, then if the certificate is found
1176 in the CRL the verify fails with a revoked error.
1177
1178 Various new CRL related callbacks added to X509_STORE_CTX structure.
1179
1180 Command line options added to 'verify' application to support this.
1181
1182 This needs some additional work, such as being able to handle multiple
1183 CRLs with different times, extension based lookup (rather than just
1184 by subject name) and ultimately more complete V2 CRL extension
1185 handling.
1186 [Steve Henson]
1187
1188 *) Add a general user interface API (crypto/ui/). This is designed
1189 to replace things like des_read_password and friends (backward
1190 compatibility functions using this new API are provided).
1191 The purpose is to remove prompting functions from the DES code
1192 section as well as provide for prompting through dialog boxes in
1193 a window system and the like.
1194 [Richard Levitte]
1195
1196 *) Add "ex_data" support to ENGINE so implementations can add state at a
1197 per-structure level rather than having to store it globally.
1198 [Geoff]
1199
1200 *) Make it possible for ENGINE structures to be copied when retrieved by
1201 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1202 This causes the "original" ENGINE structure to act like a template,
1203 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1204 operational state can be localised to each ENGINE structure, despite the
1205 fact they all share the same "methods". New ENGINE structures returned in
1206 this case have no functional references and the return value is the single
1207 structural reference. This matches the single structural reference returned
1208 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1209 ENGINE structure.
1210 [Geoff]
1211
1212 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1213 needs to match any other type at all we need to manually clear the
1214 tag cache.
1215 [Steve Henson]
1216
1217 *) Changes to the "openssl engine" utility to include;
1218 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1219 about an ENGINE's available control commands.
1220 - executing control commands from command line arguments using the
1221 '-pre' and '-post' switches. '-post' is only used if '-t' is
1222 specified and the ENGINE is successfully initialised. The syntax for
1223 the individual commands are colon-separated, for example;
1224 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1225 [Geoff]
1226
1227 *) New dynamic control command support for ENGINEs. ENGINEs can now
1228 declare their own commands (numbers), names (strings), descriptions,
1229 and input types for run-time discovery by calling applications. A
1230 subset of these commands are implicitly classed as "executable"
1231 depending on their input type, and only these can be invoked through
1232 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1233 can be based on user input, config files, etc). The distinction is
1234 that "executable" commands cannot return anything other than a boolean
1235 result and can only support numeric or string input, whereas some
1236 discoverable commands may only be for direct use through
1237 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1238 pointers, or other custom uses. The "executable" commands are to
1239 support parameterisations of ENGINE behaviour that can be
1240 unambiguously defined by ENGINEs and used consistently across any
1241 OpenSSL-based application. Commands have been added to all the
1242 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1243 control over shared-library paths without source code alterations.
1244 [Geoff]
1245
1246 *) Changed all ENGINE implementations to dynamically allocate their
1247 ENGINEs rather than declaring them statically. Apart from this being
1248 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1249 this also allows the implementations to compile without using the
1250 internal engine_int.h header.
1251 [Geoff]
1252
1253 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1254 'const' value. Any code that should be able to modify a RAND_METHOD
1255 should already have non-const pointers to it (ie. they should only
1256 modify their own ones).
1257 [Geoff]
1258
1259 *) Made a variety of little tweaks to the ENGINE code.
1260 - "atalla" and "ubsec" string definitions were moved from header files
1261 to C code. "nuron" string definitions were placed in variables
1262 rather than hard-coded - allowing parameterisation of these values
1263 later on via ctrl() commands.
1264 - Removed unused "#if 0"'d code.
1265 - Fixed engine list iteration code so it uses ENGINE_free() to release
1266 structural references.
1267 - Constified the RAND_METHOD element of ENGINE structures.
1268 - Constified various get/set functions as appropriate and added
1269 missing functions (including a catch-all ENGINE_cpy that duplicates
1270 all ENGINE values onto a new ENGINE except reference counts/state).
1271 - Removed NULL parameter checks in get/set functions. Setting a method
1272 or function to NULL is a way of cancelling out a previously set
1273 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1274 and doesn't justify the extra error symbols and code.
1275 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1276 flags from engine_int.h to engine.h.
1277 - Changed prototypes for ENGINE handler functions (init(), finish(),
1278 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1279 [Geoff]
1280
1281 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1282 to the algorithm using long division. The binary algorithm can be
1283 used only if the modulus is odd. On 32-bit systems, it is faster
1284 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1285 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1286 up to 450 bits. In 64-bit environments, the binary algorithm
1287 appears to be advantageous for much longer moduli; here we use it
1288 for moduli up to 2048 bits.
1289 [Bodo Moeller]
1290
1291 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1292 could not support the combine flag in choice fields.
1293 [Steve Henson]
1294
1295 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1296 extensions from a certificate request to the certificate.
1297 [Steve Henson]
1298
1299 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1300 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1301 file: this allows the display of the certificate about to be
1302 signed to be customised, to allow certain fields to be included
1303 or excluded and extension details. The old system didn't display
1304 multicharacter strings properly, omitted fields not in the policy
1305 and couldn't display additional details such as extensions.
1306 [Steve Henson]
1307
1308 *) Function EC_POINTs_mul for multiple scalar multiplication
1309 of an arbitrary number of elliptic curve points
1310 \sum scalars[i]*points[i],
1311 optionally including the generator defined for the EC_GROUP:
1312 scalar*generator + \sum scalars[i]*points[i].
1313
1314 EC_POINT_mul is a simple wrapper function for the typical case
1315 that the point list has just one item (besides the optional
1316 generator).
1317 [Bodo Moeller]
1318
1319 *) First EC_METHODs for curves over GF(p):
1320
1321 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1322 operations and provides various method functions that can also
1323 operate with faster implementations of modular arithmetic.
1324
1325 EC_GFp_mont_method() reuses most functions that are part of
1326 EC_GFp_simple_method, but uses Montgomery arithmetic.
1327
1328 [Bodo Moeller; point addition and point doubling
1329 implementation directly derived from source code provided by
1330 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1331
1332 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1333 crypto/ec/ec_lib.c):
1334
1335 Curves are EC_GROUP objects (with an optional group generator)
1336 based on EC_METHODs that are built into the library.
1337
1338 Points are EC_POINT objects based on EC_GROUP objects.
1339
1340 Most of the framework would be able to handle curves over arbitrary
1341 finite fields, but as there are no obvious types for fields other
1342 than GF(p), some functions are limited to that for now.
1343 [Bodo Moeller]
1344
1345 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1346 that the file contains a complete HTTP response.
1347 [Richard Levitte]
1348
1349 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1350 change the def and num file printf format specifier from "%-40sXXX"
1351 to "%-39s XXX". The latter will always guarantee a space after the
1352 field while the former will cause them to run together if the field
1353 is 40 of more characters long.
1354 [Steve Henson]
1355
1356 *) Constify the cipher and digest 'method' functions and structures
1357 and modify related functions to take constant EVP_MD and EVP_CIPHER
1358 pointers.
1359 [Steve Henson]
1360
1361 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1362 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1363 [Bodo Moeller]
1364
1365 *) Modify EVP_Digest*() routines so they now return values. Although the
1366 internal software routines can never fail additional hardware versions
1367 might.
1368 [Steve Henson]
1369
1370 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1371
1372 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1373 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1374
1375 ASN1 error codes
1376 ERR_R_NESTED_ASN1_ERROR
1377 ...
1378 ERR_R_MISSING_ASN1_EOS
1379 were 4 .. 9, conflicting with
1380 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1381 ...
1382 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1383 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1384
1385 Add new error code 'ERR_R_INTERNAL_ERROR'.
1386 [Bodo Moeller]
1387
1388 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1389 suffices.
1390 [Bodo Moeller]
1391
1392 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1393 sets the subject name for a new request or supersedes the
1394 subject name in a given request. Formats that can be parsed are
1395 'CN=Some Name, OU=myOU, C=IT'
1396 and
1397 'CN=Some Name/OU=myOU/C=IT'.
1398
1399 Add options '-batch' and '-verbose' to 'openssl req'.
1400 [Massimiliano Pala <madwolf@hackmasters.net>]
1401
1402 *) Introduce the possibility to access global variables through
1403 functions on platform were that's the best way to handle exporting
1404 global variables in shared libraries. To enable this functionality,
1405 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1406 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1407 is normally done by Configure or something similar).
1408
1409 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1410 in the source file (foo.c) like this:
1411
1412 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1413 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1414
1415 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1416 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1417
1418 OPENSSL_DECLARE_GLOBAL(int,foo);
1419 #define foo OPENSSL_GLOBAL_REF(foo)
1420 OPENSSL_DECLARE_GLOBAL(double,bar);
1421 #define bar OPENSSL_GLOBAL_REF(bar)
1422
1423 The #defines are very important, and therefore so is including the
1424 header file everywhere where the defined globals are used.
1425
1426 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1427 of ASN.1 items, but that structure is a bit different.
1428
1429 The largest change is in util/mkdef.pl which has been enhanced with
1430 better and easier to understand logic to choose which symbols should
1431 go into the Windows .def files as well as a number of fixes and code
1432 cleanup (among others, algorithm keywords are now sorted
1433 lexicographically to avoid constant rewrites).
1434 [Richard Levitte]
1435
1436 *) In BN_div() keep a copy of the sign of 'num' before writing the
1437 result to 'rm' because if rm==num the value will be overwritten
1438 and produce the wrong result if 'num' is negative: this caused
1439 problems with BN_mod() and BN_nnmod().
1440 [Steve Henson]
1441
1442 *) Function OCSP_request_verify(). This checks the signature on an
1443 OCSP request and verifies the signer certificate. The signer
1444 certificate is just checked for a generic purpose and OCSP request
1445 trust settings.
1446 [Steve Henson]
1447
1448 *) Add OCSP_check_validity() function to check the validity of OCSP
1449 responses. OCSP responses are prepared in real time and may only
1450 be a few seconds old. Simply checking that the current time lies
1451 between thisUpdate and nextUpdate max reject otherwise valid responses
1452 caused by either OCSP responder or client clock inaccuracy. Instead
1453 we allow thisUpdate and nextUpdate to fall within a certain period of
1454 the current time. The age of the response can also optionally be
1455 checked. Two new options -validity_period and -status_age added to
1456 ocsp utility.
1457 [Steve Henson]
1458
1459 *) If signature or public key algorithm is unrecognized print out its
1460 OID rather that just UNKNOWN.
1461 [Steve Henson]
1462
1463 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1464 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1465 ID to be generated from the issuer certificate alone which can then be
1466 passed to OCSP_id_issuer_cmp().
1467 [Steve Henson]
1468
1469 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1470 ASN1 modules to export functions returning ASN1_ITEM pointers
1471 instead of the ASN1_ITEM structures themselves. This adds several
1472 new macros which allow the underlying ASN1 function/structure to
1473 be accessed transparently. As a result code should not use ASN1_ITEM
1474 references directly (such as &X509_it) but instead use the relevant
1475 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1476 use of the new ASN1 code on platforms where exporting structures
1477 is problematical (for example in shared libraries) but exporting
1478 functions returning pointers to structures is not.
1479 [Steve Henson]
1480
1481 *) Add support for overriding the generation of SSL/TLS session IDs.
1482 These callbacks can be registered either in an SSL_CTX or per SSL.
1483 The purpose of this is to allow applications to control, if they wish,
1484 the arbitrary values chosen for use as session IDs, particularly as it
1485 can be useful for session caching in multiple-server environments. A
1486 command-line switch for testing this (and any client code that wishes
1487 to use such a feature) has been added to "s_server".
1488 [Geoff Thorpe, Lutz Jaenicke]
1489
1490 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1491 of the form '#if defined(...) || defined(...) || ...' and
1492 '#if !defined(...) && !defined(...) && ...'. This also avoids
1493 the growing number of special cases it was previously handling.
1494 [Richard Levitte]
1495
1496 *) Make all configuration macros available for application by making
1497 sure they are available in opensslconf.h, by giving them names starting
1498 with "OPENSSL_" to avoid conflicts with other packages and by making
1499 sure e_os2.h will cover all platform-specific cases together with
1500 opensslconf.h.
1501 Additionally, it is now possible to define configuration/platform-
1502 specific names (called "system identities"). In the C code, these
1503 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1504 macro with the name beginning with "OPENSSL_SYS_", which is determined
1505 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1506 what is available.
1507 [Richard Levitte]
1508
1509 *) New option -set_serial to 'req' and 'x509' this allows the serial
1510 number to use to be specified on the command line. Previously self
1511 signed certificates were hard coded with serial number 0 and the
1512 CA options of 'x509' had to use a serial number in a file which was
1513 auto incremented.
1514 [Steve Henson]
1515
1516 *) New options to 'ca' utility to support V2 CRL entry extensions.
1517 Currently CRL reason, invalidity date and hold instruction are
1518 supported. Add new CRL extensions to V3 code and some new objects.
1519 [Steve Henson]
1520
1521 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1522 disable standard block padding (aka PKCS#5 padding) in the EVP
1523 API, which was previously mandatory. This means that the data is
1524 not padded in any way and so the total length much be a multiple
1525 of the block size, otherwise an error occurs.
1526 [Steve Henson]
1527
1528 *) Initial (incomplete) OCSP SSL support.
1529 [Steve Henson]
1530
1531 *) New function OCSP_parse_url(). This splits up a URL into its host,
1532 port and path components: primarily to parse OCSP URLs. New -url
1533 option to ocsp utility.
1534 [Steve Henson]
1535
1536 *) New nonce behavior. The return value of OCSP_check_nonce() now
1537 reflects the various checks performed. Applications can decide
1538 whether to tolerate certain situations such as an absent nonce
1539 in a response when one was present in a request: the ocsp application
1540 just prints out a warning. New function OCSP_add1_basic_nonce()
1541 this is to allow responders to include a nonce in a response even if
1542 the request is nonce-less.
1543 [Steve Henson]
1544
1545 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1546 skipped when using openssl x509 multiple times on a single input file,
1547 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1548 [Bodo Moeller]
1549
1550 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1551 set string type: to handle setting ASN1_TIME structures. Fix ca
1552 utility to correctly initialize revocation date of CRLs.
1553 [Steve Henson]
1554
1555 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1556 the clients preferred ciphersuites and rather use its own preferences.
1557 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1558 Internet Explorer by ensuring unchanged hash method during stepup.
1559 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1560 [Lutz Jaenicke]
1561
1562 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1563 to aes and add a new 'exist' option to print out symbols that don't
1564 appear to exist.
1565 [Steve Henson]
1566
1567 *) Additional options to ocsp utility to allow flags to be set and
1568 additional certificates supplied.
1569 [Steve Henson]
1570
1571 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1572 OCSP client a number of certificate to only verify the response
1573 signature against.
1574 [Richard Levitte]
1575
1576 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1577 handle the new API. Currently only ECB, CBC modes supported. Add new
1578 AES OIDs.
1579
1580 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1581 Encryption Standard (AES) Ciphersuites for Transport Layer
1582 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1583 not enabled by default and were not part of the "ALL" ciphersuite
1584 alias because they were not yet official; they could be
1585 explicitly requested by specifying the "AESdraft" ciphersuite
1586 group alias. In the final release of OpenSSL 0.9.7, the group
1587 alias is called "AES" and is part of "ALL".)
1588 [Ben Laurie, Steve Henson, Bodo Moeller]
1589
1590 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1591 request to response.
1592 [Steve Henson]
1593
1594 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1595 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1596 extract information from a certificate request. OCSP_response_create()
1597 creates a response and optionally adds a basic response structure.
1598 OCSP_basic_add1_status() adds a complete single response to a basic
1599 response and returns the OCSP_SINGLERESP structure just added (to allow
1600 extensions to be included for example). OCSP_basic_add1_cert() adds a
1601 certificate to a basic response and OCSP_basic_sign() signs a basic
1602 response with various flags. New helper functions ASN1_TIME_check()
1603 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1604 (converts ASN1_TIME to GeneralizedTime).
1605 [Steve Henson]
1606
1607 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1608 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1609 structure from a certificate. X509_pubkey_digest() digests the public_key
1610 contents: this is used in various key identifiers.
1611 [Steve Henson]
1612
1613 *) Make sk_sort() tolerate a NULL argument.
1614 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1615
1616 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1617 passed by the function are trusted implicitly. If any of them signed the
1618 response then it is assumed to be valid and is not verified.
1619 [Steve Henson]
1620
1621 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1622 to data. This was previously part of the PKCS7 ASN1 code. This
1623 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1624 [Steve Henson, reported by Kenneth R. Robinette
1625 <support@securenetterm.com>]
1626
1627 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1628 routines: without these tracing memory leaks is very painful.
1629 Fix leaks in PKCS12 and PKCS7 routines.
1630 [Steve Henson]
1631
1632 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1633 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1634 effectively meant GeneralizedTime would never be used. Now it
1635 is initialised to -1 but X509_time_adj() now has to check the value
1636 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1637 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1638 [Steve Henson, reported by Kenneth R. Robinette
1639 <support@securenetterm.com>]
1640
1641 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1642 result in a zero length in the ASN1_INTEGER structure which was
1643 not consistent with the structure when d2i_ASN1_INTEGER() was used
1644 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1645 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1646 where it did not print out a minus for negative ASN1_INTEGER.
1647 [Steve Henson]
1648
1649 *) Add summary printout to ocsp utility. The various functions which
1650 convert status values to strings have been renamed to:
1651 OCSP_response_status_str(), OCSP_cert_status_str() and
1652 OCSP_crl_reason_str() and are no longer static. New options
1653 to verify nonce values and to disable verification. OCSP response
1654 printout format cleaned up.
1655 [Steve Henson]
1656
1657 *) Add additional OCSP certificate checks. These are those specified
1658 in RFC2560. This consists of two separate checks: the CA of the
1659 certificate being checked must either be the OCSP signer certificate
1660 or the issuer of the OCSP signer certificate. In the latter case the
1661 OCSP signer certificate must contain the OCSP signing extended key
1662 usage. This check is performed by attempting to match the OCSP
1663 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1664 in the OCSP_CERTID structures of the response.
1665 [Steve Henson]
1666
1667 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1668 and related routines. This uses the standard OpenSSL certificate
1669 verify routines to perform initial checks (just CA validity) and
1670 to obtain the certificate chain. Then additional checks will be
1671 performed on the chain. Currently the root CA is checked to see
1672 if it is explicitly trusted for OCSP signing. This is used to set
1673 a root CA as a global signing root: that is any certificate that
1674 chains to that CA is an acceptable OCSP signing certificate.
1675 [Steve Henson]
1676
1677 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1678 extensions from a separate configuration file.
1679 As when reading extensions from the main configuration file,
1680 the '-extensions ...' option may be used for specifying the
1681 section to use.
1682 [Massimiliano Pala <madwolf@comune.modena.it>]
1683
1684 *) New OCSP utility. Allows OCSP requests to be generated or
1685 read. The request can be sent to a responder and the output
1686 parsed, outputed or printed in text form. Not complete yet:
1687 still needs to check the OCSP response validity.
1688 [Steve Henson]
1689
1690 *) New subcommands for 'openssl ca':
1691 'openssl ca -status <serial>' prints the status of the cert with
1692 the given serial number (according to the index file).
1693 'openssl ca -updatedb' updates the expiry status of certificates
1694 in the index file.
1695 [Massimiliano Pala <madwolf@comune.modena.it>]
1696
1697 *) New '-newreq-nodes' command option to CA.pl. This is like
1698 '-newreq', but calls 'openssl req' with the '-nodes' option
1699 so that the resulting key is not encrypted.
1700 [Damien Miller <djm@mindrot.org>]
1701
1702 *) New configuration for the GNU Hurd.
1703 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1704
1705 *) Initial code to implement OCSP basic response verify. This
1706 is currently incomplete. Currently just finds the signer's
1707 certificate and verifies the signature on the response.
1708 [Steve Henson]
1709
1710 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1711 value of OPENSSLDIR. This is available via the new '-d' option
1712 to 'openssl version', and is also included in 'openssl version -a'.
1713 [Bodo Moeller]
1714
1715 *) Allowing defining memory allocation callbacks that will be given
1716 file name and line number information in additional arguments
1717 (a const char* and an int). The basic functionality remains, as
1718 well as the original possibility to just replace malloc(),
1719 realloc() and free() by functions that do not know about these
1720 additional arguments. To register and find out the current
1721 settings for extended allocation functions, the following
1722 functions are provided:
1723
1724 CRYPTO_set_mem_ex_functions
1725 CRYPTO_set_locked_mem_ex_functions
1726 CRYPTO_get_mem_ex_functions
1727 CRYPTO_get_locked_mem_ex_functions
1728
1729 These work the same way as CRYPTO_set_mem_functions and friends.
1730 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1731 extended allocation function is enabled.
1732 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1733 a conventional allocation function is enabled.
1734 [Richard Levitte, Bodo Moeller]
1735
1736 *) Finish off removing the remaining LHASH function pointer casts.
1737 There should no longer be any prototype-casting required when using
1738 the LHASH abstraction, and any casts that remain are "bugs". See
1739 the callback types and macros at the head of lhash.h for details
1740 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1741 [Geoff Thorpe]
1742
1743 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1744 If /dev/[u]random devices are not available or do not return enough
1745 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1746 be queried.
1747 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1748 /etc/entropy will be queried once each in this sequence, quering stops
1749 when enough entropy was collected without querying more sockets.
1750 [Lutz Jaenicke]
1751
1752 *) Change the Unix RAND_poll() variant to be able to poll several
1753 random devices, as specified by DEVRANDOM, until a sufficient amount
1754 of data has been collected. We spend at most 10 ms on each file
1755 (select timeout) and read in non-blocking mode. DEVRANDOM now
1756 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1757 (previously it was just the string "/dev/urandom"), so on typical
1758 platforms the 10 ms delay will never occur.
1759 Also separate out the Unix variant to its own file, rand_unix.c.
1760 For VMS, there's a currently-empty rand_vms.c.
1761 [Richard Levitte]
1762
1763 *) Move OCSP client related routines to ocsp_cl.c. These
1764 provide utility functions which an application needing
1765 to issue a request to an OCSP responder and analyse the
1766 response will typically need: as opposed to those which an
1767 OCSP responder itself would need which will be added later.
1768
1769 OCSP_request_sign() signs an OCSP request with an API similar
1770 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1771 response. OCSP_response_get1_basic() extracts basic response
1772 from response. OCSP_resp_find_status(): finds and extracts status
1773 information from an OCSP_CERTID structure (which will be created
1774 when the request structure is built). These are built from lower
1775 level functions which work on OCSP_SINGLERESP structures but
1776 wont normally be used unless the application wishes to examine
1777 extensions in the OCSP response for example.
1778
1779 Replace nonce routines with a pair of functions.
1780 OCSP_request_add1_nonce() adds a nonce value and optionally
1781 generates a random value. OCSP_check_nonce() checks the
1782 validity of the nonce in an OCSP response.
1783 [Steve Henson]
1784
1785 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1786 This doesn't copy the supplied OCSP_CERTID and avoids the
1787 need to free up the newly created id. Change return type
1788 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1789 This can then be used to add extensions to the request.
1790 Deleted OCSP_request_new(), since most of its functionality
1791 is now in OCSP_REQUEST_new() (and the case insensitive name
1792 clash) apart from the ability to set the request name which
1793 will be added elsewhere.
1794 [Steve Henson]
1795
1796 *) Update OCSP API. Remove obsolete extensions argument from
1797 various functions. Extensions are now handled using the new
1798 OCSP extension code. New simple OCSP HTTP function which
1799 can be used to send requests and parse the response.
1800 [Steve Henson]
1801
1802 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1803 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1804 uses the special reorder version of SET OF to sort the attributes
1805 and reorder them to match the encoded order. This resolves a long
1806 standing problem: a verify on a PKCS7 structure just after signing
1807 it used to fail because the attribute order did not match the
1808 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1809 it uses the received order. This is necessary to tolerate some broken
1810 software that does not order SET OF. This is handled by encoding
1811 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1812 to produce the required SET OF.
1813 [Steve Henson]
1814
1815 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1816 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1817 files to get correct declarations of the ASN.1 item variables.
1818 [Richard Levitte]
1819
1820 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1821 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1822 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1823 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1824 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1825 ASN1_ITEM and no wrapper functions.
1826 [Steve Henson]
1827
1828 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1829 replace the old function pointer based I/O routines. Change most of
1830 the *_d2i_bio() and *_d2i_fp() functions to use these.
1831 [Steve Henson]
1832
1833 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1834 lines, recognice more "algorithms" that can be deselected, and make
1835 it complain about algorithm deselection that isn't recognised.
1836 [Richard Levitte]
1837
1838 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1839 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1840 to use new functions. Add NO_ASN1_OLD which can be set to remove
1841 some old style ASN1 functions: this can be used to determine if old
1842 code will still work when these eventually go away.
1843 [Steve Henson]
1844
1845 *) New extension functions for OCSP structures, these follow the
1846 same conventions as certificates and CRLs.
1847 [Steve Henson]
1848
1849 *) New function X509V3_add1_i2d(). This automatically encodes and
1850 adds an extension. Its behaviour can be customised with various
1851 flags to append, replace or delete. Various wrappers added for
1852 certifcates and CRLs.
1853 [Steve Henson]
1854
1855 *) Fix to avoid calling the underlying ASN1 print routine when
1856 an extension cannot be parsed. Correct a typo in the
1857 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1858 [Steve Henson]
1859
1860 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1861 entries for variables.
1862 [Steve Henson]
1863
1864 *) Add functionality to apps/openssl.c for detecting locking
1865 problems: As the program is single-threaded, all we have
1866 to do is register a locking callback using an array for
1867 storing which locks are currently held by the program.
1868 [Bodo Moeller]
1869
1870 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1871 SSL_get_ex_data_X509_STORE_idx(), which is used in
1872 ssl_verify_cert_chain() and thus can be called at any time
1873 during TLS/SSL handshakes so that thread-safety is essential.
1874 Unfortunately, the ex_data design is not at all suited
1875 for multi-threaded use, so it probably should be abolished.
1876 [Bodo Moeller]
1877
1878 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1879 [Broadcom, tweaked and integrated by Geoff Thorpe]
1880
1881 *) Move common extension printing code to new function
1882 X509V3_print_extensions(). Reorganise OCSP print routines and
1883 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1884 [Steve Henson]
1885
1886 *) New function X509_signature_print() to remove duplication in some
1887 print routines.
1888 [Steve Henson]
1889
1890 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1891 set (this was treated exactly the same as SET OF previously). This
1892 is used to reorder the STACK representing the structure to match the
1893 encoding. This will be used to get round a problem where a PKCS7
1894 structure which was signed could not be verified because the STACK
1895 order did not reflect the encoded order.
1896 [Steve Henson]
1897
1898 *) Reimplement the OCSP ASN1 module using the new code.
1899 [Steve Henson]
1900
1901 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1902 for its ASN1 operations. The old style function pointers still exist
1903 for now but they will eventually go away.
1904 [Steve Henson]
1905
1906 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1907 completely replaces the old ASN1 functionality with a table driven
1908 encoder and decoder which interprets an ASN1_ITEM structure describing
1909 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1910 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1911 has also been converted to the new form.
1912 [Steve Henson]
1913
1914 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1915 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1916 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1917 for negative moduli.
1918 [Bodo Moeller]
1919
1920 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1921 of not touching the result's sign bit.
1922 [Bodo Moeller]
1923
1924 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1925 set.
1926 [Bodo Moeller]
1927
1928 *) Changed the LHASH code to use prototypes for callbacks, and created
1929 macros to declare and implement thin (optionally static) functions
1930 that provide type-safety and avoid function pointer casting for the
1931 type-specific callbacks.
1932 [Geoff Thorpe]
1933
1934 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1935 RFC 2712.
1936 [Veers Staats <staatsvr@asc.hpc.mil>,
1937 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1938
1939 *) Reformat the FAQ so the different questions and answers can be divided
1940 in sections depending on the subject.
1941 [Richard Levitte]
1942
1943 *) Have the zlib compression code load ZLIB.DLL dynamically under
1944 Windows.
1945 [Richard Levitte]
1946
1947 *) New function BN_mod_sqrt for computing square roots modulo a prime
1948 (using the probabilistic Tonelli-Shanks algorithm unless
1949 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1950 be handled deterministically).
1951 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1952
1953 *) Make BN_mod_inverse faster by explicitly handling small quotients
1954 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1955 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1956 [Bodo Moeller]
1957
1958 *) New function BN_kronecker.
1959 [Bodo Moeller]
1960
1961 *) Fix BN_gcd so that it works on negative inputs; the result is
1962 positive unless both parameters are zero.
1963 Previously something reasonably close to an infinite loop was
1964 possible because numbers could be growing instead of shrinking
1965 in the implementation of Euclid's algorithm.
1966 [Bodo Moeller]
1967
1968 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1969 sign of the number in question.
1970
1971 Fix BN_is_word(a,w) to work correctly for w == 0.
1972
1973 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1974 because its test if the absolute value of 'a' equals 'w'.
1975 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1976 it exists mostly for use in the implementations of BN_is_zero(),
1977 BN_is_one(), and BN_is_word().
1978 [Bodo Moeller]
1979
1980 *) New function BN_swap.
1981 [Bodo Moeller]
1982
1983 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1984 the exponentiation functions are more likely to produce reasonable
1985 results on negative inputs.
1986 [Bodo Moeller]
1987
1988 *) Change BN_mod_mul so that the result is always non-negative.
1989 Previously, it could be negative if one of the factors was negative;
1990 I don't think anyone really wanted that behaviour.
1991 [Bodo Moeller]
1992
1993 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1994 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1995 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1996 and add new functions:
1997
1998 BN_nnmod
1999 BN_mod_sqr
2000 BN_mod_add
2001 BN_mod_add_quick
2002 BN_mod_sub
2003 BN_mod_sub_quick
2004 BN_mod_lshift1
2005 BN_mod_lshift1_quick
2006 BN_mod_lshift
2007 BN_mod_lshift_quick
2008
2009 These functions always generate non-negative results.
2010
2011 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2012 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2013
2014 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2015 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2016 be reduced modulo m.
2017 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2018
2019 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2020 was actually never needed) and in BN_mul(). The removal in BN_mul()
2021 required a small change in bn_mul_part_recursive() and the addition
2022 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2023 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2024 bn_sub_words() and bn_add_words() except they take arrays with
2025 differing sizes.
2026 [Richard Levitte]
2027
2028 *) In 'openssl passwd', verify passwords read from the terminal
2029 unless the '-salt' option is used (which usually means that
2030 verification would just waste user's time since the resulting
2031 hash is going to be compared with some given password hash)
2032 or the new '-noverify' option is used.
2033
2034 This is an incompatible change, but it does not affect
2035 non-interactive use of 'openssl passwd' (passwords on the command
2036 line, '-stdin' option, '-in ...' option) and thus should not
2037 cause any problems.
2038 [Bodo Moeller]
2039
2040 *) Remove all references to RSAref, since there's no more need for it.
2041 [Richard Levitte]
2042
2043 *) Make DSO load along a path given through an environment variable
2044 (SHLIB_PATH) with shl_load().
2045 [Richard Levitte]
2046
2047 *) Constify the ENGINE code as a result of BIGNUM constification.
2048 Also constify the RSA code and most things related to it. In a
2049 few places, most notable in the depth of the ASN.1 code, ugly
2050 casts back to non-const were required (to be solved at a later
2051 time)
2052 [Richard Levitte]
2053
2054 *) Make it so the openssl application has all engines loaded by default.
2055 [Richard Levitte]
2056
2057 *) Constify the BIGNUM routines a little more.
2058 [Richard Levitte]
2059
2060 *) Add the following functions:
2061
2062 ENGINE_load_cswift()
2063 ENGINE_load_chil()
2064 ENGINE_load_atalla()
2065 ENGINE_load_nuron()
2066 ENGINE_load_builtin_engines()
2067
2068 That way, an application can itself choose if external engines that
2069 are built-in in OpenSSL shall ever be used or not. The benefit is
2070 that applications won't have to be linked with libdl or other dso
2071 libraries unless it's really needed.
2072
2073 Changed 'openssl engine' to load all engines on demand.
2074 Changed the engine header files to avoid the duplication of some
2075 declarations (they differed!).
2076 [Richard Levitte]
2077
2078 *) 'openssl engine' can now list capabilities.
2079 [Richard Levitte]
2080
2081 *) Better error reporting in 'openssl engine'.
2082 [Richard Levitte]
2083
2084 *) Never call load_dh_param(NULL) in s_server.
2085 [Bodo Moeller]
2086
2087 *) Add engine application. It can currently list engines by name and
2088 identity, and test if they are actually available.
2089 [Richard Levitte]
2090
2091 *) Improve RPM specification file by forcing symbolic linking and making
2092 sure the installed documentation is also owned by root.root.
2093 [Damien Miller <djm@mindrot.org>]
2094
2095 *) Give the OpenSSL applications more possibilities to make use of
2096 keys (public as well as private) handled by engines.
2097 [Richard Levitte]
2098
2099 *) Add OCSP code that comes from CertCo.
2100 [Richard Levitte]
2101
2102 *) Add VMS support for the Rijndael code.
2103 [Richard Levitte]
2104
2105 *) Added untested support for Nuron crypto accelerator.
2106 [Ben Laurie]
2107
2108 *) Add support for external cryptographic devices. This code was
2109 previously distributed separately as the "engine" branch.
2110 [Geoff Thorpe, Richard Levitte]
2111
2112 *) Rework the filename-translation in the DSO code. It is now possible to
2113 have far greater control over how a "name" is turned into a filename
2114 depending on the operating environment and any oddities about the
2115 different shared library filenames on each system.
2116 [Geoff Thorpe]
2117
2118 *) Support threads on FreeBSD-elf in Configure.
2119 [Richard Levitte]
2120
2121 *) Fix for SHA1 assembly problem with MASM: it produces
2122 warnings about corrupt line number information when assembling
2123 with debugging information. This is caused by the overlapping
2124 of two sections.
2125 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2126
2127 *) NCONF changes.
2128 NCONF_get_number() has no error checking at all. As a replacement,
2129 NCONF_get_number_e() is defined (_e for "error checking") and is
2130 promoted strongly. The old NCONF_get_number is kept around for
2131 binary backward compatibility.
2132 Make it possible for methods to load from something other than a BIO,
2133 by providing a function pointer that is given a name instead of a BIO.
2134 For example, this could be used to load configuration data from an
2135 LDAP server.
2136 [Richard Levitte]
2137
2138 *) Fix for non blocking accept BIOs. Added new I/O special reason
2139 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2140 with non blocking I/O was not possible because no retry code was
2141 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2142 this case.
2143 [Steve Henson]
2144
2145 *) Added the beginnings of Rijndael support.
2146 [Ben Laurie]
2147
2148 *) Fix for bug in DirectoryString mask setting. Add support for
2149 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2150 to allow certificate printing to more controllable, additional
2151 'certopt' option to 'x509' to allow new printing options to be
2152 set.
2153 [Steve Henson]
2154
2155 *) Clean old EAY MD5 hack from e_os.h.
2156 [Richard Levitte]
2157
2158 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2159
2160 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2161 memory from it's contents. This is done with a counter that will
2162 place alternating values in each byte. This can be used to solve
2163 two issues: 1) the removal of calls to memset() by highly optimizing
2164 compilers, and 2) cleansing with other values than 0, since those can
2165 be read through on certain media, for example a swap space on disk.
2166 [Geoff Thorpe]
2167
2168 *) Bugfix: client side session caching did not work with external caching,
2169 because the session->cipher setting was not restored when reloading
2170 from the external cache. This problem was masked, when
2171 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2172 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2173 [Lutz Jaenicke]
2174
2175 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2176 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2177 [Zeev Lieber <zeev-l@yahoo.com>]
2178
2179 *) Undo an undocumented change introduced in 0.9.6e which caused
2180 repeated calls to OpenSSL_add_all_ciphers() and
2181 OpenSSL_add_all_digests() to be ignored, even after calling
2182 EVP_cleanup().
2183 [Richard Levitte]
2184
2185 *) Change the default configuration reader to deal with last line not
2186 being properly terminated.
2187 [Richard Levitte]
2188
2189 *) Change X509_NAME_cmp() so it applies the special rules on handling
2190 DN values that are of type PrintableString, as well as RDNs of type
2191 emailAddress where the value has the type ia5String.
2192 [stefank@valicert.com via Richard Levitte]
2193
2194 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2195 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2196 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2197 the bitwise-OR of the two for use by the majority of applications
2198 wanting this behaviour, and update the docs. The documented
2199 behaviour and actual behaviour were inconsistent and had been
2200 changing anyway, so this is more a bug-fix than a behavioural
2201 change.
2202 [Geoff Thorpe, diagnosed by Nadav Har'El]
2203
2204 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2205 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2206 [Bodo Moeller]
2207
2208 *) Fix initialization code race conditions in
2209 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2210 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2211 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2212 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2213 ssl2_get_cipher_by_char(),
2214 ssl3_get_cipher_by_char().
2215 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2216
2217 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2218 the cached sessions are flushed, as the remove_cb() might use ex_data
2219 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2220 (see [openssl.org #212]).
2221 [Geoff Thorpe, Lutz Jaenicke]
2222
2223 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2224 length, instead of the encoding length to d2i_ASN1_OBJECT.
2225 [Steve Henson]
2226
2227 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2228
2229 *) [In 0.9.6g-engine release:]
2230 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2231 [Lynn Gazis <lgazis@rainbow.com>]
2232
2233 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2234
2235 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2236 and get fix the header length calculation.
2237 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2238 Alon Kantor <alonk@checkpoint.com> (and others),
2239 Steve Henson]
2240
2241 *) Use proper error handling instead of 'assertions' in buffer
2242 overflow checks added in 0.9.6e. This prevents DoS (the
2243 assertions could call abort()).
2244 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2245
2246 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2247
2248 *) Add various sanity checks to asn1_get_length() to reject
2249 the ASN1 length bytes if they exceed sizeof(long), will appear
2250 negative or the content length exceeds the length of the
2251 supplied buffer.
2252 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2253
2254 *) Fix cipher selection routines: ciphers without encryption had no flags
2255 for the cipher strength set and where therefore not handled correctly
2256 by the selection routines (PR #130).
2257 [Lutz Jaenicke]
2258
2259 *) Fix EVP_dsa_sha macro.
2260 [Nils Larsch]
2261
2262 *) New option
2263 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2264 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2265 that was added in OpenSSL 0.9.6d.
2266
2267 As the countermeasure turned out to be incompatible with some
2268 broken SSL implementations, the new option is part of SSL_OP_ALL.
2269 SSL_OP_ALL is usually employed when compatibility with weird SSL
2270 implementations is desired (e.g. '-bugs' option to 's_client' and
2271 's_server'), so the new option is automatically set in many
2272 applications.
2273 [Bodo Moeller]
2274
2275 *) Changes in security patch:
2276
2277 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2278 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2279 Air Force Materiel Command, USAF, under agreement number
2280 F30602-01-2-0537.
2281
2282 *) Add various sanity checks to asn1_get_length() to reject
2283 the ASN1 length bytes if they exceed sizeof(long), will appear
2284 negative or the content length exceeds the length of the
2285 supplied buffer. (CAN-2002-0659)
2286 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2287
2288 *) Assertions for various potential buffer overflows, not known to
2289 happen in practice.
2290 [Ben Laurie (CHATS)]
2291
2292 *) Various temporary buffers to hold ASCII versions of integers were
2293 too small for 64 bit platforms. (CAN-2002-0655)
2294 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2295
2296 *) Remote buffer overflow in SSL3 protocol - an attacker could
2297 supply an oversized session ID to a client. (CAN-2002-0656)
2298 [Ben Laurie (CHATS)]
2299
2300 *) Remote buffer overflow in SSL2 protocol - an attacker could
2301 supply an oversized client master key. (CAN-2002-0656)
2302 [Ben Laurie (CHATS)]
2303
2304 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2305
2306 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2307 encoded as NULL) with id-dsa-with-sha1.
2308 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2309
2310 *) Check various X509_...() return values in apps/req.c.
2311 [Nils Larsch <nla@trustcenter.de>]
2312
2313 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2314 an end-of-file condition would erronously be flagged, when the CRLF
2315 was just at the end of a processed block. The bug was discovered when
2316 processing data through a buffering memory BIO handing the data to a
2317 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2318 <ptsekov@syntrex.com> and Nedelcho Stanev.
2319 [Lutz Jaenicke]
2320
2321 *) Implement a countermeasure against a vulnerability recently found
2322 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2323 before application data chunks to avoid the use of known IVs
2324 with data potentially chosen by the attacker.
2325 [Bodo Moeller]
2326
2327 *) Fix length checks in ssl3_get_client_hello().
2328 [Bodo Moeller]
2329
2330 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2331 to prevent ssl3_read_internal() from incorrectly assuming that
2332 ssl3_read_bytes() found application data while handshake
2333 processing was enabled when in fact s->s3->in_read_app_data was
2334 merely automatically cleared during the initial handshake.
2335 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2336
2337 *) Fix object definitions for Private and Enterprise: they were not
2338 recognized in their shortname (=lowercase) representation. Extend
2339 obj_dat.pl to issue an error when using undefined keywords instead
2340 of silently ignoring the problem (Svenning Sorensen
2341 <sss@sss.dnsalias.net>).
2342 [Lutz Jaenicke]
2343
2344 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2345 generators, i.e. generators other than 2 and 5. (Previously, the
2346 code did not properly initialise the 'add' and 'rem' values to
2347 BN_generate_prime().)
2348
2349 In the new general case, we do not insist that 'generator' is
2350 actually a primitive root: This requirement is rather pointless;
2351 a generator of the order-q subgroup is just as good, if not
2352 better.
2353 [Bodo Moeller]
2354
2355 *) Map new X509 verification errors to alerts. Discovered and submitted by
2356 Tom Wu <tom@arcot.com>.
2357 [Lutz Jaenicke]
2358
2359 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2360 returning non-zero before the data has been completely received
2361 when using non-blocking I/O.
2362 [Bodo Moeller; problem pointed out by John Hughes]
2363
2364 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2365 [Ben Laurie, Lutz Jaenicke]
2366
2367 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2368 Yoram Zahavi <YoramZ@gilian.com>).
2369 [Lutz Jaenicke]
2370
2371 *) Add information about CygWin 1.3 and on, and preserve proper
2372 configuration for the versions before that.
2373 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2374
2375 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2376 check whether we deal with a copy of a session and do not delete from
2377 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2378 <izhar@checkpoint.com>.
2379 [Lutz Jaenicke]
2380
2381 *) Do not store session data into the internal session cache, if it
2382 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2383 flag is set). Proposed by Aslam <aslam@funk.com>.
2384 [Lutz Jaenicke]
2385
2386 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2387 value is 0.
2388 [Richard Levitte]
2389
2390 *) [In 0.9.6d-engine release:]
2391 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2392 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2393
2394 *) Add the configuration target linux-s390x.
2395 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2396
2397 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2398 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2399 variable as an indication that a ClientHello message has been
2400 received. As the flag value will be lost between multiple
2401 invocations of ssl3_accept when using non-blocking I/O, the
2402 function may not be aware that a handshake has actually taken
2403 place, thus preventing a new session from being added to the
2404 session cache.
2405
2406 To avoid this problem, we now set s->new_session to 2 instead of
2407 using a local variable.
2408 [Lutz Jaenicke, Bodo Moeller]
2409
2410 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2411 if the SSL_R_LENGTH_MISMATCH error is detected.
2412 [Geoff Thorpe, Bodo Moeller]
2413
2414 *) New 'shared_ldflag' column in Configure platform table.
2415 [Richard Levitte]
2416
2417 *) Fix EVP_CIPHER_mode macro.
2418 ["Dan S. Camper" <dan@bti.net>]
2419
2420 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2421 type, we must throw them away by setting rr->length to 0.
2422 [D P Chang <dpc@qualys.com>]
2423
2424 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2425
2426 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2427 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2428 worked incorrectly for those cases where range = 10..._2 and
2429 3*range is two bits longer than range.)
2430 [Bodo Moeller]
2431
2432 *) Only add signing time to PKCS7 structures if it is not already
2433 present.
2434 [Steve Henson]
2435
2436 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2437 OBJ_ld_ce should be OBJ_id_ce.
2438 Also some ip-pda OIDs in crypto/objects/objects.txt were
2439 incorrect (cf. RFC 3039).
2440 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2441
2442 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2443 returns early because it has nothing to do.
2444 [Andy Schneider <andy.schneider@bjss.co.uk>]
2445
2446 *) [In 0.9.6c-engine release:]
2447 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2448 [Andy Schneider <andy.schneider@bjss.co.uk>]
2449
2450 *) [In 0.9.6c-engine release:]
2451 Add support for Cryptographic Appliance's keyserver technology.
2452 (Use engine 'keyclient')
2453 [Cryptographic Appliances and Geoff Thorpe]
2454
2455 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2456 is called via tools/c89.sh because arguments have to be
2457 rearranged (all '-L' options must appear before the first object
2458 modules).
2459 [Richard Shapiro <rshapiro@abinitio.com>]
2460
2461 *) [In 0.9.6c-engine release:]
2462 Add support for Broadcom crypto accelerator cards, backported
2463 from 0.9.7.
2464 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2465
2466 *) [In 0.9.6c-engine release:]
2467 Add support for SureWare crypto accelerator cards from
2468 Baltimore Technologies. (Use engine 'sureware')
2469 [Baltimore Technologies and Mark Cox]
2470
2471 *) [In 0.9.6c-engine release:]
2472 Add support for crypto accelerator cards from Accelerated
2473 Encryption Processing, www.aep.ie. (Use engine 'aep')
2474 [AEP Inc. and Mark Cox]
2475
2476 *) Add a configuration entry for gcc on UnixWare.
2477 [Gary Benson <gbenson@redhat.com>]
2478
2479 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2480 messages are stored in a single piece (fixed-length part and
2481 variable-length part combined) and fix various bugs found on the way.
2482 [Bodo Moeller]
2483
2484 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2485 instead. BIO_gethostbyname() does not know what timeouts are
2486 appropriate, so entries would stay in cache even when they have
2487 become invalid.
2488 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2489
2490 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2491 faced with a pathologically small ClientHello fragment that does
2492 not contain client_version: Instead of aborting with an error,
2493 simply choose the highest available protocol version (i.e.,
2494 TLS 1.0 unless it is disabled). In practice, ClientHello
2495 messages are never sent like this, but this change gives us
2496 strictly correct behaviour at least for TLS.
2497 [Bodo Moeller]
2498
2499 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2500 never resets s->method to s->ctx->method when called from within
2501 one of the SSL handshake functions.
2502 [Bodo Moeller; problem pointed out by Niko Baric]
2503
2504 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2505 (sent using the client's version number) if client_version is
2506 smaller than the protocol version in use. Also change
2507 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2508 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2509 the client will at least see that alert.
2510 [Bodo Moeller]
2511
2512 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2513 correctly.
2514 [Bodo Moeller]
2515
2516 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2517 client receives HelloRequest while in a handshake.
2518 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2519
2520 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2521 should end in 'break', not 'goto end' which circuments various
2522 cleanups done in state SSL_ST_OK. But session related stuff
2523 must be disabled for SSL_ST_OK in the case that we just sent a
2524 HelloRequest.
2525
2526 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2527 before just sending a HelloRequest.
2528 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2529
2530 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2531 reveal whether illegal block cipher padding was found or a MAC
2532 verification error occured. (Neither SSLerr() codes nor alerts
2533 are directly visible to potential attackers, but the information
2534 may leak via logfiles.)
2535
2536 Similar changes are not required for the SSL 2.0 implementation
2537 because the number of padding bytes is sent in clear for SSL 2.0,
2538 and the extra bytes are just ignored. However ssl/s2_pkt.c
2539 failed to verify that the purported number of padding bytes is in
2540 the legal range.
2541 [Bodo Moeller]
2542
2543 *) Add OpenUNIX-8 support including shared libraries
2544 (Boyd Lynn Gerber <gerberb@zenez.com>).
2545 [Lutz Jaenicke]
2546
2547 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2548 'wristwatch attack' using huge encoding parameters (cf.
2549 James H. Manger's CRYPTO 2001 paper). Note that the
2550 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2551 encoding parameters and hence was not vulnerable.
2552 [Bodo Moeller]
2553
2554 *) BN_sqr() bug fix.
2555 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2556
2557 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2558 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2559 followed by modular reduction.
2560 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2561
2562 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2563 equivalent based on BN_pseudo_rand() instead of BN_rand().
2564 [Bodo Moeller]
2565
2566 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2567 This function was broken, as the check for a new client hello message
2568 to handle SGC did not allow these large messages.
2569 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2570 [Lutz Jaenicke]
2571
2572 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2573 [Lutz Jaenicke]
2574
2575 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2576 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2577 [Lutz Jaenicke]
2578
2579 *) Rework the configuration and shared library support for Tru64 Unix.
2580 The configuration part makes use of modern compiler features and
2581 still retains old compiler behavior for those that run older versions
2582 of the OS. The shared library support part includes a variant that
2583 uses the RPATH feature, and is available through the special
2584 configuration target "alpha-cc-rpath", which will never be selected
2585 automatically.
2586 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2587
2588 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2589 with the same message size as in ssl3_get_certificate_request().
2590 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2591 messages might inadvertently be reject as too long.
2592 [Petr Lampa <lampa@fee.vutbr.cz>]
2593
2594 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2595 [Andy Polyakov]
2596
2597 *) Modified SSL library such that the verify_callback that has been set
2598 specificly for an SSL object with SSL_set_verify() is actually being
2599 used. Before the change, a verify_callback set with this function was
2600 ignored and the verify_callback() set in the SSL_CTX at the time of
2601 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2602 to allow the necessary settings.
2603 [Lutz Jaenicke]
2604
2605 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2606 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2607 done automatically (in contradiction to the requirements of the C
2608 standard). This made problems when used from OpenSSH.
2609 [Lutz Jaenicke]
2610
2611 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2612 dh->length and always used
2613
2614 BN_rand_range(priv_key, dh->p).
2615
2616 BN_rand_range() is not necessary for Diffie-Hellman, and this
2617 specific range makes Diffie-Hellman unnecessarily inefficient if
2618 dh->length (recommended exponent length) is much smaller than the
2619 length of dh->p. We could use BN_rand_range() if the order of
2620 the subgroup was stored in the DH structure, but we only have
2621 dh->length.
2622
2623 So switch back to
2624
2625 BN_rand(priv_key, l, ...)
2626
2627 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2628 otherwise.
2629 [Bodo Moeller]
2630
2631 *) In
2632
2633 RSA_eay_public_encrypt
2634 RSA_eay_private_decrypt
2635 RSA_eay_private_encrypt (signing)
2636 RSA_eay_public_decrypt (signature verification)
2637
2638 (default implementations for RSA_public_encrypt,
2639 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2640 always reject numbers >= n.
2641 [Bodo Moeller]
2642
2643 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2644 to synchronize access to 'locking_thread'. This is necessary on
2645 systems where access to 'locking_thread' (an 'unsigned long'
2646 variable) is not atomic.
2647 [Bodo Moeller]
2648
2649 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2650 *before* setting the 'crypto_lock_rand' flag. The previous code had
2651 a race condition if 0 is a valid thread ID.
2652 [Travis Vitek <vitek@roguewave.com>]
2653
2654 *) Add support for shared libraries under Irix.
2655 [Albert Chin-A-Young <china@thewrittenword.com>]
2656
2657 *) Add configuration option to build on Linux on both big-endian and
2658 little-endian MIPS.
2659 [Ralf Baechle <ralf@uni-koblenz.de>]
2660
2661 *) Add the possibility to create shared libraries on HP-UX.
2662 [Richard Levitte]
2663
2664 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2665
2666 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2667 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2668 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2669 PRNG state recovery was possible based on the output of
2670 one PRNG request appropriately sized to gain knowledge on
2671 'md' followed by enough consecutive 1-byte PRNG requests
2672 to traverse all of 'state'.
2673
2674 1. When updating 'md_local' (the current thread's copy of 'md')
2675 during PRNG output generation, hash all of the previous
2676 'md_local' value, not just the half used for PRNG output.
2677
2678 2. Make the number of bytes from 'state' included into the hash
2679 independent from the number of PRNG bytes requested.
2680
2681 The first measure alone would be sufficient to avoid
2682 Markku-Juhani's attack. (Actually it had never occurred
2683 to me that the half of 'md_local' used for chaining was the
2684 half from which PRNG output bytes were taken -- I had always
2685 assumed that the secret half would be used.) The second
2686 measure makes sure that additional data from 'state' is never
2687 mixed into 'md_local' in small portions; this heuristically
2688 further strengthens the PRNG.
2689 [Bodo Moeller]
2690
2691 *) Fix crypto/bn/asm/mips3.s.
2692 [Andy Polyakov]
2693
2694 *) When only the key is given to "enc", the IV is undefined. Print out
2695 an error message in this case.
2696 [Lutz Jaenicke]
2697
2698 *) Handle special case when X509_NAME is empty in X509 printing routines.
2699 [Steve Henson]
2700
2701 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2702 positive and less than q.
2703 [Bodo Moeller]
2704
2705 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2706 used: it isn't thread safe and the add_lock_callback should handle
2707 that itself.
2708 [Paul Rose <Paul.Rose@bridge.com>]
2709
2710 *) Verify that incoming data obeys the block size in
2711 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2712 [Bodo Moeller]
2713
2714 *) Fix OAEP check.
2715 [Ulf Möller, Bodo Möller]
2716
2717 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2718 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2719 when fixing the server behaviour for backwards-compatible 'client
2720 hello' messages. (Note that the attack is impractical against
2721 SSL 3.0 and TLS 1.0 anyway because length and version checking
2722 means that the probability of guessing a valid ciphertext is
2723 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2724 paper.)
2725
2726 Before 0.9.5, the countermeasure (hide the error by generating a
2727 random 'decryption result') did not work properly because
2728 ERR_clear_error() was missing, meaning that SSL_get_error() would
2729 detect the supposedly ignored error.
2730
2731 Both problems are now fixed.
2732 [Bodo Moeller]
2733
2734 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2735 (previously it was 1024).
2736 [Bodo Moeller]
2737
2738 *) Fix for compatibility mode trust settings: ignore trust settings
2739 unless some valid trust or reject settings are present.
2740 [Steve Henson]
2741
2742 *) Fix for blowfish EVP: its a variable length cipher.
2743 [Steve Henson]
2744
2745 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2746 parameters in DSA public key structures and return an error in the
2747 DSA routines if parameters are absent.
2748 [Steve Henson]
2749
2750 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2751 in the current directory if neither $RANDFILE nor $HOME was set.
2752 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2753 caused some confusion to Windows users who haven't defined $HOME.
2754 Thus RAND_file_name() is changed again: e_os.h can define a
2755 DEFAULT_HOME, which will be used if $HOME is not set.
2756 For Windows, we use "C:"; on other platforms, we still require
2757 environment variables.
2758
2759 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2760 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2761 having multiple threads call RAND_poll() concurrently.
2762 [Bodo Moeller]
2763
2764 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2765 combination of a flag and a thread ID variable.
2766 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2767 flag), *other* threads can enter ssleay_add_bytes without obeying
2768 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2769 that they do not hold after the first thread unsets add_do_not_lock).
2770 [Bodo Moeller]
2771
2772 *) Change bctest again: '-x' expressions are not available in all
2773 versions of 'test'.
2774 [Bodo Moeller]
2775
2776 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2777
2778 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2779 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2780
2781 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2782 the default extension for executables, if any. Also, make the perl
2783 scripts that use symlink() to test if it really exists and use "cp"
2784 if it doesn't. All this made OpenSSL compilable and installable in
2785 CygWin.
2786 [Richard Levitte]
2787
2788 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2789 If SEQUENCE is length is indefinite just set c->slen to the total
2790 amount of data available.
2791 [Steve Henson, reported by shige@FreeBSD.org]
2792 [This change does not apply to 0.9.7.]
2793
2794 *) Change bctest to avoid here-documents inside command substitution
2795 (workaround for FreeBSD /bin/sh bug).
2796 For compatibility with Ultrix, avoid shell functions (introduced
2797 in the bctest version that searches along $PATH).
2798 [Bodo Moeller]
2799
2800 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2801 with des_encrypt() defined on some operating systems, like Solaris
2802 and UnixWare.
2803 [Richard Levitte]
2804
2805 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2806 On the Importance of Eliminating Errors in Cryptographic
2807 Computations, J. Cryptology 14 (2001) 2, 101-119,
2808 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2809 [Ulf Moeller]
2810
2811 *) MIPS assembler BIGNUM division bug fix.
2812 [Andy Polyakov]
2813
2814 *) Disabled incorrect Alpha assembler code.
2815 [Richard Levitte]
2816
2817 *) Fix PKCS#7 decode routines so they correctly update the length
2818 after reading an EOC for the EXPLICIT tag.
2819 [Steve Henson]
2820 [This change does not apply to 0.9.7.]
2821
2822 *) Fix bug in PKCS#12 key generation routines. This was triggered
2823 if a 3DES key was generated with a 0 initial byte. Include
2824 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2825 (but broken) behaviour.
2826 [Steve Henson]
2827
2828 *) Enhance bctest to search for a working bc along $PATH and print
2829 it when found.
2830 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2831
2832 *) Fix memory leaks in err.c: free err_data string if necessary;
2833 don't write to the wrong index in ERR_set_error_data.
2834 [Bodo Moeller]
2835
2836 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2837 did not exist.
2838 [Bodo Moeller]
2839
2840 *) Replace rdtsc with _emit statements for VC++ version 5.
2841 [Jeremy Cooper <jeremy@baymoo.org>]
2842
2843 *) Make it possible to reuse SSLv2 sessions.
2844 [Richard Levitte]
2845
2846 *) In copy_email() check for >= 0 as a return value for
2847 X509_NAME_get_index_by_NID() since 0 is a valid index.
2848 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2849
2850 *) Avoid coredump with unsupported or invalid public keys by checking if
2851 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2852 PKCS7_verify() fails with non detached data.
2853 [Steve Henson]
2854
2855 *) Don't use getenv in library functions when run as setuid/setgid.
2856 New function OPENSSL_issetugid().
2857 [Ulf Moeller]
2858
2859 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2860 due to incorrect handling of multi-threading:
2861
2862 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2863
2864 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2865
2866 3. Count how many times MemCheck_off() has been called so that
2867 nested use can be treated correctly. This also avoids
2868 inband-signalling in the previous code (which relied on the
2869 assumption that thread ID 0 is impossible).
2870 [Bodo Moeller]
2871
2872 *) Add "-rand" option also to s_client and s_server.
2873 [Lutz Jaenicke]
2874
2875 *) Fix CPU detection on Irix 6.x.
2876 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2877 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2878
2879 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2880 was empty.
2881 [Steve Henson]
2882 [This change does not apply to 0.9.7.]
2883
2884 *) Use the cached encoding of an X509_NAME structure rather than
2885 copying it. This is apparently the reason for the libsafe "errors"
2886 but the code is actually correct.
2887 [Steve Henson]
2888
2889 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2890 Bleichenbacher's DSA attack.
2891 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2892 to be set and top=0 forces the highest bit to be set; top=-1 is new
2893 and leaves the highest bit random.
2894 [Ulf Moeller, Bodo Moeller]
2895
2896 *) In the NCONF_...-based implementations for CONF_... queries
2897 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2898 a temporary CONF structure with the data component set to NULL
2899 (which gives segmentation faults in lh_retrieve).
2900 Instead, use NULL for the CONF pointer in CONF_get_string and
2901 CONF_get_number (which may use environment variables) and directly
2902 return NULL from CONF_get_section.
2903 [Bodo Moeller]
2904
2905 *) Fix potential buffer overrun for EBCDIC.
2906 [Ulf Moeller]
2907
2908 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2909 keyUsage if basicConstraints absent for a CA.
2910 [Steve Henson]
2911
2912 *) Make SMIME_write_PKCS7() write mail header values with a format that
2913 is more generally accepted (no spaces before the semicolon), since
2914 some programs can't parse those values properly otherwise. Also make
2915 sure BIO's that break lines after each write do not create invalid
2916 headers.
2917 [Richard Levitte]
2918
2919 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2920 macros previously used would not encode an empty SEQUENCE OF
2921 and break the signature.
2922 [Steve Henson]
2923 [This change does not apply to 0.9.7.]
2924
2925 *) Zero the premaster secret after deriving the master secret in
2926 DH ciphersuites.
2927 [Steve Henson]
2928
2929 *) Add some EVP_add_digest_alias registrations (as found in
2930 OpenSSL_add_all_digests()) to SSL_library_init()
2931 aka OpenSSL_add_ssl_algorithms(). This provides improved
2932 compatibility with peers using X.509 certificates
2933 with unconventional AlgorithmIdentifier OIDs.
2934 [Bodo Moeller]
2935
2936 *) Fix for Irix with NO_ASM.
2937 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2938
2939 *) ./config script fixes.
2940 [Ulf Moeller, Richard Levitte]
2941
2942 *) Fix 'openssl passwd -1'.
2943 [Bodo Moeller]
2944
2945 *) Change PKCS12_key_gen_asc() so it can cope with non null
2946 terminated strings whose length is passed in the passlen
2947 parameter, for example from PEM callbacks. This was done
2948 by adding an extra length parameter to asc2uni().
2949 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2950
2951 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2952 call failed, free the DSA structure.
2953 [Bodo Moeller]
2954
2955 *) Fix to uni2asc() to cope with zero length Unicode strings.
2956 These are present in some PKCS#12 files.
2957 [Steve Henson]
2958
2959 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2960 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2961 when writing a 32767 byte record.
2962 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2963
2964 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2965 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2966
2967 (RSA objects have a reference count access to which is protected
2968 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2969 so they are meant to be shared between threads.)
2970 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2971 "Reddie, Steven" <Steven.Reddie@ca.com>]
2972
2973 *) Fix a deadlock in CRYPTO_mem_leaks().
2974 [Bodo Moeller]
2975
2976 *) Use better test patterns in bntest.
2977 [Ulf Möller]
2978
2979 *) rand_win.c fix for Borland C.
2980 [Ulf Möller]
2981
2982 *) BN_rshift bugfix for n == 0.
2983 [Bodo Moeller]
2984
2985 *) Add a 'bctest' script that checks for some known 'bc' bugs
2986 so that 'make test' does not abort just because 'bc' is broken.
2987 [Bodo Moeller]
2988
2989 *) Store verify_result within SSL_SESSION also for client side to
2990 avoid potential security hole. (Re-used sessions on the client side
2991 always resulted in verify_result==X509_V_OK, not using the original
2992 result of the server certificate verification.)
2993 [Lutz Jaenicke]
2994
2995 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2996 SSL3_RT_APPLICATION_DATA, return 0.
2997 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2998 [Bodo Moeller]
2999
3000 *) Fix SSL_peek:
3001 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3002 releases, have been re-implemented by renaming the previous
3003 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3004 and ssl3_read_internal, respectively, and adding 'peek' parameters
3005 to them. The new ssl[23]_{read,peek} functions are calls to
3006 ssl[23]_read_internal with the 'peek' flag set appropriately.
3007 A 'peek' parameter has also been added to ssl3_read_bytes, which
3008 does the actual work for ssl3_read_internal.
3009 [Bodo Moeller]
3010
3011 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3012 the method-specific "init()" handler. Also clean up ex_data after
3013 calling the method-specific "finish()" handler. Previously, this was
3014 happening the other way round.
3015 [Geoff Thorpe]
3016
3017 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3018 The previous value, 12, was not always sufficient for BN_mod_exp().
3019 [Bodo Moeller]
3020
3021 *) Make sure that shared libraries get the internal name engine with
3022 the full version number and not just 0. This should mark the
3023 shared libraries as not backward compatible. Of course, this should
3024 be changed again when we can guarantee backward binary compatibility.
3025 [Richard Levitte]
3026
3027 *) Fix typo in get_cert_by_subject() in by_dir.c
3028 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3029
3030 *) Rework the system to generate shared libraries:
3031
3032 - Make note of the expected extension for the shared libraries and
3033 if there is a need for symbolic links from for example libcrypto.so.0
3034 to libcrypto.so.0.9.7. There is extended info in Configure for
3035 that.
3036
3037 - Make as few rebuilds of the shared libraries as possible.
3038
3039 - Still avoid linking the OpenSSL programs with the shared libraries.
3040
3041 - When installing, install the shared libraries separately from the
3042 static ones.
3043 [Richard Levitte]
3044
3045 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3046
3047 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3048 and not in SSL_clear because the latter is also used by the
3049 accept/connect functions; previously, the settings made by
3050 SSL_set_read_ahead would be lost during the handshake.
3051 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3052
3053 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3054 Previously, it would create entries for disableed algorithms no
3055 matter what.
3056 [Richard Levitte]
3057
3058 *) Added several new manual pages for SSL_* function.
3059 [Lutz Jaenicke]
3060
3061 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3062
3063 *) In ssl23_get_client_hello, generate an error message when faced
3064 with an initial SSL 3.0/TLS record that is too small to contain the
3065 first two bytes of the ClientHello message, i.e. client_version.
3066 (Note that this is a pathologic case that probably has never happened
3067 in real life.) The previous approach was to use the version number
3068 from the record header as a substitute; but our protocol choice
3069 should not depend on that one because it is not authenticated
3070 by the Finished messages.
3071 [Bodo Moeller]
3072
3073 *) More robust randomness gathering functions for Windows.
3074 [Jeffrey Altman <jaltman@columbia.edu>]
3075
3076 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3077 not set then we don't setup the error code for issuer check errors
3078 to avoid possibly overwriting other errors which the callback does
3079 handle. If an application does set the flag then we assume it knows
3080 what it is doing and can handle the new informational codes
3081 appropriately.
3082 [Steve Henson]
3083
3084 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3085 a general "ANY" type, as such it should be able to decode anything
3086 including tagged types. However it didn't check the class so it would
3087 wrongly interpret tagged types in the same way as their universal
3088 counterpart and unknown types were just rejected. Changed so that the
3089 tagged and unknown types are handled in the same way as a SEQUENCE:
3090 that is the encoding is stored intact. There is also a new type
3091 "V_ASN1_OTHER" which is used when the class is not universal, in this
3092 case we have no idea what the actual type is so we just lump them all
3093 together.
3094 [Steve Henson]
3095
3096 *) On VMS, stdout may very well lead to a file that is written to
3097 in a record-oriented fashion. That means that every write() will
3098 write a separate record, which will be read separately by the
3099 programs trying to read from it. This can be very confusing.
3100
3101 The solution is to put a BIO filter in the way that will buffer
3102 text until a linefeed is reached, and then write everything a
3103 line at a time, so every record written will be an actual line,
3104 not chunks of lines and not (usually doesn't happen, but I've
3105 seen it once) several lines in one record. BIO_f_linebuffer() is
3106 the answer.
3107
3108 Currently, it's a VMS-only method, because that's where it has
3109 been tested well enough.
3110 [Richard Levitte]
3111
3112 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3113 it can return incorrect results.
3114 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3115 but it was in 0.9.6-beta[12].)
3116 [Bodo Moeller]
3117
3118 *) Disable the check for content being present when verifying detached
3119 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3120 include zero length content when signing messages.
3121 [Steve Henson]
3122
3123 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3124 BIO_ctrl (for BIO pairs).
3125 [Bodo Möller]
3126
3127 *) Add DSO method for VMS.
3128 [Richard Levitte]
3129
3130 *) Bug fix: Montgomery multiplication could produce results with the
3131 wrong sign.
3132 [Ulf Möller]
3133
3134 *) Add RPM specification openssl.spec and modify it to build three
3135 packages. The default package contains applications, application
3136 documentation and run-time libraries. The devel package contains
3137 include files, static libraries and function documentation. The
3138 doc package contains the contents of the doc directory. The original
3139 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3140 [Richard Levitte]
3141
3142 *) Add a large number of documentation files for many SSL routines.
3143 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3144
3145 *) Add a configuration entry for Sony News 4.
3146 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3147
3148 *) Don't set the two most significant bits to one when generating a
3149 random number < q in the DSA library.
3150 [Ulf Möller]
3151
3152 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3153 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3154 the underlying transport is blocking) if a handshake took place.
3155 (The default behaviour is needed by applications such as s_client
3156 and s_server that use select() to determine when to use SSL_read;
3157 but for applications that know in advance when to expect data, it
3158 just makes things more complicated.)
3159 [Bodo Moeller]
3160
3161 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3162 from EGD.
3163 [Ben Laurie]
3164
3165 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3166 work better on such systems.
3167 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3168
3169 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3170 Update PKCS12_parse() so it copies the friendlyName and the
3171 keyid to the certificates aux info.
3172 [Steve Henson]
3173
3174 *) Fix bug in PKCS7_verify() which caused an infinite loop
3175 if there was more than one signature.
3176 [Sven Uszpelkat <su@celocom.de>]
3177
3178 *) Major change in util/mkdef.pl to include extra information
3179 about each symbol, as well as presentig variables as well
3180 as functions. This change means that there's n more need
3181 to rebuild the .num files when some algorithms are excluded.
3182 [Richard Levitte]
3183
3184 *) Allow the verify time to be set by an application,
3185 rather than always using the current time.
3186 [Steve Henson]
3187
3188 *) Phase 2 verify code reorganisation. The certificate
3189 verify code now looks up an issuer certificate by a
3190 number of criteria: subject name, authority key id
3191 and key usage. It also verifies self signed certificates
3192 by the same criteria. The main comparison function is
3193 X509_check_issued() which performs these checks.
3194
3195 Lot of changes were necessary in order to support this
3196 without completely rewriting the lookup code.
3197
3198 Authority and subject key identifier are now cached.
3199
3200 The LHASH 'certs' is X509_STORE has now been replaced
3201 by a STACK_OF(X509_OBJECT). This is mainly because an
3202 LHASH can't store or retrieve multiple objects with
3203 the same hash value.
3204
3205 As a result various functions (which were all internal
3206 use only) have changed to handle the new X509_STORE
3207 structure. This will break anything that messed round
3208 with X509_STORE internally.
3209
3210 The functions X509_STORE_add_cert() now checks for an
3211 exact match, rather than just subject name.
3212
3213 The X509_STORE API doesn't directly support the retrieval
3214 of multiple certificates matching a given criteria, however
3215 this can be worked round by performing a lookup first
3216 (which will fill the cache with candidate certificates)
3217 and then examining the cache for matches. This is probably
3218 the best we can do without throwing out X509_LOOKUP
3219 entirely (maybe later...).
3220
3221 The X509_VERIFY_CTX structure has been enhanced considerably.
3222
3223 All certificate lookup operations now go via a get_issuer()
3224 callback. Although this currently uses an X509_STORE it
3225 can be replaced by custom lookups. This is a simple way
3226 to bypass the X509_STORE hackery necessary to make this
3227 work and makes it possible to use more efficient techniques
3228 in future. A very simple version which uses a simple
3229 STACK for its trusted certificate store is also provided
3230 using X509_STORE_CTX_trusted_stack().
3231
3232 The verify_cb() and verify() callbacks now have equivalents
3233 in the X509_STORE_CTX structure.
3234
3235 X509_STORE_CTX also has a 'flags' field which can be used
3236 to customise the verify behaviour.
3237 [Steve Henson]
3238
3239 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3240 excludes S/MIME capabilities.
3241 [Steve Henson]
3242
3243 *) When a certificate request is read in keep a copy of the
3244 original encoding of the signed data and use it when outputing
3245 again. Signatures then use the original encoding rather than
3246 a decoded, encoded version which may cause problems if the
3247 request is improperly encoded.
3248 [Steve Henson]
3249
3250 *) For consistency with other BIO_puts implementations, call
3251 buffer_write(b, ...) directly in buffer_puts instead of calling
3252 BIO_write(b, ...).
3253
3254 In BIO_puts, increment b->num_write as in BIO_write.
3255 [Peter.Sylvester@EdelWeb.fr]
3256
3257 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3258 BN_zero, we may not return a BIGNUM with an array consisting of
3259 words set to zero.)
3260 [Bodo Moeller]
3261
3262 *) Avoid calling abort() from within the library when problems are
3263 detected, except if preprocessor symbols have been defined
3264 (such as REF_CHECK, BN_DEBUG etc.).
3265 [Bodo Moeller]
3266
3267 *) New openssl application 'rsautl'. This utility can be
3268 used for low level RSA operations. DER public key
3269 BIO/fp routines also added.
3270 [Steve Henson]
3271
3272 *) New Configure entry and patches for compiling on QNX 4.
3273 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3274
3275 *) A demo state-machine implementation was sponsored by
3276 Nuron (http://www.nuron.com/) and is now available in
3277 demos/state_machine.
3278 [Ben Laurie]
3279
3280 *) New options added to the 'dgst' utility for signature
3281 generation and verification.
3282 [Steve Henson]
3283
3284 *) Unrecognized PKCS#7 content types are now handled via a
3285 catch all ASN1_TYPE structure. This allows unsupported
3286 types to be stored as a "blob" and an application can
3287 encode and decode it manually.
3288 [Steve Henson]
3289
3290 *) Fix various signed/unsigned issues to make a_strex.c
3291 compile under VC++.
3292 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3293
3294 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3295 length if passed a buffer. ASN1_INTEGER_to_BN failed
3296 if passed a NULL BN and its argument was negative.
3297 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3298
3299 *) Modification to PKCS#7 encoding routines to output definite
3300 length encoding. Since currently the whole structures are in
3301 memory there's not real point in using indefinite length
3302 constructed encoding. However if OpenSSL is compiled with
3303 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3304 [Steve Henson]
3305
3306 *) Added BIO_vprintf() and BIO_vsnprintf().
3307 [Richard Levitte]
3308
3309 *) Added more prefixes to parse for in the the strings written
3310 through a logging bio, to cover all the levels that are available
3311 through syslog. The prefixes are now:
3312
3313 PANIC, EMERG, EMR => LOG_EMERG
3314 ALERT, ALR => LOG_ALERT
3315 CRIT, CRI => LOG_CRIT
3316 ERROR, ERR => LOG_ERR
3317 WARNING, WARN, WAR => LOG_WARNING
3318 NOTICE, NOTE, NOT => LOG_NOTICE
3319 INFO, INF => LOG_INFO
3320 DEBUG, DBG => LOG_DEBUG
3321
3322 and as before, if none of those prefixes are present at the
3323 beginning of the string, LOG_ERR is chosen.
3324
3325 On Win32, the LOG_* levels are mapped according to this:
3326
3327 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3328 LOG_WARNING => EVENTLOG_WARNING_TYPE
3329 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3330
3331 [Richard Levitte]
3332
3333 *) Made it possible to reconfigure with just the configuration
3334 argument "reconf" or "reconfigure". The command line arguments
3335 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3336 and are retrieved from there when reconfiguring.
3337 [Richard Levitte]
3338
3339 *) MD4 implemented.
3340 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3341
3342 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3343 [Richard Levitte]
3344
3345 *) The obj_dat.pl script was messing up the sorting of object
3346 names. The reason was that it compared the quoted version
3347 of strings as a result "OCSP" > "OCSP Signing" because
3348 " > SPACE. Changed script to store unquoted versions of
3349 names and add quotes on output. It was also omitting some
3350 names from the lookup table if they were given a default
3351 value (that is if SN is missing it is given the same
3352 value as LN and vice versa), these are now added on the
3353 grounds that if an object has a name we should be able to
3354 look it up. Finally added warning output when duplicate
3355 short or long names are found.
3356 [Steve Henson]
3357
3358 *) Changes needed for Tandem NSK.
3359 [Scott Uroff <scott@xypro.com>]
3360
3361 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3362 RSA_padding_check_SSLv23(), special padding was never detected
3363 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3364 version rollback attacks was not effective.
3365
3366 In s23_clnt.c, don't use special rollback-attack detection padding
3367 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3368 client; similarly, in s23_srvr.c, don't do the rollback check if
3369 SSL 2.0 is the only protocol enabled in the server.
3370 [Bodo Moeller]
3371
3372 *) Make it possible to get hexdumps of unprintable data with 'openssl
3373 asn1parse'. By implication, the functions ASN1_parse_dump() and
3374 BIO_dump_indent() are added.
3375 [Richard Levitte]
3376
3377 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3378 these print out strings and name structures based on various
3379 flags including RFC2253 support and proper handling of
3380 multibyte characters. Added options to the 'x509' utility
3381 to allow the various flags to be set.
3382 [Steve Henson]
3383
3384 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3385 Also change the functions X509_cmp_current_time() and
3386 X509_gmtime_adj() work with an ASN1_TIME structure,
3387 this will enable certificates using GeneralizedTime in validity
3388 dates to be checked.
3389 [Steve Henson]
3390
3391 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3392 negative public key encodings) on by default,
3393 NO_NEG_PUBKEY_BUG can be set to disable it.
3394 [Steve Henson]
3395
3396 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3397 content octets. An i2c_ASN1_OBJECT is unnecessary because
3398 the encoding can be trivially obtained from the structure.
3399 [Steve Henson]
3400
3401 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3402 not read locks (CRYPTO_r_[un]lock).
3403 [Bodo Moeller]
3404
3405 *) A first attempt at creating official support for shared
3406 libraries through configuration. I've kept it so the
3407 default is static libraries only, and the OpenSSL programs
3408 are always statically linked for now, but there are
3409 preparations for dynamic linking in place.
3410 This has been tested on Linux and Tru64.
3411 [Richard Levitte]
3412
3413 *) Randomness polling function for Win9x, as described in:
3414 Peter Gutmann, Software Generation of Practically Strong
3415 Random Numbers.
3416 [Ulf Möller]
3417
3418 *) Fix so PRNG is seeded in req if using an already existing
3419 DSA key.
3420 [Steve Henson]
3421
3422 *) New options to smime application. -inform and -outform
3423 allow alternative formats for the S/MIME message including
3424 PEM and DER. The -content option allows the content to be
3425 specified separately. This should allow things like Netscape
3426 form signing output easier to verify.
3427 [Steve Henson]
3428
3429 *) Fix the ASN1 encoding of tags using the 'long form'.
3430 [Steve Henson]
3431
3432 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3433 STRING types. These convert content octets to and from the
3434 underlying type. The actual tag and length octets are
3435 already assumed to have been read in and checked. These
3436 are needed because all other string types have virtually
3437 identical handling apart from the tag. By having versions
3438 of the ASN1 functions that just operate on content octets
3439 IMPLICIT tagging can be handled properly. It also allows
3440 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3441 and ASN1_INTEGER are identical apart from the tag.
3442 [Steve Henson]
3443
3444 *) Change the handling of OID objects as follows:
3445
3446 - New object identifiers are inserted in objects.txt, following
3447 the syntax given in objects.README.
3448 - objects.pl is used to process obj_mac.num and create a new
3449 obj_mac.h.
3450 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3451 obj_mac.h.
3452
3453 This is currently kind of a hack, and the perl code in objects.pl
3454 isn't very elegant, but it works as I intended. The simplest way
3455 to check that it worked correctly is to look in obj_dat.h and
3456 check the array nid_objs and make sure the objects haven't moved
3457 around (this is important!). Additions are OK, as well as
3458 consistent name changes.
3459 [Richard Levitte]
3460
3461 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3462 [Bodo Moeller]
3463
3464 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3465 The given file adds to whatever has already been seeded into the
3466 random pool through the RANDFILE configuration file option or
3467 environment variable, or the default random state file.
3468 [Richard Levitte]
3469
3470 *) mkstack.pl now sorts each macro group into lexical order.
3471 Previously the output order depended on the order the files
3472 appeared in the directory, resulting in needless rewriting
3473 of safestack.h .
3474 [Steve Henson]
3475
3476 *) Patches to make OpenSSL compile under Win32 again. Mostly
3477 work arounds for the VC++ problem that it treats func() as
3478 func(void). Also stripped out the parts of mkdef.pl that
3479 added extra typesafe functions: these no longer exist.
3480 [Steve Henson]
3481
3482 *) Reorganisation of the stack code. The macros are now all
3483 collected in safestack.h . Each macro is defined in terms of
3484 a "stack macro" of the form SKM_<name>(type, a, b). The
3485 DEBUG_SAFESTACK is now handled in terms of function casts,
3486 this has the advantage of retaining type safety without the
3487 use of additional functions. If DEBUG_SAFESTACK is not defined
3488 then the non typesafe macros are used instead. Also modified the
3489 mkstack.pl script to handle the new form. Needs testing to see
3490 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3491 the default if no major problems. Similar behaviour for ASN1_SET_OF
3492 and PKCS12_STACK_OF.
3493 [Steve Henson]
3494
3495 *) When some versions of IIS use the 'NET' form of private key the
3496 key derivation algorithm is different. Normally MD5(password) is
3497 used as a 128 bit RC4 key. In the modified case
3498 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3499 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3500 as the old Netscape_RSA functions except they have an additional
3501 'sgckey' parameter which uses the modified algorithm. Also added
3502 an -sgckey command line option to the rsa utility. Thanks to
3503 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3504 algorithm to openssl-dev.
3505 [Steve Henson]
3506
3507 *) The evp_local.h macros were using 'c.##kname' which resulted in
3508 invalid expansion on some systems (SCO 5.0.5 for example).
3509 Corrected to 'c.kname'.
3510 [Phillip Porch <root@theporch.com>]
3511
3512 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3513 a STACK of email addresses from a certificate or request, these look
3514 in the subject name and the subject alternative name extensions and
3515 omit any duplicate addresses.
3516 [Steve Henson]
3517
3518 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3519 This makes DSA verification about 2 % faster.
3520 [Bodo Moeller]
3521
3522 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3523 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3524 plus overhead for 1024 bit moduli).
3525 This makes exponentiations about 0.5 % faster for 1024 bit
3526 exponents (as measured by "openssl speed rsa2048").
3527 [Bodo Moeller]
3528
3529 *) Rename memory handling macros to avoid conflicts with other
3530 software:
3531 Malloc => OPENSSL_malloc
3532 Malloc_locked => OPENSSL_malloc_locked
3533 Realloc => OPENSSL_realloc
3534 Free => OPENSSL_free
3535 [Richard Levitte]
3536
3537 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3538 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3539 [Bodo Moeller]
3540
3541 *) CygWin32 support.
3542 [John Jarvie <jjarvie@newsguy.com>]
3543
3544 *) The type-safe stack code has been rejigged. It is now only compiled
3545 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3546 by default all type-specific stack functions are "#define"d back to
3547 standard stack functions. This results in more streamlined output
3548 but retains the type-safety checking possibilities of the original
3549 approach.
3550 [Geoff Thorpe]
3551
3552 *) The STACK code has been cleaned up, and certain type declarations
3553 that didn't make a lot of sense have been brought in line. This has
3554 also involved a cleanup of sorts in safestack.h to more correctly
3555 map type-safe stack functions onto their plain stack counterparts.
3556 This work has also resulted in a variety of "const"ifications of
3557 lots of the code, especially "_cmp" operations which should normally
3558 be prototyped with "const" parameters anyway.
3559 [Geoff Thorpe]
3560
3561 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3562 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3563 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3564 where all of 'md' is used each time the PRNG is used, but 'state'
3565 is used only indexed by a cyclic counter. As entropy may not be
3566 well distributed from the beginning, 'md' is important as a
3567 chaining variable. However, the output function chains only half
3568 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3569 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3570 in all of 'state' being rewritten, with the new values depending
3571 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3572 [Bodo Moeller]
3573
3574 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3575 the handshake is continued after ssl_verify_cert_chain();
3576 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3577 can lead to 'unexplainable' connection aborts later.
3578 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3579
3580 *) Major EVP API cipher revision.
3581 Add hooks for extra EVP features. This allows various cipher
3582 parameters to be set in the EVP interface. Support added for variable
3583 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3584 setting of RC2 and RC5 parameters.
3585
3586 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3587 ciphers.
3588
3589 Remove lots of duplicated code from the EVP library. For example *every*
3590 cipher init() function handles the 'iv' in the same way according to the
3591 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3592 for CFB and OFB modes they zero ctx->num.
3593
3594 New functionality allows removal of S/MIME code RC2 hack.
3595
3596 Most of the routines have the same form and so can be declared in terms
3597 of macros.
3598
3599 By shifting this to the top level EVP_CipherInit() it can be removed from
3600 all individual ciphers. If the cipher wants to handle IVs or keys
3601 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3602 flags.
3603
3604 Change lots of functions like EVP_EncryptUpdate() to now return a
3605 value: although software versions of the algorithms cannot fail
3606 any installed hardware versions can.
3607 [Steve Henson]
3608
3609 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3610 this option is set, tolerate broken clients that send the negotiated
3611 protocol version number instead of the requested protocol version
3612 number.
3613 [Bodo Moeller]
3614
3615 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3616 i.e. non-zero for export ciphersuites, zero otherwise.
3617 Previous versions had this flag inverted, inconsistent with
3618 rsa_tmp_cb (..._TMP_RSA_CB).
3619 [Bodo Moeller; problem reported by Amit Chopra]
3620
3621 *) Add missing DSA library text string. Work around for some IIS
3622 key files with invalid SEQUENCE encoding.
3623 [Steve Henson]
3624
3625 *) Add a document (doc/standards.txt) that list all kinds of standards
3626 and so on that are implemented in OpenSSL.
3627 [Richard Levitte]
3628
3629 *) Enhance c_rehash script. Old version would mishandle certificates
3630 with the same subject name hash and wouldn't handle CRLs at all.
3631 Added -fingerprint option to crl utility, to support new c_rehash
3632 features.
3633 [Steve Henson]
3634
3635 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3636 [Ulf Möller]
3637
3638 *) Fix for SSL server purpose checking. Server checking was
3639 rejecting certificates which had extended key usage present
3640 but no ssl client purpose.
3641 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3642
3643 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3644 is a little unclear about how a blank password is handled.
3645 Since the password in encoded as a BMPString with terminating
3646 double NULL a zero length password would end up as just the
3647 double NULL. However no password at all is different and is
3648 handled differently in the PKCS#12 key generation code. NS
3649 treats a blank password as zero length. MSIE treats it as no
3650 password on export: but it will try both on import. We now do
3651 the same: PKCS12_parse() tries zero length and no password if
3652 the password is set to "" or NULL (NULL is now a valid password:
3653 it wasn't before) as does the pkcs12 application.
3654 [Steve Henson]
3655
3656 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3657 perror when PEM_read_bio_X509_REQ fails, the error message must
3658 be obtained from the error queue.
3659 [Bodo Moeller]
3660
3661 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3662 it in ERR_remove_state if appropriate, and change ERR_get_state
3663 accordingly to avoid race conditions (this is necessary because
3664 thread_hash is no longer constant once set).
3665 [Bodo Moeller]
3666
3667 *) Bugfix for linux-elf makefile.one.
3668 [Ulf Möller]
3669
3670 *) RSA_get_default_method() will now cause a default
3671 RSA_METHOD to be chosen if one doesn't exist already.
3672 Previously this was only set during a call to RSA_new()
3673 or RSA_new_method(NULL) meaning it was possible for
3674 RSA_get_default_method() to return NULL.
3675 [Geoff Thorpe]
3676
3677 *) Added native name translation to the existing DSO code
3678 that will convert (if the flag to do so is set) filenames
3679 that are sufficiently small and have no path information
3680 into a canonical native form. Eg. "blah" converted to
3681 "libblah.so" or "blah.dll" etc.
3682 [Geoff Thorpe]
3683
3684 *) New function ERR_error_string_n(e, buf, len) which is like
3685 ERR_error_string(e, buf), but writes at most 'len' bytes
3686 including the 0 terminator. For ERR_error_string_n, 'buf'
3687 may not be NULL.
3688 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3689
3690 *) CONF library reworked to become more general. A new CONF
3691 configuration file reader "class" is implemented as well as a
3692 new functions (NCONF_*, for "New CONF") to handle it. The now
3693 old CONF_* functions are still there, but are reimplemented to
3694 work in terms of the new functions. Also, a set of functions
3695 to handle the internal storage of the configuration data is
3696 provided to make it easier to write new configuration file
3697 reader "classes" (I can definitely see something reading a
3698 configuration file in XML format, for example), called _CONF_*,
3699 or "the configuration storage API"...
3700
3701 The new configuration file reading functions are:
3702
3703 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3704 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3705
3706 NCONF_default, NCONF_WIN32
3707
3708 NCONF_dump_fp, NCONF_dump_bio
3709
3710 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3711 NCONF_new creates a new CONF object. This works in the same way
3712 as other interfaces in OpenSSL, like the BIO interface.
3713 NCONF_dump_* dump the internal storage of the configuration file,
3714 which is useful for debugging. All other functions take the same
3715 arguments as the old CONF_* functions wth the exception of the
3716 first that must be a `CONF *' instead of a `LHASH *'.
3717
3718 To make it easer to use the new classes with the old CONF_* functions,
3719 the function CONF_set_default_method is provided.
3720 [Richard Levitte]
3721
3722 *) Add '-tls1' option to 'openssl ciphers', which was already
3723 mentioned in the documentation but had not been implemented.
3724 (This option is not yet really useful because even the additional
3725 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3726 [Bodo Moeller]
3727
3728 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3729 OpenSSL-based applications) load shared libraries and bind to
3730 them in a portable way.
3731 [Geoff Thorpe, with contributions from Richard Levitte]
3732
3733 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3734
3735 *) Make sure _lrotl and _lrotr are only used with MSVC.
3736
3737 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3738 (the default implementation of RAND_status).
3739
3740 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3741 to '-clrext' (= clear extensions), as intended and documented.
3742 [Bodo Moeller; inconsistency pointed out by Michael Attili
3743 <attili@amaxo.com>]
3744
3745 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3746 was larger than the MD block size.
3747 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3748
3749 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3750 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3751 using the passed key: if the passed key was a private key the result
3752 of X509_print(), for example, would be to print out all the private key
3753 components.
3754 [Steve Henson]
3755
3756 *) des_quad_cksum() byte order bug fix.
3757 [Ulf Möller, using the problem description in krb4-0.9.7, where
3758 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3759
3760 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3761 discouraged.
3762 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3763
3764 *) For easily testing in shell scripts whether some command
3765 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3766 returns with exit code 0 iff no command of the given name is available.
3767 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3768 the output goes to stdout and nothing is printed to stderr.
3769 Additional arguments are always ignored.
3770
3771 Since for each cipher there is a command of the same name,
3772 the 'no-cipher' compilation switches can be tested this way.
3773
3774 ('openssl no-XXX' is not able to detect pseudo-commands such
3775 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3776 [Bodo Moeller]
3777
3778 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3779 [Bodo Moeller]
3780
3781 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3782 is set; it will be thrown away anyway because each handshake creates
3783 its own key.
3784 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3785 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3786 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3787 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3788 [Bodo Moeller]
3789
3790 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3791 'Q' and 'R' lose their special meanings (quit/renegotiate).
3792 This is part of what -quiet does; unlike -quiet, -ign_eof
3793 does not suppress any output.
3794 [Richard Levitte]
3795
3796 *) Add compatibility options to the purpose and trust code. The
3797 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3798 accepts a certificate or CA, this was the previous behaviour,
3799 with all the associated security issues.
3800
3801 X509_TRUST_COMPAT is the old trust behaviour: only and
3802 automatically trust self signed roots in certificate store. A
3803 new trust setting X509_TRUST_DEFAULT is used to specify that
3804 a purpose has no associated trust setting and it should instead
3805 use the value in the default purpose.
3806 [Steve Henson]
3807
3808 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3809 and fix a memory leak.
3810 [Steve Henson]
3811
3812 *) In util/mkerr.pl (which implements 'make errors'), preserve
3813 reason strings from the previous version of the .c file, as
3814 the default to have only downcase letters (and digits) in
3815 automatically generated reasons codes is not always appropriate.
3816 [Bodo Moeller]
3817
3818 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3819 using strerror. Previously, ERR_reason_error_string() returned
3820 library names as reason strings for SYSerr; but SYSerr is a special
3821 case where small numbers are errno values, not library numbers.
3822 [Bodo Moeller]
3823
3824 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3825 converts DSA parameters into DH parameters. (When creating parameters,
3826 DSA_generate_parameters is used.)
3827 [Bodo Moeller]
3828
3829 *) Include 'length' (recommended exponent length) in C code generated
3830 by 'openssl dhparam -C'.
3831 [Bodo Moeller]
3832
3833 *) The second argument to set_label in perlasm was already being used
3834 so couldn't be used as a "file scope" flag. Moved to third argument
3835 which was free.
3836 [Steve Henson]
3837
3838 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3839 instead of RAND_bytes for encryption IVs and salts.
3840 [Bodo Moeller]
3841
3842 *) Include RAND_status() into RAND_METHOD instead of implementing
3843 it only for md_rand.c Otherwise replacing the PRNG by calling
3844 RAND_set_rand_method would be impossible.
3845 [Bodo Moeller]
3846
3847 *) Don't let DSA_generate_key() enter an infinite loop if the random
3848 number generation fails.
3849 [Bodo Moeller]
3850
3851 *) New 'rand' application for creating pseudo-random output.
3852 [Bodo Moeller]
3853
3854 *) Added configuration support for Linux/IA64
3855 [Rolf Haberrecker <rolf@suse.de>]
3856
3857 *) Assembler module support for Mingw32.
3858 [Ulf Möller]
3859
3860 *) Shared library support for HPUX (in shlib/).
3861 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3862
3863 *) Shared library support for Solaris gcc.
3864 [Lutz Behnke <behnke@trustcenter.de>]
3865
3866 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3867
3868 *) PKCS7_encrypt() was adding text MIME headers twice because they
3869 were added manually and by SMIME_crlf_copy().
3870 [Steve Henson]
3871
3872 *) In bntest.c don't call BN_rand with zero bits argument.
3873 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3874
3875 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3876 case was implemented. This caused BN_div_recp() to fail occasionally.
3877 [Ulf Möller]
3878
3879 *) Add an optional second argument to the set_label() in the perl
3880 assembly language builder. If this argument exists and is set
3881 to 1 it signals that the assembler should use a symbol whose
3882 scope is the entire file, not just the current function. This
3883 is needed with MASM which uses the format label:: for this scope.
3884 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3885
3886 *) Change the ASN1 types so they are typedefs by default. Before
3887 almost all types were #define'd to ASN1_STRING which was causing
3888 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3889 for example.
3890 [Steve Henson]
3891
3892 *) Change names of new functions to the new get1/get0 naming
3893 convention: After 'get1', the caller owns a reference count
3894 and has to call ..._free; 'get0' returns a pointer to some
3895 data structure without incrementing reference counters.
3896 (Some of the existing 'get' functions increment a reference
3897 counter, some don't.)
3898 Similarly, 'set1' and 'add1' functions increase reference
3899 counters or duplicate objects.
3900 [Steve Henson]
3901
3902 *) Allow for the possibility of temp RSA key generation failure:
3903 the code used to assume it always worked and crashed on failure.
3904 [Steve Henson]
3905
3906 *) Fix potential buffer overrun problem in BIO_printf().
3907 [Ulf Möller, using public domain code by Patrick Powell; problem
3908 pointed out by David Sacerdote <das33@cornell.edu>]
3909
3910 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3911 RAND_egd() and RAND_status(). In the command line application,
3912 the EGD socket can be specified like a seed file using RANDFILE
3913 or -rand.
3914 [Ulf Möller]
3915
3916 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3917 Some CAs (e.g. Verisign) distribute certificates in this form.
3918 [Steve Henson]
3919
3920 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3921 list to exclude them. This means that no special compilation option
3922 is needed to use anonymous DH: it just needs to be included in the
3923 cipher list.
3924 [Steve Henson]
3925
3926 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3927 EVP_MD_type. The old functionality is available in a new macro called
3928 EVP_MD_md(). Change code that uses it and update docs.
3929 [Steve Henson]
3930
3931 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3932 where the 'void *' argument is replaced by a function pointer argument.
3933 Previously 'void *' was abused to point to functions, which works on
3934 many platforms, but is not correct. As these functions are usually
3935 called by macros defined in OpenSSL header files, most source code
3936 should work without changes.
3937 [Richard Levitte]
3938
3939 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3940 sections with information on -D... compiler switches used for
3941 compiling the library so that applications can see them. To enable
3942 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3943 must be defined. E.g.,
3944 #define OPENSSL_ALGORITHM_DEFINES
3945 #include <openssl/opensslconf.h>
3946 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3947 [Richard Levitte, Ulf and Bodo Möller]
3948
3949 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3950 record layer.
3951 [Bodo Moeller]
3952
3953 *) Change the 'other' type in certificate aux info to a STACK_OF
3954 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3955 the required ASN1 format: arbitrary types determined by an OID.
3956 [Steve Henson]
3957
3958 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3959 argument to 'req'. This is not because the function is newer or
3960 better than others it just uses the work 'NEW' in the certificate
3961 request header lines. Some software needs this.
3962 [Steve Henson]
3963
3964 *) Reorganise password command line arguments: now passwords can be
3965 obtained from various sources. Delete the PEM_cb function and make
3966 it the default behaviour: i.e. if the callback is NULL and the
3967 usrdata argument is not NULL interpret it as a null terminated pass
3968 phrase. If usrdata and the callback are NULL then the pass phrase
3969 is prompted for as usual.
3970 [Steve Henson]
3971
3972 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3973 the support is automatically enabled. The resulting binaries will
3974 autodetect the card and use it if present.
3975 [Ben Laurie and Compaq Inc.]
3976
3977 *) Work around for Netscape hang bug. This sends certificate request
3978 and server done in one record. Since this is perfectly legal in the
3979 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3980 the bugs/SSLv3 entry for more info.
3981 [Steve Henson]
3982
3983 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3984 [Andy Polyakov]
3985
3986 *) Add -rand argument to smime and pkcs12 applications and read/write
3987 of seed file.
3988 [Steve Henson]
3989
3990 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3991 [Bodo Moeller]
3992
3993 *) Add command line password options to the remaining applications.
3994 [Steve Henson]
3995
3996 *) Bug fix for BN_div_recp() for numerators with an even number of
3997 bits.
3998 [Ulf Möller]
3999
4000 *) More tests in bntest.c, and changed test_bn output.
4001 [Ulf Möller]
4002
4003 *) ./config recognizes MacOS X now.
4004 [Andy Polyakov]
4005
4006 *) Bug fix for BN_div() when the first words of num and divsor are
4007 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4008 [Ulf Möller]
4009
4010 *) Add support for various broken PKCS#8 formats, and command line
4011 options to produce them.
4012 [Steve Henson]
4013
4014 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4015 get temporary BIGNUMs from a BN_CTX.
4016 [Ulf Möller]
4017
4018 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4019 for p == 0.
4020 [Ulf Möller]
4021
4022 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4023 include a #define from the old name to the new. The original intent
4024 was that statically linked binaries could for example just call
4025 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4026 link with digests. This never worked becayse SSLeay_add_all_digests()
4027 and SSLeay_add_all_ciphers() were in the same source file so calling
4028 one would link with the other. They are now in separate source files.
4029 [Steve Henson]
4030
4031 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4032 [Steve Henson]
4033
4034 *) Use a less unusual form of the Miller-Rabin primality test (it used
4035 a binary algorithm for exponentiation integrated into the Miller-Rabin
4036 loop, our standard modexp algorithms are faster).
4037 [Bodo Moeller]
4038
4039 *) Support for the EBCDIC character set completed.
4040 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4041
4042 *) Source code cleanups: use const where appropriate, eliminate casts,
4043 use void * instead of char * in lhash.
4044 [Ulf Möller]
4045
4046 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4047 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4048 this the server could overwrite ephemeral keys that the client
4049 has already seen).
4050 [Bodo Moeller]
4051
4052 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4053 using 50 iterations of the Rabin-Miller test.
4054
4055 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4056 iterations of the Rabin-Miller test as required by the appendix
4057 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4058 As BN_is_prime_fasttest includes trial division, DSA parameter
4059 generation becomes much faster.
4060
4061 This implies a change for the callback functions in DSA_is_prime
4062 and DSA_generate_parameters: The callback function is called once
4063 for each positive witness in the Rabin-Miller test, not just
4064 occasionally in the inner loop; and the parameters to the
4065 callback function now provide an iteration count for the outer
4066 loop rather than for the current invocation of the inner loop.
4067 DSA_generate_parameters additionally can call the callback
4068 function with an 'iteration count' of -1, meaning that a
4069 candidate has passed the trial division test (when q is generated
4070 from an application-provided seed, trial division is skipped).
4071 [Bodo Moeller]
4072
4073 *) New function BN_is_prime_fasttest that optionally does trial
4074 division before starting the Rabin-Miller test and has
4075 an additional BN_CTX * argument (whereas BN_is_prime always
4076 has to allocate at least one BN_CTX).
4077 'callback(1, -1, cb_arg)' is called when a number has passed the
4078 trial division stage.
4079 [Bodo Moeller]
4080
4081 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4082 as ASN1_TIME.
4083 [Steve Henson]
4084
4085 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4086 [Steve Henson]
4087
4088 *) New function BN_pseudo_rand().
4089 [Ulf Möller]
4090
4091 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4092 bignum version of BN_from_montgomery() with the working code from
4093 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4094 the comments.
4095 [Ulf Möller]
4096
4097 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4098 made it impossible to use the same SSL_SESSION data structure in
4099 SSL2 clients in multiple threads.
4100 [Bodo Moeller]
4101
4102 *) The return value of RAND_load_file() no longer counts bytes obtained
4103 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4104 to seed the PRNG (previously an explicit byte count was required).
4105 [Ulf Möller, Bodo Möller]
4106
4107 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4108 used (char *) instead of (void *) and had casts all over the place.
4109 [Steve Henson]
4110
4111 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4112 [Ulf Möller]
4113
4114 *) Retain source code compatibility for BN_prime_checks macro:
4115 BN_is_prime(..., BN_prime_checks, ...) now uses
4116 BN_prime_checks_for_size to determine the appropriate number of
4117 Rabin-Miller iterations.
4118 [Ulf Möller]
4119
4120 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4121 DH_CHECK_P_NOT_SAFE_PRIME.
4122 (Check if this is true? OpenPGP calls them "strong".)
4123 [Ulf Möller]
4124
4125 *) Merge the functionality of "dh" and "gendh" programs into a new program
4126 "dhparam". The old programs are retained for now but will handle DH keys
4127 (instead of parameters) in future.
4128 [Steve Henson]
4129
4130 *) Make the ciphers, s_server and s_client programs check the return values
4131 when a new cipher list is set.
4132 [Steve Henson]
4133
4134 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4135 ciphers. Before when the 56bit ciphers were enabled the sorting was
4136 wrong.
4137
4138 The syntax for the cipher sorting has been extended to support sorting by
4139 cipher-strength (using the strength_bits hard coded in the tables).
4140 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4141
4142 Fix a bug in the cipher-command parser: when supplying a cipher command
4143 string with an "undefined" symbol (neither command nor alphanumeric
4144 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4145 an error is flagged.
4146
4147 Due to the strength-sorting extension, the code of the
4148 ssl_create_cipher_list() function was completely rearranged. I hope that
4149 the readability was also increased :-)
4150 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4151
4152 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4153 for the first serial number and places 2 in the serial number file. This
4154 avoids problems when the root CA is created with serial number zero and
4155 the first user certificate has the same issuer name and serial number
4156 as the root CA.
4157 [Steve Henson]
4158
4159 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4160 the new code. Add documentation for this stuff.
4161 [Steve Henson]
4162
4163 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4164 X509_*() to X509at_*() on the grounds that they don't handle X509
4165 structures and behave in an analagous way to the X509v3 functions:
4166 they shouldn't be called directly but wrapper functions should be used
4167 instead.
4168
4169 So we also now have some wrapper functions that call the X509at functions
4170 when passed certificate requests. (TO DO: similar things can be done with
4171 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4172 things. Some of these need some d2i or i2d and print functionality
4173 because they handle more complex structures.)
4174 [Steve Henson]
4175
4176 *) Add missing #ifndefs that caused missing symbols when building libssl
4177 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4178 NO_RSA in ssl/s2*.c.
4179 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4180
4181 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4182 has a return value which indicates the quality of the random data
4183 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4184 error queue. New function RAND_pseudo_bytes() generates output that is
4185 guaranteed to be unique but not unpredictable. RAND_add is like
4186 RAND_seed, but takes an extra argument for an entropy estimate
4187 (RAND_seed always assumes full entropy).
4188 [Ulf Möller]
4189
4190 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4191 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4192 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4193 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4194 false-positive rate of at most 2^-80 for random input.
4195 [Bodo Moeller]
4196
4197 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4198 [Bodo Moeller]
4199
4200 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4201 in the 0.9.5 release), this returns the chain
4202 from an X509_CTX structure with a dup of the stack and all
4203 the X509 reference counts upped: so the stack will exist
4204 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4205 to use this.
4206
4207 Also make SSL_SESSION_print() print out the verify return
4208 code.
4209 [Steve Henson]
4210
4211 *) Add manpage for the pkcs12 command. Also change the default
4212 behaviour so MAC iteration counts are used unless the new
4213 -nomaciter option is used. This improves file security and
4214 only older versions of MSIE (4.0 for example) need it.
4215 [Steve Henson]
4216
4217 *) Honor the no-xxx Configure options when creating .DEF files.
4218 [Ulf Möller]
4219
4220 *) Add PKCS#10 attributes to field table: challengePassword,
4221 unstructuredName and unstructuredAddress. These are taken from
4222 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4223 international characters are used.
4224
4225 More changes to X509_ATTRIBUTE code: allow the setting of types
4226 based on strings. Remove the 'loc' parameter when adding
4227 attributes because these will be a SET OF encoding which is sorted
4228 in ASN1 order.
4229 [Steve Henson]
4230
4231 *) Initial changes to the 'req' utility to allow request generation
4232 automation. This will allow an application to just generate a template
4233 file containing all the field values and have req construct the
4234 request.
4235
4236 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4237 used all over the place including certificate requests and PKCS#7
4238 structures. They are currently handled manually where necessary with
4239 some primitive wrappers for PKCS#7. The new functions behave in a
4240 manner analogous to the X509 extension functions: they allow
4241 attributes to be looked up by NID and added.
4242
4243 Later something similar to the X509V3 code would be desirable to
4244 automatically handle the encoding, decoding and printing of the
4245 more complex types. The string types like challengePassword can
4246 be handled by the string table functions.
4247
4248 Also modified the multi byte string table handling. Now there is
4249 a 'global mask' which masks out certain types. The table itself
4250 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4251 is useful when for example there is only one permissible type
4252 (as in countryName) and using the mask might result in no valid
4253 types at all.
4254 [Steve Henson]
4255
4256 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4257 SSL_get_peer_finished to allow applications to obtain the latest
4258 Finished messages sent to the peer or expected from the peer,
4259 respectively. (SSL_get_peer_finished is usually the Finished message
4260 actually received from the peer, otherwise the protocol will be aborted.)
4261
4262 As the Finished message are message digests of the complete handshake
4263 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4264 be used for external authentication procedures when the authentication
4265 provided by SSL/TLS is not desired or is not enough.
4266 [Bodo Moeller]
4267
4268 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4269 the host supports BWX extension and if Compaq C is present on the
4270 $PATH. Just exploiting of the BWX extension results in 20-30%
4271 performance kick for some algorithms, e.g. DES and RC4 to mention
4272 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4273 SHA1.
4274 [Andy Polyakov]
4275
4276 *) Add support for MS "fast SGC". This is arguably a violation of the
4277 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4278 weak crypto and after checking the certificate is SGC a second one
4279 with strong crypto. MS SGC stops the first handshake after receiving
4280 the server certificate message and sends a second client hello. Since
4281 a server will typically do all the time consuming operations before
4282 expecting any further messages from the client (server key exchange
4283 is the most expensive) there is little difference between the two.
4284
4285 To get OpenSSL to support MS SGC we have to permit a second client
4286 hello message after we have sent server done. In addition we have to
4287 reset the MAC if we do get this second client hello.
4288 [Steve Henson]
4289
4290 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4291 if a DER encoded private key is RSA or DSA traditional format. Changed
4292 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4293 format DER encoded private key. Newer code should use PKCS#8 format which
4294 has the key type encoded in the ASN1 structure. Added DER private key
4295 support to pkcs8 application.
4296 [Steve Henson]
4297
4298 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4299 ciphersuites has been selected (as required by the SSL 3/TLS 1
4300 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4301 is set, we interpret this as a request to violate the specification
4302 (the worst that can happen is a handshake failure, and 'correct'
4303 behaviour would result in a handshake failure anyway).
4304 [Bodo Moeller]
4305
4306 *) In SSL_CTX_add_session, take into account that there might be multiple
4307 SSL_SESSION structures with the same session ID (e.g. when two threads
4308 concurrently obtain them from an external cache).
4309 The internal cache can handle only one SSL_SESSION with a given ID,
4310 so if there's a conflict, we now throw out the old one to achieve
4311 consistency.
4312 [Bodo Moeller]
4313
4314 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4315 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4316 some routines that use cipher OIDs: some ciphers do not have OIDs
4317 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4318 example.
4319 [Steve Henson]
4320
4321 *) Simplify the trust setting structure and code. Now we just have
4322 two sequences of OIDs for trusted and rejected settings. These will
4323 typically have values the same as the extended key usage extension
4324 and any application specific purposes.
4325
4326 The trust checking code now has a default behaviour: it will just
4327 check for an object with the same NID as the passed id. Functions can
4328 be provided to override either the default behaviour or the behaviour
4329 for a given id. SSL client, server and email already have functions
4330 in place for compatibility: they check the NID and also return "trusted"
4331 if the certificate is self signed.
4332 [Steve Henson]
4333
4334 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4335 traditional format into an EVP_PKEY structure.
4336 [Steve Henson]
4337
4338 *) Add a password callback function PEM_cb() which either prompts for
4339 a password if usr_data is NULL or otherwise assumes it is a null
4340 terminated password. Allow passwords to be passed on command line
4341 environment or config files in a few more utilities.
4342 [Steve Henson]
4343
4344 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4345 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4346 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4347 Update documentation.
4348 [Steve Henson]
4349
4350 *) Support for ASN1 "NULL" type. This could be handled before by using
4351 ASN1_TYPE but there wasn't any function that would try to read a NULL
4352 and produce an error if it couldn't. For compatibility we also have
4353 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4354 don't allocate anything because they don't need to.
4355 [Steve Henson]
4356
4357 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4358 for details.
4359 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4360
4361 *) Rebuild of the memory allocation routines used by OpenSSL code and
4362 possibly others as well. The purpose is to make an interface that
4363 provide hooks so anyone can build a separate set of allocation and
4364 deallocation routines to be used by OpenSSL, for example memory
4365 pool implementations, or something else, which was previously hard
4366 since Malloc(), Realloc() and Free() were defined as macros having
4367 the values malloc, realloc and free, respectively (except for Win32
4368 compilations). The same is provided for memory debugging code.
4369 OpenSSL already comes with functionality to find memory leaks, but
4370 this gives people a chance to debug other memory problems.
4371
4372 With these changes, a new set of functions and macros have appeared:
4373
4374 CRYPTO_set_mem_debug_functions() [F]
4375 CRYPTO_get_mem_debug_functions() [F]
4376 CRYPTO_dbg_set_options() [F]
4377 CRYPTO_dbg_get_options() [F]
4378 CRYPTO_malloc_debug_init() [M]
4379
4380 The memory debug functions are NULL by default, unless the library
4381 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4382 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4383 gives the standard debugging functions that come with OpenSSL) or
4384 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4385 provided by the library user) must be used. When the standard
4386 debugging functions are used, CRYPTO_dbg_set_options can be used to
4387 request additional information:
4388 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4389 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4390
4391 Also, things like CRYPTO_set_mem_functions will always give the
4392 expected result (the new set of functions is used for allocation
4393 and deallocation) at all times, regardless of platform and compiler
4394 options.
4395
4396 To finish it up, some functions that were never use in any other
4397 way than through macros have a new API and new semantic:
4398
4399 CRYPTO_dbg_malloc()
4400 CRYPTO_dbg_realloc()
4401 CRYPTO_dbg_free()
4402
4403 All macros of value have retained their old syntax.
4404 [Richard Levitte and Bodo Moeller]
4405
4406 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4407 ordering of SMIMECapabilities wasn't in "strength order" and there
4408 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4409 algorithm.
4410 [Steve Henson]
4411
4412 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4413 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4414 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4415
4416 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4417 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4418 functionality to handle multipart/signed properly) and a utility
4419 called 'smime' to call all this stuff. This is based on code I
4420 originally wrote for Celo who have kindly allowed it to be
4421 included in OpenSSL.
4422 [Steve Henson]
4423
4424 *) Add variants des_set_key_checked and des_set_key_unchecked of
4425 des_set_key (aka des_key_sched). Global variable des_check_key
4426 decides which of these is called by des_set_key; this way
4427 des_check_key behaves as it always did, but applications and
4428 the library itself, which was buggy for des_check_key == 1,
4429 have a cleaner way to pick the version they need.
4430 [Bodo Moeller]
4431
4432 *) New function PKCS12_newpass() which changes the password of a
4433 PKCS12 structure.
4434 [Steve Henson]
4435
4436 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4437 dynamic mix. In both cases the ids can be used as an index into the
4438 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4439 functions so they accept a list of the field values and the
4440 application doesn't need to directly manipulate the X509_TRUST
4441 structure.
4442 [Steve Henson]
4443
4444 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4445 need initialising.
4446 [Steve Henson]
4447
4448 *) Modify the way the V3 extension code looks up extensions. This now
4449 works in a similar way to the object code: we have some "standard"
4450 extensions in a static table which is searched with OBJ_bsearch()
4451 and the application can add dynamic ones if needed. The file
4452 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4453 updated whenever a new extension is added to the core code and kept
4454 in ext_nid order. There is a simple program 'tabtest.c' which checks
4455 this. New extensions are not added too often so this file can readily
4456 be maintained manually.
4457
4458 There are two big advantages in doing things this way. The extensions
4459 can be looked up immediately and no longer need to be "added" using
4460 X509V3_add_standard_extensions(): this function now does nothing.
4461 [Side note: I get *lots* of email saying the extension code doesn't
4462 work because people forget to call this function]
4463 Also no dynamic allocation is done unless new extensions are added:
4464 so if we don't add custom extensions there is no need to call
4465 X509V3_EXT_cleanup().
4466 [Steve Henson]
4467
4468 *) Modify enc utility's salting as follows: make salting the default. Add a
4469 magic header, so unsalted files fail gracefully instead of just decrypting
4470 to garbage. This is because not salting is a big security hole, so people
4471 should be discouraged from doing it.
4472 [Ben Laurie]
4473
4474 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4475 digest to be passed on the command line but it only used this
4476 parameter when signing a certificate. Modified so all relevant
4477 operations are affected by the digest parameter including the
4478 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4479 DSA key was used because it didn't fix the digest.
4480 [Steve Henson]
4481
4482 *) Initial certificate chain verify code. Currently tests the untrusted
4483 certificates for consistency with the verify purpose (which is set
4484 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4485
4486 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4487 this is because it will reject chains with invalid extensions whereas
4488 every previous version of OpenSSL and SSLeay made no checks at all.
4489
4490 Trust code: checks the root CA for the relevant trust settings. Trust
4491 settings have an initial value consistent with the verify purpose: e.g.
4492 if the verify purpose is for SSL client use it expects the CA to be
4493 trusted for SSL client use. However the default value can be changed to
4494 permit custom trust settings: one example of this would be to only trust
4495 certificates from a specific "secure" set of CAs.
4496
4497 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4498 which should be used for version portability: especially since the
4499 verify structure is likely to change more often now.
4500
4501 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4502 to set them. If not set then assume SSL clients will verify SSL servers
4503 and vice versa.
4504
4505 Two new options to the verify program: -untrusted allows a set of
4506 untrusted certificates to be passed in and -purpose which sets the
4507 intended purpose of the certificate. If a purpose is set then the
4508 new chain verify code is used to check extension consistency.
4509 [Steve Henson]
4510
4511 *) Support for the authority information access extension.
4512 [Steve Henson]
4513
4514 *) Modify RSA and DSA PEM read routines to transparently handle
4515 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4516 public keys in a format compatible with certificate
4517 SubjectPublicKeyInfo structures. Unfortunately there were already
4518 functions called *_PublicKey_* which used various odd formats so
4519 these are retained for compatibility: however the DSA variants were
4520 never in a public release so they have been deleted. Changed dsa/rsa
4521 utilities to handle the new format: note no releases ever handled public
4522 keys so we should be OK.
4523
4524 The primary motivation for this change is to avoid the same fiasco
4525 that dogs private keys: there are several incompatible private key
4526 formats some of which are standard and some OpenSSL specific and
4527 require various evil hacks to allow partial transparent handling and
4528 even then it doesn't work with DER formats. Given the option anything
4529 other than PKCS#8 should be dumped: but the other formats have to
4530 stay in the name of compatibility.
4531
4532 With public keys and the benefit of hindsight one standard format
4533 is used which works with EVP_PKEY, RSA or DSA structures: though
4534 it clearly returns an error if you try to read the wrong kind of key.
4535
4536 Added a -pubkey option to the 'x509' utility to output the public key.
4537 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4538 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4539 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4540 that do the same as the EVP_PKEY_assign_*() except they up the
4541 reference count of the added key (they don't "swallow" the
4542 supplied key).
4543 [Steve Henson]
4544
4545 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4546 CRLs would fail if the file contained no certificates or no CRLs:
4547 added a new function to read in both types and return the number
4548 read: this means that if none are read it will be an error. The
4549 DER versions of the certificate and CRL reader would always fail
4550 because it isn't possible to mix certificates and CRLs in DER format
4551 without choking one or the other routine. Changed this to just read
4552 a certificate: this is the best we can do. Also modified the code
4553 in apps/verify.c to take notice of return codes: it was previously
4554 attempting to read in certificates from NULL pointers and ignoring
4555 any errors: this is one reason why the cert and CRL reader seemed
4556 to work. It doesn't check return codes from the default certificate
4557 routines: these may well fail if the certificates aren't installed.
4558 [Steve Henson]
4559
4560 *) Code to support otherName option in GeneralName.
4561 [Steve Henson]
4562
4563 *) First update to verify code. Change the verify utility
4564 so it warns if it is passed a self signed certificate:
4565 for consistency with the normal behaviour. X509_verify
4566 has been modified to it will now verify a self signed
4567 certificate if *exactly* the same certificate appears
4568 in the store: it was previously impossible to trust a
4569 single self signed certificate. This means that:
4570 openssl verify ss.pem
4571 now gives a warning about a self signed certificate but
4572 openssl verify -CAfile ss.pem ss.pem
4573 is OK.
4574 [Steve Henson]
4575
4576 *) For servers, store verify_result in SSL_SESSION data structure
4577 (and add it to external session representation).
4578 This is needed when client certificate verifications fails,
4579 but an application-provided verification callback (set by
4580 SSL_CTX_set_cert_verify_callback) allows accepting the session
4581 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4582 but returns 1): When the session is reused, we have to set
4583 ssl->verify_result to the appropriate error code to avoid
4584 security holes.
4585 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4586
4587 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4588 case in PKCS7_dataInit() where the signed PKCS7 structure
4589 didn't contain any existing data because it was being created.
4590 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4591
4592 *) Add a salt to the key derivation routines in enc.c. This
4593 forms the first 8 bytes of the encrypted file. Also add a
4594 -S option to allow a salt to be input on the command line.
4595 [Steve Henson]
4596
4597 *) New function X509_cmp(). Oddly enough there wasn't a function
4598 to compare two certificates. We do this by working out the SHA1
4599 hash and comparing that. X509_cmp() will be needed by the trust
4600 code.
4601 [Steve Henson]
4602
4603 *) SSL_get1_session() is like SSL_get_session(), but increments
4604 the reference count in the SSL_SESSION returned.
4605 [Geoff Thorpe <geoff@eu.c2.net>]
4606
4607 *) Fix for 'req': it was adding a null to request attributes.
4608 Also change the X509_LOOKUP and X509_INFO code to handle
4609 certificate auxiliary information.
4610 [Steve Henson]
4611
4612 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4613 the 'enc' command.
4614 [Steve Henson]
4615
4616 *) Add the possibility to add extra information to the memory leak
4617 detecting output, to form tracebacks, showing from where each
4618 allocation was originated: CRYPTO_push_info("constant string") adds
4619 the string plus current file name and line number to a per-thread
4620 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4621 is like calling CYRPTO_pop_info() until the stack is empty.
4622 Also updated memory leak detection code to be multi-thread-safe.
4623 [Richard Levitte]
4624
4625 *) Add options -text and -noout to pkcs7 utility and delete the
4626 encryption options which never did anything. Update docs.
4627 [Steve Henson]
4628
4629 *) Add options to some of the utilities to allow the pass phrase
4630 to be included on either the command line (not recommended on
4631 OSes like Unix) or read from the environment. Update the
4632 manpages and fix a few bugs.
4633 [Steve Henson]
4634
4635 *) Add a few manpages for some of the openssl commands.
4636 [Steve Henson]
4637
4638 *) Fix the -revoke option in ca. It was freeing up memory twice,
4639 leaking and not finding already revoked certificates.
4640 [Steve Henson]
4641
4642 *) Extensive changes to support certificate auxiliary information.
4643 This involves the use of X509_CERT_AUX structure and X509_AUX
4644 functions. An X509_AUX function such as PEM_read_X509_AUX()
4645 can still read in a certificate file in the usual way but it
4646 will also read in any additional "auxiliary information". By
4647 doing things this way a fair degree of compatibility can be
4648 retained: existing certificates can have this information added
4649 using the new 'x509' options.
4650
4651 Current auxiliary information includes an "alias" and some trust
4652 settings. The trust settings will ultimately be used in enhanced
4653 certificate chain verification routines: currently a certificate
4654 can only be trusted if it is self signed and then it is trusted
4655 for all purposes.
4656 [Steve Henson]
4657
4658 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4659 The problem was that one of the replacement routines had not been working
4660 since SSLeay releases. For now the offending routine has been replaced
4661 with non-optimised assembler. Even so, this now gives around 95%
4662 performance improvement for 1024 bit RSA signs.
4663 [Mark Cox]
4664
4665 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4666 handling. Most clients have the effective key size in bits equal to
4667 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4668 A few however don't do this and instead use the size of the decrypted key
4669 to determine the RC2 key length and the AlgorithmIdentifier to determine
4670 the effective key length. In this case the effective key length can still
4671 be 40 bits but the key length can be 168 bits for example. This is fixed
4672 by manually forcing an RC2 key into the EVP_PKEY structure because the
4673 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4674 the key length and effective key length are equal.
4675 [Steve Henson]
4676
4677 *) Add a bunch of functions that should simplify the creation of
4678 X509_NAME structures. Now you should be able to do:
4679 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4680 and have it automatically work out the correct field type and fill in
4681 the structures. The more adventurous can try:
4682 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4683 and it will (hopefully) work out the correct multibyte encoding.
4684 [Steve Henson]
4685
4686 *) Change the 'req' utility to use the new field handling and multibyte
4687 copy routines. Before the DN field creation was handled in an ad hoc
4688 way in req, ca, and x509 which was rather broken and didn't support
4689 BMPStrings or UTF8Strings. Since some software doesn't implement
4690 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4691 using the dirstring_type option. See the new comment in the default
4692 openssl.cnf for more info.
4693 [Steve Henson]
4694
4695 *) Make crypto/rand/md_rand.c more robust:
4696 - Assure unique random numbers after fork().
4697 - Make sure that concurrent threads access the global counter and
4698 md serializably so that we never lose entropy in them
4699 or use exactly the same state in multiple threads.
4700 Access to the large state is not always serializable because
4701 the additional locking could be a performance killer, and
4702 md should be large enough anyway.
4703 [Bodo Moeller]
4704
4705 *) New file apps/app_rand.c with commonly needed functionality
4706 for handling the random seed file.
4707
4708 Use the random seed file in some applications that previously did not:
4709 ca,
4710 dsaparam -genkey (which also ignored its '-rand' option),
4711 s_client,
4712 s_server,
4713 x509 (when signing).
4714 Except on systems with /dev/urandom, it is crucial to have a random
4715 seed file at least for key creation, DSA signing, and for DH exchanges;
4716 for RSA signatures we could do without one.
4717
4718 gendh and gendsa (unlike genrsa) used to read only the first byte
4719 of each file listed in the '-rand' option. The function as previously
4720 found in genrsa is now in app_rand.c and is used by all programs
4721 that support '-rand'.
4722 [Bodo Moeller]
4723
4724 *) In RAND_write_file, use mode 0600 for creating files;
4725 don't just chmod when it may be too late.
4726 [Bodo Moeller]
4727
4728 *) Report an error from X509_STORE_load_locations
4729 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4730 [Bill Perry]
4731
4732 *) New function ASN1_mbstring_copy() this copies a string in either
4733 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4734 into an ASN1_STRING type. A mask of permissible types is passed
4735 and it chooses the "minimal" type to use or an error if not type
4736 is suitable.
4737 [Steve Henson]
4738
4739 *) Add function equivalents to the various macros in asn1.h. The old
4740 macros are retained with an M_ prefix. Code inside the library can
4741 use the M_ macros. External code (including the openssl utility)
4742 should *NOT* in order to be "shared library friendly".
4743 [Steve Henson]
4744
4745 *) Add various functions that can check a certificate's extensions
4746 to see if it usable for various purposes such as SSL client,
4747 server or S/MIME and CAs of these types. This is currently
4748 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4749 verification. Also added a -purpose flag to x509 utility to
4750 print out all the purposes.
4751 [Steve Henson]
4752
4753 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4754 functions.
4755 [Steve Henson]
4756
4757 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4758 for, obtain and decode and extension and obtain its critical flag.
4759 This allows all the necessary extension code to be handled in a
4760 single function call.
4761 [Steve Henson]
4762
4763 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4764 platforms. See crypto/rc4/rc4_enc.c for further details.
4765 [Andy Polyakov]
4766
4767 *) New -noout option to asn1parse. This causes no output to be produced
4768 its main use is when combined with -strparse and -out to extract data
4769 from a file (which may not be in ASN.1 format).
4770 [Steve Henson]
4771
4772 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4773 when producing the local key id.
4774 [Richard Levitte <levitte@stacken.kth.se>]
4775
4776 *) New option -dhparam in s_server. This allows a DH parameter file to be
4777 stated explicitly. If it is not stated then it tries the first server
4778 certificate file. The previous behaviour hard coded the filename
4779 "server.pem".
4780 [Steve Henson]
4781
4782 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4783 a public key to be input or output. For example:
4784 openssl rsa -in key.pem -pubout -out pubkey.pem
4785 Also added necessary DSA public key functions to handle this.
4786 [Steve Henson]
4787
4788 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4789 in the message. This was handled by allowing
4790 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4791 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4792
4793 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4794 to the end of the strings whereas this didn't. This would cause problems
4795 if strings read with d2i_ASN1_bytes() were later modified.
4796 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4797
4798 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4799 data and it contains EOF it will end up returning an error. This is
4800 caused by input 46 bytes long. The cause is due to the way base64
4801 BIOs find the start of base64 encoded data. They do this by trying a
4802 trial decode on each line until they find one that works. When they
4803 do a flag is set and it starts again knowing it can pass all the
4804 data directly through the decoder. Unfortunately it doesn't reset
4805 the context it uses. This means that if EOF is reached an attempt
4806 is made to pass two EOFs through the context and this causes the
4807 resulting error. This can also cause other problems as well. As is
4808 usual with these problems it takes *ages* to find and the fix is
4809 trivial: move one line.
4810 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4811
4812 *) Ugly workaround to get s_client and s_server working under Windows. The
4813 old code wouldn't work because it needed to select() on sockets and the
4814 tty (for keypresses and to see if data could be written). Win32 only
4815 supports select() on sockets so we select() with a 1s timeout on the
4816 sockets and then see if any characters are waiting to be read, if none
4817 are present then we retry, we also assume we can always write data to
4818 the tty. This isn't nice because the code then blocks until we've
4819 received a complete line of data and it is effectively polling the
4820 keyboard at 1s intervals: however it's quite a bit better than not
4821 working at all :-) A dedicated Windows application might handle this
4822 with an event loop for example.
4823 [Steve Henson]
4824
4825 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4826 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4827 will be called when RSA_sign() and RSA_verify() are used. This is useful
4828 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4829 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4830 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4831 This necessitated the support of an extra signature type NID_md5_sha1
4832 for SSL signatures and modifications to the SSL library to use it instead
4833 of calling RSA_public_decrypt() and RSA_private_encrypt().
4834 [Steve Henson]
4835
4836 *) Add new -verify -CAfile and -CApath options to the crl program, these
4837 will lookup a CRL issuers certificate and verify the signature in a
4838 similar way to the verify program. Tidy up the crl program so it
4839 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4840 less strict. It will now permit CRL extensions even if it is not
4841 a V2 CRL: this will allow it to tolerate some broken CRLs.
4842 [Steve Henson]
4843
4844 *) Initialize all non-automatic variables each time one of the openssl
4845 sub-programs is started (this is necessary as they may be started
4846 multiple times from the "OpenSSL>" prompt).
4847 [Lennart Bang, Bodo Moeller]
4848
4849 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4850 removing all other RSA functionality (this is what NO_RSA does). This
4851 is so (for example) those in the US can disable those operations covered
4852 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4853 key generation.
4854 [Steve Henson]
4855
4856 *) Non-copying interface to BIO pairs.
4857 (still largely untested)
4858 [Bodo Moeller]
4859
4860 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4861 ASCII string. This was handled independently in various places before.
4862 [Steve Henson]
4863
4864 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4865 UTF8 strings a character at a time.
4866 [Steve Henson]
4867
4868 *) Use client_version from client hello to select the protocol
4869 (s23_srvr.c) and for RSA client key exchange verification
4870 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4871 [Bodo Moeller]
4872
4873 *) Add various utility functions to handle SPKACs, these were previously
4874 handled by poking round in the structure internals. Added new function
4875 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4876 print, verify and generate SPKACs. Based on an original idea from
4877 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4878 [Steve Henson]
4879
4880 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4881 [Andy Polyakov]
4882
4883 *) Allow the config file extension section to be overwritten on the
4884 command line. Based on an original idea from Massimiliano Pala
4885 <madwolf@comune.modena.it>. The new option is called -extensions
4886 and can be applied to ca, req and x509. Also -reqexts to override
4887 the request extensions in req and -crlexts to override the crl extensions
4888 in ca.
4889 [Steve Henson]
4890
4891 *) Add new feature to the SPKAC handling in ca. Now you can include
4892 the same field multiple times by preceding it by "XXXX." for example:
4893 1.OU="Unit name 1"
4894 2.OU="Unit name 2"
4895 this is the same syntax as used in the req config file.
4896 [Steve Henson]
4897
4898 *) Allow certificate extensions to be added to certificate requests. These
4899 are specified in a 'req_extensions' option of the req section of the
4900 config file. They can be printed out with the -text option to req but
4901 are otherwise ignored at present.
4902 [Steve Henson]
4903
4904 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4905 data read consists of only the final block it would not decrypted because
4906 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4907 A misplaced 'break' also meant the decrypted final block might not be
4908 copied until the next read.
4909 [Steve Henson]
4910
4911 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4912 a few extra parameters to the DH structure: these will be useful if
4913 for example we want the value of 'q' or implement X9.42 DH.
4914 [Steve Henson]
4915
4916 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4917 provides hooks that allow the default DSA functions or functions on a
4918 "per key" basis to be replaced. This allows hardware acceleration and
4919 hardware key storage to be handled without major modification to the
4920 library. Also added low level modexp hooks and CRYPTO_EX structure and
4921 associated functions.
4922 [Steve Henson]
4923
4924 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4925 as "read only": it can't be written to and the buffer it points to will
4926 not be freed. Reading from a read only BIO is much more efficient than
4927 a normal memory BIO. This was added because there are several times when
4928 an area of memory needs to be read from a BIO. The previous method was
4929 to create a memory BIO and write the data to it, this results in two
4930 copies of the data and an O(n^2) reading algorithm. There is a new
4931 function BIO_new_mem_buf() which creates a read only memory BIO from
4932 an area of memory. Also modified the PKCS#7 routines to use read only
4933 memory BIOs.
4934 [Steve Henson]
4935
4936 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4937 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4938 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4939 but a retry condition occured while trying to read the rest.
4940 [Bodo Moeller]
4941
4942 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4943 NID_pkcs7_encrypted by default: this was wrong since this should almost
4944 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4945 the encrypted data type: this is a more sensible place to put it and it
4946 allows the PKCS#12 code to be tidied up that duplicated this
4947 functionality.
4948 [Steve Henson]
4949
4950 *) Changed obj_dat.pl script so it takes its input and output files on
4951 the command line. This should avoid shell escape redirection problems
4952 under Win32.
4953 [Steve Henson]
4954
4955 *) Initial support for certificate extension requests, these are included
4956 in things like Xenroll certificate requests. Included functions to allow
4957 extensions to be obtained and added.
4958 [Steve Henson]
4959
4960 *) -crlf option to s_client and s_server for sending newlines as
4961 CRLF (as required by many protocols).
4962 [Bodo Moeller]
4963
4964 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4965
4966 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4967 [Ralf S. Engelschall]
4968
4969 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4970 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4971
4972 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4973 program.
4974 [Steve Henson]
4975
4976 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4977 DH parameters/keys (q is lost during that conversion, but the resulting
4978 DH parameters contain its length).
4979
4980 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4981 much faster than DH_generate_parameters (which creates parameters
4982 where p = 2*q + 1), and also the smaller q makes DH computations
4983 much more efficient (160-bit exponentiation instead of 1024-bit
4984 exponentiation); so this provides a convenient way to support DHE
4985 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4986 utter importance to use
4987 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4988 or
4989 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4990 when such DH parameters are used, because otherwise small subgroup
4991 attacks may become possible!
4992 [Bodo Moeller]
4993
4994 *) Avoid memory leak in i2d_DHparams.
4995 [Bodo Moeller]
4996
4997 *) Allow the -k option to be used more than once in the enc program:
4998 this allows the same encrypted message to be read by multiple recipients.
4999 [Steve Henson]
5000
5001 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5002 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5003 it will always use the numerical form of the OID, even if it has a short
5004 or long name.
5005 [Steve Henson]
5006
5007 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5008 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5009 otherwise bn_mod_exp was called. In the case of hardware keys for example
5010 no private key components need be present and it might store extra data
5011 in the RSA structure, which cannot be accessed from bn_mod_exp.
5012 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5013 private key operations.
5014 [Steve Henson]
5015
5016 *) Added support for SPARC Linux.
5017 [Andy Polyakov]
5018
5019 *) pem_password_cb function type incompatibly changed from
5020 typedef int pem_password_cb(char *buf, int size, int rwflag);
5021 to
5022 ....(char *buf, int size, int rwflag, void *userdata);
5023 so that applications can pass data to their callbacks:
5024 The PEM[_ASN1]_{read,write}... functions and macros now take an
5025 additional void * argument, which is just handed through whenever
5026 the password callback is called.
5027 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5028
5029 New function SSL_CTX_set_default_passwd_cb_userdata.
5030
5031 Compatibility note: As many C implementations push function arguments
5032 onto the stack in reverse order, the new library version is likely to
5033 interoperate with programs that have been compiled with the old
5034 pem_password_cb definition (PEM_whatever takes some data that
5035 happens to be on the stack as its last argument, and the callback
5036 just ignores this garbage); but there is no guarantee whatsoever that
5037 this will work.
5038
5039 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5040 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5041 problems not only on Windows, but also on some Unix platforms.
5042 To avoid problematic command lines, these definitions are now in an
5043 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5044 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5045 [Bodo Moeller]
5046
5047 *) MIPS III/IV assembler module is reimplemented.
5048 [Andy Polyakov]
5049
5050 *) More DES library cleanups: remove references to srand/rand and
5051 delete an unused file.
5052 [Ulf Möller]
5053
5054 *) Add support for the the free Netwide assembler (NASM) under Win32,
5055 since not many people have MASM (ml) and it can be hard to obtain.
5056 This is currently experimental but it seems to work OK and pass all
5057 the tests. Check out INSTALL.W32 for info.
5058 [Steve Henson]
5059
5060 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5061 without temporary keys kept an extra copy of the server key,
5062 and connections with temporary keys did not free everything in case
5063 of an error.
5064 [Bodo Moeller]
5065
5066 *) New function RSA_check_key and new openssl rsa option -check
5067 for verifying the consistency of RSA keys.
5068 [Ulf Moeller, Bodo Moeller]
5069
5070 *) Various changes to make Win32 compile work:
5071 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5072 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5073 comparison" warnings.
5074 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5075 [Steve Henson]
5076
5077 *) Add a debugging option to PKCS#5 v2 key generation function: when
5078 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5079 derived keys are printed to stderr.
5080 [Steve Henson]
5081
5082 *) Copy the flags in ASN1_STRING_dup().
5083 [Roman E. Pavlov <pre@mo.msk.ru>]
5084
5085 *) The x509 application mishandled signing requests containing DSA
5086 keys when the signing key was also DSA and the parameters didn't match.
5087
5088 It was supposed to omit the parameters when they matched the signing key:
5089 the verifying software was then supposed to automatically use the CA's
5090 parameters if they were absent from the end user certificate.
5091
5092 Omitting parameters is no longer recommended. The test was also
5093 the wrong way round! This was probably due to unusual behaviour in
5094 EVP_cmp_parameters() which returns 1 if the parameters match.
5095 This meant that parameters were omitted when they *didn't* match and
5096 the certificate was useless. Certificates signed with 'ca' didn't have
5097 this bug.
5098 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5099
5100 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5101 The interface is as follows:
5102 Applications can use
5103 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5104 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5105 "off" is now the default.
5106 The library internally uses
5107 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5108 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5109 to disable memory-checking temporarily.
5110
5111 Some inconsistent states that previously were possible (and were
5112 even the default) are now avoided.
5113
5114 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5115 with each memory chunk allocated; this is occasionally more helpful
5116 than just having a counter.
5117
5118 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5119
5120 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5121 extensions.
5122 [Bodo Moeller]
5123
5124 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5125 which largely parallels "options", but is for changing API behaviour,
5126 whereas "options" are about protocol behaviour.
5127 Initial "mode" flags are:
5128
5129 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5130 a single record has been written.
5131 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5132 retries use the same buffer location.
5133 (But all of the contents must be
5134 copied!)
5135 [Bodo Moeller]
5136
5137 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5138 worked.
5139
5140 *) Fix problems with no-hmac etc.
5141 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5142
5143 *) New functions RSA_get_default_method(), RSA_set_method() and
5144 RSA_get_method(). These allows replacement of RSA_METHODs without having
5145 to mess around with the internals of an RSA structure.
5146 [Steve Henson]
5147
5148 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5149 Also really enable memory leak checks in openssl.c and in some
5150 test programs.
5151 [Chad C. Mulligan, Bodo Moeller]
5152
5153 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5154 up the length of negative integers. This has now been simplified to just
5155 store the length when it is first determined and use it later, rather
5156 than trying to keep track of where data is copied and updating it to
5157 point to the end.
5158 [Steve Henson, reported by Brien Wheeler
5159 <bwheeler@authentica-security.com>]
5160
5161 *) Add a new function PKCS7_signatureVerify. This allows the verification
5162 of a PKCS#7 signature but with the signing certificate passed to the
5163 function itself. This contrasts with PKCS7_dataVerify which assumes the
5164 certificate is present in the PKCS#7 structure. This isn't always the
5165 case: certificates can be omitted from a PKCS#7 structure and be
5166 distributed by "out of band" means (such as a certificate database).
5167 [Steve Henson]
5168
5169 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5170 function prototypes in pem.h, also change util/mkdef.pl to add the
5171 necessary function names.
5172 [Steve Henson]
5173
5174 *) mk1mf.pl (used by Windows builds) did not properly read the
5175 options set by Configure in the top level Makefile, and Configure
5176 was not even able to write more than one option correctly.
5177 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5178 [Bodo Moeller]
5179
5180 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5181 file to be loaded from a BIO or FILE pointer. The BIO version will
5182 for example allow memory BIOs to contain config info.
5183 [Steve Henson]
5184
5185 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5186 Whoever hopes to achieve shared-library compatibility across versions
5187 must use this, not the compile-time macro.
5188 (Exercise 0.9.4: Which is the minimum library version required by
5189 such programs?)
5190 Note: All this applies only to multi-threaded programs, others don't
5191 need locks.
5192 [Bodo Moeller]
5193
5194 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5195 through a BIO pair triggered the default case, i.e.
5196 SSLerr(...,SSL_R_UNKNOWN_STATE).
5197 [Bodo Moeller]
5198
5199 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5200 can use the SSL library even if none of the specific BIOs is
5201 appropriate.
5202 [Bodo Moeller]
5203
5204 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5205 for the encoded length.
5206 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5207
5208 *) Add initial documentation of the X509V3 functions.
5209 [Steve Henson]
5210
5211 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5212 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5213 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5214 secure PKCS#8 private key format with a high iteration count.
5215 [Steve Henson]
5216
5217 *) Fix determination of Perl interpreter: A perl or perl5
5218 _directory_ in $PATH was also accepted as the interpreter.
5219 [Ralf S. Engelschall]
5220
5221 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5222 wrong with it but it was very old and did things like calling
5223 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5224 unusual formatting.
5225 [Steve Henson]
5226
5227 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5228 to use the new extension code.
5229 [Steve Henson]
5230
5231 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5232 with macros. This should make it easier to change their form, add extra
5233 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5234 constant.
5235 [Steve Henson]
5236
5237 *) Add to configuration table a new entry that can specify an alternative
5238 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5239 according to Mark Crispin <MRC@Panda.COM>.
5240 [Bodo Moeller]
5241
5242 #if 0
5243 *) DES CBC did not update the IV. Weird.
5244 [Ben Laurie]
5245 #else
5246 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5247 Changing the behaviour of the former might break existing programs --
5248 where IV updating is needed, des_ncbc_encrypt can be used.
5249 #endif
5250
5251 *) When bntest is run from "make test" it drives bc to check its
5252 calculations, as well as internally checking them. If an internal check
5253 fails, it needs to cause bc to give a non-zero result or make test carries
5254 on without noticing the failure. Fixed.
5255 [Ben Laurie]
5256
5257 *) DES library cleanups.
5258 [Ulf Möller]
5259
5260 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5261 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5262 ciphers. NOTE: although the key derivation function has been verified
5263 against some published test vectors it has not been extensively tested
5264 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5265 of v2.0.
5266 [Steve Henson]
5267
5268 *) Instead of "mkdir -p", which is not fully portable, use new
5269 Perl script "util/mkdir-p.pl".
5270 [Bodo Moeller]
5271
5272 *) Rewrite the way password based encryption (PBE) is handled. It used to
5273 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5274 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5275 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5276 the 'parameter' field of the AlgorithmIdentifier is passed to the
5277 underlying key generation function so it must do its own ASN1 parsing.
5278 This has also changed the EVP_PBE_CipherInit() function which now has a
5279 'parameter' argument instead of literal salt and iteration count values
5280 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5281 [Steve Henson]
5282
5283 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5284 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5285 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5286 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5287 value was just used as a "magic string" and not used directly its
5288 value doesn't matter.
5289 [Steve Henson]
5290
5291 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5292 support mutable.
5293 [Ben Laurie]
5294
5295 *) "linux-sparc64" configuration (ultrapenguin).
5296 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5297 "linux-sparc" configuration.
5298 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5299
5300 *) config now generates no-xxx options for missing ciphers.
5301 [Ulf Möller]
5302
5303 *) Support the EBCDIC character set (work in progress).
5304 File ebcdic.c not yet included because it has a different license.
5305 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5306
5307 *) Support BS2000/OSD-POSIX.
5308 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5309
5310 *) Make callbacks for key generation use void * instead of char *.
5311 [Ben Laurie]
5312
5313 *) Make S/MIME samples compile (not yet tested).
5314 [Ben Laurie]
5315
5316 *) Additional typesafe stacks.
5317 [Ben Laurie]
5318
5319 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5320 [Bodo Moeller]
5321
5322
5323 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5324
5325 *) New configuration variant "sco5-gcc".
5326
5327 *) Updated some demos.
5328 [Sean O Riordain, Wade Scholine]
5329
5330 *) Add missing BIO_free at exit of pkcs12 application.
5331 [Wu Zhigang]
5332
5333 *) Fix memory leak in conf.c.
5334 [Steve Henson]
5335
5336 *) Updates for Win32 to assembler version of MD5.
5337 [Steve Henson]
5338
5339 *) Set #! path to perl in apps/der_chop to where we found it
5340 instead of using a fixed path.
5341 [Bodo Moeller]
5342
5343 *) SHA library changes for irix64-mips4-cc.
5344 [Andy Polyakov]
5345
5346 *) Improvements for VMS support.
5347 [Richard Levitte]
5348
5349
5350 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5351
5352 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5353 This also avoids the problems with SC4.2 and unpatched SC5.
5354 [Andy Polyakov <appro@fy.chalmers.se>]
5355
5356 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5357 These are required because of the typesafe stack would otherwise break
5358 existing code. If old code used a structure member which used to be STACK
5359 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5360 sk_num or sk_value it would produce an error because the num, data members
5361 are not present in STACK_OF. Now it just produces a warning. sk_set
5362 replaces the old method of assigning a value to sk_value
5363 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5364 that does this will no longer work (and should use sk_set instead) but
5365 this could be regarded as a "questionable" behaviour anyway.
5366 [Steve Henson]
5367
5368 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5369 correctly handle encrypted S/MIME data.
5370 [Steve Henson]
5371
5372 *) Change type of various DES function arguments from des_cblock
5373 (which means, in function argument declarations, pointer to char)
5374 to des_cblock * (meaning pointer to array with 8 char elements),
5375 which allows the compiler to do more typechecking; it was like
5376 that back in SSLeay, but with lots of ugly casts.
5377
5378 Introduce new type const_des_cblock.
5379 [Bodo Moeller]
5380
5381 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5382 problems: find RecipientInfo structure that matches recipient certificate
5383 and initialise the ASN1 structures properly based on passed cipher.
5384 [Steve Henson]
5385
5386 *) Belatedly make the BN tests actually check the results.
5387 [Ben Laurie]
5388
5389 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5390 to and from BNs: it was completely broken. New compilation option
5391 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5392 key elements as negative integers.
5393 [Steve Henson]
5394
5395 *) Reorganize and speed up MD5.
5396 [Andy Polyakov <appro@fy.chalmers.se>]
5397
5398 *) VMS support.
5399 [Richard Levitte <richard@levitte.org>]
5400
5401 *) New option -out to asn1parse to allow the parsed structure to be
5402 output to a file. This is most useful when combined with the -strparse
5403 option to examine the output of things like OCTET STRINGS.
5404 [Steve Henson]
5405
5406 *) Make SSL library a little more fool-proof by not requiring any longer
5407 that SSL_set_{accept,connect}_state be called before
5408 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5409 in many applications because usually everything *appeared* to work as
5410 intended anyway -- now it really works as intended).
5411 [Bodo Moeller]
5412
5413 *) Move openssl.cnf out of lib/.
5414 [Ulf Möller]
5415
5416 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5417 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5418 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5419 [Ralf S. Engelschall]
5420
5421 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5422 handle PKCS#7 enveloped data properly.
5423 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5424
5425 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5426 copying pointers. The cert_st handling is changed by this in
5427 various ways (and thus what used to be known as ctx->default_cert
5428 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5429 any longer when s->cert does not give us what we need).
5430 ssl_cert_instantiate becomes obsolete by this change.
5431 As soon as we've got the new code right (possibly it already is?),
5432 we have solved a couple of bugs of the earlier code where s->cert
5433 was used as if it could not have been shared with other SSL structures.
5434
5435 Note that using the SSL API in certain dirty ways now will result
5436 in different behaviour than observed with earlier library versions:
5437 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5438 does not influence s as it used to.
5439
5440 In order to clean up things more thoroughly, inside SSL_SESSION
5441 we don't use CERT any longer, but a new structure SESS_CERT
5442 that holds per-session data (if available); currently, this is
5443 the peer's certificate chain and, for clients, the server's certificate
5444 and temporary key. CERT holds only those values that can have
5445 meaningful defaults in an SSL_CTX.
5446 [Bodo Moeller]
5447
5448 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5449 from the internal representation. Various PKCS#7 fixes: remove some
5450 evil casts and set the enc_dig_alg field properly based on the signing
5451 key type.
5452 [Steve Henson]
5453
5454 *) Allow PKCS#12 password to be set from the command line or the
5455 environment. Let 'ca' get its config file name from the environment
5456 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5457 and 'x509').
5458 [Steve Henson]
5459
5460 *) Allow certificate policies extension to use an IA5STRING for the
5461 organization field. This is contrary to the PKIX definition but
5462 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5463 extension option.
5464 [Steve Henson]
5465
5466 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5467 without disallowing inline assembler and the like for non-pedantic builds.
5468 [Ben Laurie]
5469
5470 *) Support Borland C++ builder.
5471 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5472
5473 *) Support Mingw32.
5474 [Ulf Möller]
5475
5476 *) SHA-1 cleanups and performance enhancements.
5477 [Andy Polyakov <appro@fy.chalmers.se>]
5478
5479 *) Sparc v8plus assembler for the bignum library.
5480 [Andy Polyakov <appro@fy.chalmers.se>]
5481
5482 *) Accept any -xxx and +xxx compiler options in Configure.
5483 [Ulf Möller]
5484
5485 *) Update HPUX configuration.
5486 [Anonymous]
5487
5488 *) Add missing sk_<type>_unshift() function to safestack.h
5489 [Ralf S. Engelschall]
5490
5491 *) New function SSL_CTX_use_certificate_chain_file that sets the
5492 "extra_cert"s in addition to the certificate. (This makes sense
5493 only for "PEM" format files, as chains as a whole are not
5494 DER-encoded.)
5495 [Bodo Moeller]
5496
5497 *) Support verify_depth from the SSL API.
5498 x509_vfy.c had what can be considered an off-by-one-error:
5499 Its depth (which was not part of the external interface)
5500 was actually counting the number of certificates in a chain;
5501 now it really counts the depth.
5502 [Bodo Moeller]
5503
5504 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5505 instead of X509err, which often resulted in confusing error
5506 messages since the error codes are not globally unique
5507 (e.g. an alleged error in ssl3_accept when a certificate
5508 didn't match the private key).
5509
5510 *) New function SSL_CTX_set_session_id_context that allows to set a default
5511 value (so that you don't need SSL_set_session_id_context for each
5512 connection using the SSL_CTX).
5513 [Bodo Moeller]
5514
5515 *) OAEP decoding bug fix.
5516 [Ulf Möller]
5517
5518 *) Support INSTALL_PREFIX for package builders, as proposed by
5519 David Harris.
5520 [Bodo Moeller]
5521
5522 *) New Configure options "threads" and "no-threads". For systems
5523 where the proper compiler options are known (currently Solaris
5524 and Linux), "threads" is the default.
5525 [Bodo Moeller]
5526
5527 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5528 [Bodo Moeller]
5529
5530 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5531 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5532 such as /usr/local/bin.
5533 [Bodo Moeller]
5534
5535 *) "make linux-shared" to build shared libraries.
5536 [Niels Poppe <niels@netbox.org>]
5537
5538 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5539 [Ulf Möller]
5540
5541 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5542 extension adding in x509 utility.
5543 [Steve Henson]
5544
5545 *) Remove NOPROTO sections and error code comments.
5546 [Ulf Möller]
5547
5548 *) Partial rewrite of the DEF file generator to now parse the ANSI
5549 prototypes.
5550 [Steve Henson]
5551
5552 *) New Configure options --prefix=DIR and --openssldir=DIR.
5553 [Ulf Möller]
5554
5555 *) Complete rewrite of the error code script(s). It is all now handled
5556 by one script at the top level which handles error code gathering,
5557 header rewriting and C source file generation. It should be much better
5558 than the old method: it now uses a modified version of Ulf's parser to
5559 read the ANSI prototypes in all header files (thus the old K&R definitions
5560 aren't needed for error creation any more) and do a better job of
5561 translating function codes into names. The old 'ASN1 error code imbedded
5562 in a comment' is no longer necessary and it doesn't use .err files which
5563 have now been deleted. Also the error code call doesn't have to appear all
5564 on one line (which resulted in some large lines...).
5565 [Steve Henson]
5566
5567 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5568 [Bodo Moeller]
5569
5570 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5571 0 (which usually indicates a closed connection), but continue reading.
5572 [Bodo Moeller]
5573
5574 *) Fix some race conditions.
5575 [Bodo Moeller]
5576
5577 *) Add support for CRL distribution points extension. Add Certificate
5578 Policies and CRL distribution points documentation.
5579 [Steve Henson]
5580
5581 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5582 [Ulf Möller]
5583
5584 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5585 8 of keying material. Merlin has also confirmed interop with this fix
5586 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5587 [Merlin Hughes <merlin@baltimore.ie>]
5588
5589 *) Fix lots of warnings.
5590 [Richard Levitte <levitte@stacken.kth.se>]
5591
5592 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5593 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5594 [Richard Levitte <levitte@stacken.kth.se>]
5595
5596 *) Fix problems with sizeof(long) == 8.
5597 [Andy Polyakov <appro@fy.chalmers.se>]
5598
5599 *) Change functions to ANSI C.
5600 [Ulf Möller]
5601
5602 *) Fix typos in error codes.
5603 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5604
5605 *) Remove defunct assembler files from Configure.
5606 [Ulf Möller]
5607
5608 *) SPARC v8 assembler BIGNUM implementation.
5609 [Andy Polyakov <appro@fy.chalmers.se>]
5610
5611 *) Support for Certificate Policies extension: both print and set.
5612 Various additions to support the r2i method this uses.
5613 [Steve Henson]
5614
5615 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5616 return a const string when you are expecting an allocated buffer.
5617 [Ben Laurie]
5618
5619 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5620 types DirectoryString and DisplayText.
5621 [Steve Henson]
5622
5623 *) Add code to allow r2i extensions to access the configuration database,
5624 add an LHASH database driver and add several ctx helper functions.
5625 [Steve Henson]
5626
5627 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5628 fail when they extended the size of a BIGNUM.
5629 [Steve Henson]
5630
5631 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5632 support typesafe stack.
5633 [Steve Henson]
5634
5635 *) Fix typo in SSL_[gs]et_options().
5636 [Nils Frostberg <nils@medcom.se>]
5637
5638 *) Delete various functions and files that belonged to the (now obsolete)
5639 old X509V3 handling code.
5640 [Steve Henson]
5641
5642 *) New Configure option "rsaref".
5643 [Ulf Möller]
5644
5645 *) Don't auto-generate pem.h.
5646 [Bodo Moeller]
5647
5648 *) Introduce type-safe ASN.1 SETs.
5649 [Ben Laurie]
5650
5651 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5652 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5653
5654 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5655 that links with OpenSSL (well at least cause lots of warnings), but fear
5656 not: the conversion is trivial, and it eliminates loads of evil casts. A
5657 few STACKed things have been converted already. Feel free to convert more.
5658 In the fullness of time, I'll do away with the STACK type altogether.
5659 [Ben Laurie]
5660
5661 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5662 specified in <certfile> by updating the entry in the index.txt file.
5663 This way one no longer has to edit the index.txt file manually for
5664 revoking a certificate. The -revoke option does the gory details now.
5665 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5666
5667 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5668 `-text' option at all and this way the `-noout -text' combination was
5669 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5670 [Ralf S. Engelschall]
5671
5672 *) Make sure a corresponding plain text error message exists for the
5673 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5674 verify callback function determined that a certificate was revoked.
5675 [Ralf S. Engelschall]
5676
5677 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5678 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5679 all available cipers including rc5, which was forgotten until now.
5680 In order to let the testing shell script know which algorithms
5681 are available, a new (up to now undocumented) command
5682 "openssl list-cipher-commands" is used.
5683 [Bodo Moeller]
5684
5685 *) Bugfix: s_client occasionally would sleep in select() when
5686 it should have checked SSL_pending() first.
5687 [Bodo Moeller]
5688
5689 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5690 the raw DSA values prior to ASN.1 encoding.
5691 [Ulf Möller]
5692
5693 *) Tweaks to Configure
5694 [Niels Poppe <niels@netbox.org>]
5695
5696 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5697 yet...
5698 [Steve Henson]
5699
5700 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5701 [Ulf Möller]
5702
5703 *) New config option to avoid instructions that are illegal on the 80386.
5704 The default code is faster, but requires at least a 486.
5705 [Ulf Möller]
5706
5707 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5708 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5709 same as SSL2_VERSION anyway.
5710 [Bodo Moeller]
5711
5712 *) New "-showcerts" option for s_client.
5713 [Bodo Moeller]
5714
5715 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5716 application. Various cleanups and fixes.
5717 [Steve Henson]
5718
5719 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5720 modify error routines to work internally. Add error codes and PBE init
5721 to library startup routines.
5722 [Steve Henson]
5723
5724 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5725 packing functions to asn1 and evp. Changed function names and error
5726 codes along the way.
5727 [Steve Henson]
5728
5729 *) PKCS12 integration: and so it begins... First of several patches to
5730 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5731 objects to objects.h
5732 [Steve Henson]
5733
5734 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5735 and display support for Thawte strong extranet extension.
5736 [Steve Henson]
5737
5738 *) Add LinuxPPC support.
5739 [Jeff Dubrule <igor@pobox.org>]
5740
5741 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5742 bn_div_words in alpha.s.
5743 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5744
5745 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5746 OAEP isn't supported when OpenSSL is built with RSAref.
5747 [Ulf Moeller <ulf@fitug.de>]
5748
5749 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5750 so they no longer are missing under -DNOPROTO.
5751 [Soren S. Jorvang <soren@t.dk>]
5752
5753
5754 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5755
5756 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5757 doesn't work when the session is reused. Coming soon!
5758 [Ben Laurie]
5759
5760 *) Fix a security hole, that allows sessions to be reused in the wrong
5761 context thus bypassing client cert protection! All software that uses
5762 client certs and session caches in multiple contexts NEEDS PATCHING to
5763 allow session reuse! A fuller solution is in the works.
5764 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5765
5766 *) Some more source tree cleanups (removed obsolete files
5767 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5768 permission on "config" script to be executable) and a fix for the INSTALL
5769 document.
5770 [Ulf Moeller <ulf@fitug.de>]
5771
5772 *) Remove some legacy and erroneous uses of malloc, free instead of
5773 Malloc, Free.
5774 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5775
5776 *) Make rsa_oaep_test return non-zero on error.
5777 [Ulf Moeller <ulf@fitug.de>]
5778
5779 *) Add support for native Solaris shared libraries. Configure
5780 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5781 if someone would make that last step automatic.
5782 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5783
5784 *) ctx_size was not built with the right compiler during "make links". Fixed.
5785 [Ben Laurie]
5786
5787 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5788 except NULL ciphers". This means the default cipher list will no longer
5789 enable NULL ciphers. They need to be specifically enabled e.g. with
5790 the string "DEFAULT:eNULL".
5791 [Steve Henson]
5792
5793 *) Fix to RSA private encryption routines: if p < q then it would
5794 occasionally produce an invalid result. This will only happen with
5795 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5796 [Steve Henson]
5797
5798 *) Be less restrictive and allow also `perl util/perlpath.pl
5799 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5800 because this way one can also use an interpreter named `perl5' (which is
5801 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5802 installed as `perl').
5803 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5804
5805 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5806 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5807
5808 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5809 advapi32.lib to Win32 build and change the pem test comparision
5810 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5811 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5812 and crypto/des/ede_cbcm_enc.c.
5813 [Steve Henson]
5814
5815 *) DES quad checksum was broken on big-endian architectures. Fixed.
5816 [Ben Laurie]
5817
5818 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5819 Win32 test batch file so it (might) work again. The Win32 test batch file
5820 is horrible: I feel ill....
5821 [Steve Henson]
5822
5823 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5824 in e_os.h. Audit of header files to check ANSI and non ANSI
5825 sections: 10 functions were absent from non ANSI section and not exported
5826 from Windows DLLs. Fixed up libeay.num for new functions.
5827 [Steve Henson]
5828
5829 *) Make `openssl version' output lines consistent.
5830 [Ralf S. Engelschall]
5831
5832 *) Fix Win32 symbol export lists for BIO functions: Added
5833 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5834 to ms/libeay{16,32}.def.
5835 [Ralf S. Engelschall]
5836
5837 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5838 fine under Unix and passes some trivial tests I've now added. But the
5839 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5840 added to make sure no one expects that this stuff really works in the
5841 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5842 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5843 openssl_bio.xs.
5844 [Ralf S. Engelschall]
5845
5846 *) Fix the generation of two part addresses in perl.
5847 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5848
5849 *) Add config entry for Linux on MIPS.
5850 [John Tobey <jtobey@channel1.com>]
5851
5852 *) Make links whenever Configure is run, unless we are on Windoze.
5853 [Ben Laurie]
5854
5855 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5856 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5857 in CRLs.
5858 [Steve Henson]
5859
5860 *) Add a useful kludge to allow package maintainers to specify compiler and
5861 other platforms details on the command line without having to patch the
5862 Configure script everytime: One now can use ``perl Configure
5863 <id>:<details>'', i.e. platform ids are allowed to have details appended
5864 to them (seperated by colons). This is treated as there would be a static
5865 pre-configured entry in Configure's %table under key <id> with value
5866 <details> and ``perl Configure <id>'' is called. So, when you want to
5867 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5868 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5869 now, which overrides the FreeBSD-elf entry on-the-fly.
5870 [Ralf S. Engelschall]
5871
5872 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5873 [Ben Laurie]
5874
5875 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5876 on the `perl Configure ...' command line. This way one can compile
5877 OpenSSL libraries with Position Independent Code (PIC) which is needed
5878 for linking it into DSOs.
5879 [Ralf S. Engelschall]
5880
5881 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5882 Fixed.
5883 [Ben Laurie]
5884
5885 *) Cleaned up the LICENSE document: The official contact for any license
5886 questions now is the OpenSSL core team under openssl-core@openssl.org.
5887 And add a paragraph about the dual-license situation to make sure people
5888 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5889 to the OpenSSL toolkit.
5890 [Ralf S. Engelschall]
5891
5892 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5893 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5894 Additonally cleaned up the `make links' target: Remove unnecessary
5895 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5896 to speed processing and no longer clutter the display with confusing
5897 stuff. Instead only the actually done links are displayed.
5898 [Ralf S. Engelschall]
5899
5900 *) Permit null encryption ciphersuites, used for authentication only. It used
5901 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5902 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5903 encryption.
5904 [Ben Laurie]
5905
5906 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5907 signed attributes when verifying signatures (this would break them),
5908 the detached data encoding was wrong and public keys obtained using
5909 X509_get_pubkey() weren't freed.
5910 [Steve Henson]
5911
5912 *) Add text documentation for the BUFFER functions. Also added a work around
5913 to a Win95 console bug. This was triggered by the password read stuff: the
5914 last character typed gets carried over to the next fread(). If you were
5915 generating a new cert request using 'req' for example then the last
5916 character of the passphrase would be CR which would then enter the first
5917 field as blank.
5918 [Steve Henson]
5919
5920 *) Added the new `Includes OpenSSL Cryptography Software' button as
5921 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5922 button and can be used by applications based on OpenSSL to show the
5923 relationship to the OpenSSL project.
5924 [Ralf S. Engelschall]
5925
5926 *) Remove confusing variables in function signatures in files
5927 ssl/ssl_lib.c and ssl/ssl.h.
5928 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5929
5930 *) Don't install bss_file.c under PREFIX/include/
5931 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5932
5933 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5934 functions that return function pointers and has support for NT specific
5935 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5936 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5937 unsigned to signed types: this was killing the Win32 compile.
5938 [Steve Henson]
5939
5940 *) Add new certificate file to stack functions,
5941 SSL_add_dir_cert_subjects_to_stack() and
5942 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5943 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5944 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5945 This means that Apache-SSL and similar packages don't have to mess around
5946 to add as many CAs as they want to the preferred list.
5947 [Ben Laurie]
5948
5949 *) Experiment with doxygen documentation. Currently only partially applied to
5950 ssl/ssl_lib.c.
5951 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5952 openssl.doxy as the configuration file.
5953 [Ben Laurie]
5954
5955 *) Get rid of remaining C++-style comments which strict C compilers hate.
5956 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5957
5958 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5959 compiled in by default: it has problems with large keys.
5960 [Steve Henson]
5961
5962 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5963 DH private keys and/or callback functions which directly correspond to
5964 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5965 is needed for applications which have to configure certificates on a
5966 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5967 (e.g. s_server).
5968 For the RSA certificate situation is makes no difference, but
5969 for the DSA certificate situation this fixes the "no shared cipher"
5970 problem where the OpenSSL cipher selection procedure failed because the
5971 temporary keys were not overtaken from the context and the API provided
5972 no way to reconfigure them.
5973 The new functions now let applications reconfigure the stuff and they
5974 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5975 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5976 non-public-API function ssl_cert_instantiate() is used as a helper
5977 function and also to reduce code redundancy inside ssl_rsa.c.
5978 [Ralf S. Engelschall]
5979
5980 *) Move s_server -dcert and -dkey options out of the undocumented feature
5981 area because they are useful for the DSA situation and should be
5982 recognized by the users.
5983 [Ralf S. Engelschall]
5984
5985 *) Fix the cipher decision scheme for export ciphers: the export bits are
5986 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5987 SSL_EXP_MASK. So, the original variable has to be used instead of the
5988 already masked variable.
5989 [Richard Levitte <levitte@stacken.kth.se>]
5990
5991 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5992 [Richard Levitte <levitte@stacken.kth.se>]
5993
5994 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5995 from `int' to `unsigned int' because it's a length and initialized by
5996 EVP_DigestFinal() which expects an `unsigned int *'.
5997 [Richard Levitte <levitte@stacken.kth.se>]
5998
5999 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6000 script. Instead use the usual Shell->Perl transition trick.
6001 [Ralf S. Engelschall]
6002
6003 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6004 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6005 -noout -modulus' as it's already the case for `openssl rsa -noout
6006 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6007 currently the public key is printed (a decision which was already done by
6008 `openssl dsa -modulus' in the past) which serves a similar purpose.
6009 Additionally the NO_RSA no longer completely removes the whole -modulus
6010 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6011 now, too.
6012 [Ralf S. Engelschall]
6013
6014 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6015 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6016 [Arne Ansper <arne@ats.cyber.ee>]
6017
6018 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6019 to be added. Now both 'req' and 'ca' can use new objects defined in the
6020 config file.
6021 [Steve Henson]
6022
6023 *) Add cool BIO that does syslog (or event log on NT).
6024 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6025
6026 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6027 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6028 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6029 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6030 [Ben Laurie]
6031
6032 *) Add preliminary config info for new extension code.
6033 [Steve Henson]
6034
6035 *) Make RSA_NO_PADDING really use no padding.
6036 [Ulf Moeller <ulf@fitug.de>]
6037
6038 *) Generate errors when private/public key check is done.
6039 [Ben Laurie]
6040
6041 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6042 for some CRL extensions and new objects added.
6043 [Steve Henson]
6044
6045 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6046 key usage extension and fuller support for authority key id.
6047 [Steve Henson]
6048
6049 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6050 padding method for RSA, which is recommended for new applications in PKCS
6051 #1 v2.0 (RFC 2437, October 1998).
6052 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6053 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6054 against Bleichbacher's attack on RSA.
6055 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6056 Ben Laurie]
6057
6058 *) Updates to the new SSL compression code
6059 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6060
6061 *) Fix so that the version number in the master secret, when passed
6062 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6063 (because the server will not accept higher), that the version number
6064 is 0x03,0x01, not 0x03,0x00
6065 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6066
6067 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6068 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6069 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6070 [Steve Henson]
6071
6072 *) Support for RAW extensions where an arbitrary extension can be
6073 created by including its DER encoding. See apps/openssl.cnf for
6074 an example.
6075 [Steve Henson]
6076
6077 *) Make sure latest Perl versions don't interpret some generated C array
6078 code as Perl array code in the crypto/err/err_genc.pl script.
6079 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6080
6081 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6082 not many people have the assembler. Various Win32 compilation fixes and
6083 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6084 build instructions.
6085 [Steve Henson]
6086
6087 *) Modify configure script 'Configure' to automatically create crypto/date.h
6088 file under Win32 and also build pem.h from pem.org. New script
6089 util/mkfiles.pl to create the MINFO file on environments that can't do a
6090 'make files': perl util/mkfiles.pl >MINFO should work.
6091 [Steve Henson]
6092
6093 *) Major rework of DES function declarations, in the pursuit of correctness
6094 and purity. As a result, many evil casts evaporated, and some weirdness,
6095 too. You may find this causes warnings in your code. Zapping your evil
6096 casts will probably fix them. Mostly.
6097 [Ben Laurie]
6098
6099 *) Fix for a typo in asn1.h. Bug fix to object creation script
6100 obj_dat.pl. It considered a zero in an object definition to mean
6101 "end of object": none of the objects in objects.h have any zeros
6102 so it wasn't spotted.
6103 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6104
6105 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6106 Masking (CBCM). In the absence of test vectors, the best I have been able
6107 to do is check that the decrypt undoes the encrypt, so far. Send me test
6108 vectors if you have them.
6109 [Ben Laurie]
6110
6111 *) Correct calculation of key length for export ciphers (too much space was
6112 allocated for null ciphers). This has not been tested!
6113 [Ben Laurie]
6114
6115 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6116 message is now correct (it understands "crypto" and "ssl" on its
6117 command line). There is also now an "update" option. This will update
6118 the util/ssleay.num and util/libeay.num files with any new functions.
6119 If you do a:
6120 perl util/mkdef.pl crypto ssl update
6121 it will update them.
6122 [Steve Henson]
6123
6124 *) Overhauled the Perl interface (perl/*):
6125 - ported BN stuff to OpenSSL's different BN library
6126 - made the perl/ source tree CVS-aware
6127 - renamed the package from SSLeay to OpenSSL (the files still contain
6128 their history because I've copied them in the repository)
6129 - removed obsolete files (the test scripts will be replaced
6130 by better Test::Harness variants in the future)
6131 [Ralf S. Engelschall]
6132
6133 *) First cut for a very conservative source tree cleanup:
6134 1. merge various obsolete readme texts into doc/ssleay.txt
6135 where we collect the old documents and readme texts.
6136 2. remove the first part of files where I'm already sure that we no
6137 longer need them because of three reasons: either they are just temporary
6138 files which were left by Eric or they are preserved original files where
6139 I've verified that the diff is also available in the CVS via "cvs diff
6140 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6141 the crypto/md/ stuff).
6142 [Ralf S. Engelschall]
6143
6144 *) More extension code. Incomplete support for subject and issuer alt
6145 name, issuer and authority key id. Change the i2v function parameters
6146 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6147 what that's for :-) Fix to ASN1 macro which messed up
6148 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6149 [Steve Henson]
6150
6151 *) Preliminary support for ENUMERATED type. This is largely copied from the
6152 INTEGER code.
6153 [Steve Henson]
6154
6155 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6156 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6157
6158 *) Make sure `make rehash' target really finds the `openssl' program.
6159 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6160
6161 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6162 like to hear about it if this slows down other processors.
6163 [Ben Laurie]
6164
6165 *) Add CygWin32 platform information to Configure script.
6166 [Alan Batie <batie@aahz.jf.intel.com>]
6167
6168 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6169 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6170
6171 *) New program nseq to manipulate netscape certificate sequences
6172 [Steve Henson]
6173
6174 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6175 few typos.
6176 [Steve Henson]
6177
6178 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6179 but the BN code had some problems that would cause failures when
6180 doing certificate verification and some other functions.
6181 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6182
6183 *) Add ASN1 and PEM code to support netscape certificate sequences.
6184 [Steve Henson]
6185
6186 *) Add ASN1 and PEM code to support netscape certificate sequences.
6187 [Steve Henson]
6188
6189 *) Add several PKIX and private extended key usage OIDs.
6190 [Steve Henson]
6191
6192 *) Modify the 'ca' program to handle the new extension code. Modify
6193 openssl.cnf for new extension format, add comments.
6194 [Steve Henson]
6195
6196 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6197 and add a sample to openssl.cnf so req -x509 now adds appropriate
6198 CA extensions.
6199 [Steve Henson]
6200
6201 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6202 error code, add initial support to X509_print() and x509 application.
6203 [Steve Henson]
6204
6205 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6206 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6207 stuff is currently isolated and isn't even compiled yet.
6208 [Steve Henson]
6209
6210 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6211 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6212 Removed the versions check from X509 routines when loading extensions:
6213 this allows certain broken certificates that don't set the version
6214 properly to be processed.
6215 [Steve Henson]
6216
6217 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6218 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6219 can still be regenerated with "make depend".
6220 [Ben Laurie]
6221
6222 *) Spelling mistake in C version of CAST-128.
6223 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6224
6225 *) Changes to the error generation code. The perl script err-code.pl
6226 now reads in the old error codes and retains the old numbers, only
6227 adding new ones if necessary. It also only changes the .err files if new
6228 codes are added. The makefiles have been modified to only insert errors
6229 when needed (to avoid needlessly modifying header files). This is done
6230 by only inserting errors if the .err file is newer than the auto generated
6231 C file. To rebuild all the error codes from scratch (the old behaviour)
6232 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6233 or delete all the .err files.
6234 [Steve Henson]
6235
6236 *) CAST-128 was incorrectly implemented for short keys. The C version has
6237 been fixed, but is untested. The assembler versions are also fixed, but
6238 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6239 to regenerate it if needed.
6240 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6241 Hagino <itojun@kame.net>]
6242
6243 *) File was opened incorrectly in randfile.c.
6244 [Ulf Möller <ulf@fitug.de>]
6245
6246 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6247 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6248 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6249 al: it's just almost always a UTCTime. Note this patch adds new error
6250 codes so do a "make errors" if there are problems.
6251 [Steve Henson]
6252
6253 *) Correct Linux 1 recognition in config.
6254 [Ulf Möller <ulf@fitug.de>]
6255
6256 *) Remove pointless MD5 hash when using DSA keys in ca.
6257 [Anonymous <nobody@replay.com>]
6258
6259 *) Generate an error if given an empty string as a cert directory. Also
6260 generate an error if handed NULL (previously returned 0 to indicate an
6261 error, but didn't set one).
6262 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6263
6264 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6265 [Ben Laurie]
6266
6267 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6268 parameters. This was causing a warning which killed off the Win32 compile.
6269 [Steve Henson]
6270
6271 *) Remove C++ style comments from crypto/bn/bn_local.h.
6272 [Neil Costigan <neil.costigan@celocom.com>]
6273
6274 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6275 based on a text string, looking up short and long names and finally
6276 "dot" format. The "dot" format stuff didn't work. Added new function
6277 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6278 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6279 OID is not part of the table.
6280 [Steve Henson]
6281
6282 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6283 X509_LOOKUP_by_alias().
6284 [Ben Laurie]
6285
6286 *) Sort openssl functions by name.
6287 [Ben Laurie]
6288
6289 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6290 encryption from sample DSA keys (in case anyone is interested the password
6291 was "1234").
6292 [Steve Henson]
6293
6294 *) Make _all_ *_free functions accept a NULL pointer.
6295 [Frans Heymans <fheymans@isaserver.be>]
6296
6297 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6298 NULL pointers.
6299 [Anonymous <nobody@replay.com>]
6300
6301 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6302 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6303
6304 *) Don't blow it for numeric -newkey arguments to apps/req.
6305 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6306
6307 *) Temp key "for export" tests were wrong in s3_srvr.c.
6308 [Anonymous <nobody@replay.com>]
6309
6310 *) Add prototype for temp key callback functions
6311 SSL_CTX_set_tmp_{rsa,dh}_callback().
6312 [Ben Laurie]
6313
6314 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6315 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6316 [Steve Henson]
6317
6318 *) X509_name_add_entry() freed the wrong thing after an error.
6319 [Arne Ansper <arne@ats.cyber.ee>]
6320
6321 *) rsa_eay.c would attempt to free a NULL context.
6322 [Arne Ansper <arne@ats.cyber.ee>]
6323
6324 *) BIO_s_socket() had a broken should_retry() on Windoze.
6325 [Arne Ansper <arne@ats.cyber.ee>]
6326
6327 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6328 [Arne Ansper <arne@ats.cyber.ee>]
6329
6330 *) Make sure the already existing X509_STORE->depth variable is initialized
6331 in X509_STORE_new(), but document the fact that this variable is still
6332 unused in the certificate verification process.
6333 [Ralf S. Engelschall]
6334
6335 *) Fix the various library and apps files to free up pkeys obtained from
6336 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6337 [Steve Henson]
6338
6339 *) Fix reference counting in X509_PUBKEY_get(). This makes
6340 demos/maurice/example2.c work, amongst others, probably.
6341 [Steve Henson and Ben Laurie]
6342
6343 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6344 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6345 are no longer created. This way we have a single and consistent command
6346 line interface `openssl <command>', similar to `cvs <command>'.
6347 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6348
6349 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6350 BIT STRING wrapper always have zero unused bits.
6351 [Steve Henson]
6352
6353 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6354 [Steve Henson]
6355
6356 *) Make the top-level INSTALL documentation easier to understand.
6357 [Paul Sutton]
6358
6359 *) Makefiles updated to exit if an error occurs in a sub-directory
6360 make (including if user presses ^C) [Paul Sutton]
6361
6362 *) Make Montgomery context stuff explicit in RSA data structure.
6363 [Ben Laurie]
6364
6365 *) Fix build order of pem and err to allow for generated pem.h.
6366 [Ben Laurie]
6367
6368 *) Fix renumbering bug in X509_NAME_delete_entry().
6369 [Ben Laurie]
6370
6371 *) Enhanced the err-ins.pl script so it makes the error library number
6372 global and can add a library name. This is needed for external ASN1 and
6373 other error libraries.
6374 [Steve Henson]
6375
6376 *) Fixed sk_insert which never worked properly.
6377 [Steve Henson]
6378
6379 *) Fix ASN1 macros so they can handle indefinite length construted
6380 EXPLICIT tags. Some non standard certificates use these: they can now
6381 be read in.
6382 [Steve Henson]
6383
6384 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6385 into a single doc/ssleay.txt bundle. This way the information is still
6386 preserved but no longer messes up this directory. Now it's new room for
6387 the new set of documenation files.
6388 [Ralf S. Engelschall]
6389
6390 *) SETs were incorrectly DER encoded. This was a major pain, because they
6391 shared code with SEQUENCEs, which aren't coded the same. This means that
6392 almost everything to do with SETs or SEQUENCEs has either changed name or
6393 number of arguments.
6394 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6395
6396 *) Fix test data to work with the above.
6397 [Ben Laurie]
6398
6399 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6400 was already fixed by Eric for 0.9.1 it seems.
6401 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6402
6403 *) Autodetect FreeBSD3.
6404 [Ben Laurie]
6405
6406 *) Fix various bugs in Configure. This affects the following platforms:
6407 nextstep
6408 ncr-scde
6409 unixware-2.0
6410 unixware-2.0-pentium
6411 sco5-cc.
6412 [Ben Laurie]
6413
6414 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6415 before they are needed.
6416 [Ben Laurie]
6417
6418 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6419 [Ben Laurie]
6420
6421
6422 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6423
6424 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6425 changed SSLeay to OpenSSL in version strings.
6426 [Ralf S. Engelschall]
6427
6428 *) Some fixups to the top-level documents.
6429 [Paul Sutton]
6430
6431 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6432 because the symlink to include/ was missing.
6433 [Ralf S. Engelschall]
6434
6435 *) Incorporated the popular no-RSA/DSA-only patches
6436 which allow to compile a RSA-free SSLeay.
6437 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6438
6439 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6440 when "ssleay" is still not found.
6441 [Ralf S. Engelschall]
6442
6443 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6444 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6445
6446 *) Updated the README file.
6447 [Ralf S. Engelschall]
6448
6449 *) Added various .cvsignore files in the CVS repository subdirs
6450 to make a "cvs update" really silent.
6451 [Ralf S. Engelschall]
6452
6453 *) Recompiled the error-definition header files and added
6454 missing symbols to the Win32 linker tables.
6455 [Ralf S. Engelschall]
6456
6457 *) Cleaned up the top-level documents;
6458 o new files: CHANGES and LICENSE
6459 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6460 o merged COPYRIGHT into LICENSE
6461 o removed obsolete TODO file
6462 o renamed MICROSOFT to INSTALL.W32
6463 [Ralf S. Engelschall]
6464
6465 *) Removed dummy files from the 0.9.1b source tree:
6466 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6467 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6468 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6469 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6470 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6471 [Ralf S. Engelschall]
6472
6473 *) Added various platform portability fixes.
6474 [Mark J. Cox]
6475
6476 *) The Genesis of the OpenSSL rpject:
6477 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6478 Young and Tim J. Hudson created while they were working for C2Net until
6479 summer 1998.
6480 [The OpenSSL Project]
6481
6482
6483 Changes between 0.9.0b and 0.9.1b [not released]
6484
6485 *) Updated a few CA certificates under certs/
6486 [Eric A. Young]
6487
6488 *) Changed some BIGNUM api stuff.
6489 [Eric A. Young]
6490
6491 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6492 DGUX x86, Linux Alpha, etc.
6493 [Eric A. Young]
6494
6495 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6496 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6497 available).
6498 [Eric A. Young]
6499
6500 *) Add -strparse option to asn1pars program which parses nested
6501 binary structures
6502 [Dr Stephen Henson <shenson@bigfoot.com>]
6503
6504 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6505 [Eric A. Young]
6506
6507 *) DSA fix for "ca" program.
6508 [Eric A. Young]
6509
6510 *) Added "-genkey" option to "dsaparam" program.
6511 [Eric A. Young]
6512
6513 *) Added RIPE MD160 (rmd160) message digest.
6514 [Eric A. Young]
6515
6516 *) Added -a (all) option to "ssleay version" command.
6517 [Eric A. Young]
6518
6519 *) Added PLATFORM define which is the id given to Configure.
6520 [Eric A. Young]
6521
6522 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6523 [Eric A. Young]
6524
6525 *) Extended the ASN.1 parser routines.
6526 [Eric A. Young]
6527
6528 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6529 [Eric A. Young]
6530
6531 *) Added a BN_CTX to the BN library.
6532 [Eric A. Young]
6533
6534 *) Fixed the weak key values in DES library
6535 [Eric A. Young]
6536
6537 *) Changed API in EVP library for cipher aliases.
6538 [Eric A. Young]
6539
6540 *) Added support for RC2/64bit cipher.
6541 [Eric A. Young]
6542
6543 *) Converted the lhash library to the crypto/mem.c functions.
6544 [Eric A. Young]
6545
6546 *) Added more recognized ASN.1 object ids.
6547 [Eric A. Young]
6548
6549 *) Added more RSA padding checks for SSL/TLS.
6550 [Eric A. Young]
6551
6552 *) Added BIO proxy/filter functionality.
6553 [Eric A. Young]
6554
6555 *) Added extra_certs to SSL_CTX which can be used
6556 send extra CA certificates to the client in the CA cert chain sending
6557 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6558 [Eric A. Young]
6559
6560 *) Now Fortezza is denied in the authentication phase because
6561 this is key exchange mechanism is not supported by SSLeay at all.
6562 [Eric A. Young]
6563
6564 *) Additional PKCS1 checks.
6565 [Eric A. Young]
6566
6567 *) Support the string "TLSv1" for all TLS v1 ciphers.
6568 [Eric A. Young]
6569
6570 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6571 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6572 [Eric A. Young]
6573
6574 *) Fixed a few memory leaks.
6575 [Eric A. Young]
6576
6577 *) Fixed various code and comment typos.
6578 [Eric A. Young]
6579
6580 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6581 bytes sent in the client random.
6582 [Edward Bishop <ebishop@spyglass.com>]
6583