5 Changes between 0.9.7a and 0.9.8 [xx XXX xxxx]
7 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
8 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9 will now compute a table of multiples of the generator that
10 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11 faster (notably in the case of a single point multiplication,
13 [Nils Larsch, Bodo Moeller]
15 *) IPv6 support for certificate extensions. The various extensions
16 which use the IP:a.b.c.d can now take IPv6 addresses using the
17 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
21 *) Added an ENGINE that implements RSA by performing private key
22 exponentiations with the GMP library. The conversions to and from
23 GMP's mpz_t format aren't optimised nor are any montgomery forms
24 cached, and on x86 it appears OpenSSL's own performance has caught up.
25 However there are likely to be other architectures where GMP could
26 provide a boost. This ENGINE is not built in by default, but it can be
27 specified at Configure time and should be accompanied by the necessary
29 ./config -DOPENSSL_USE_GMP -lgmp
32 *) "openssl engine" will not display ENGINE/DSO load failure errors when
33 testing availability of engines with "-t" - the old behaviour is
34 produced by increasing the feature's verbosity with "-tt".
37 *) ECDSA routines: under certain error conditions uninitialized BN objects
38 could be freed. Solution: make sure initialization is performed early
39 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
43 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
44 and DH_METHOD (eg. by ENGINE implementations) to override the normal
45 software implementations. For DSA and DH, parameter generation can
46 also be overriden by providing the appropriate method callbacks.
49 *) Change the "progress" mechanism used in key-generation and
50 primality testing to functions that take a new BN_GENCB pointer in
51 place of callback/argument pairs. The new API functions have "_ex"
52 postfixes and the older functions are reimplemented as wrappers for
53 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
54 declarations of the old functions to help (graceful) attempts to
55 migrate to the new functions. Also, the new key-generation API
56 functions operate on a caller-supplied key-structure and return
57 success/failure rather than returning a key or NULL - this is to
58 help make "keygen" another member function of RSA_METHOD etc.
60 Example for using the new callback interface:
62 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
66 BN_GENCB_set(&my_cb, my_callback, my_arg);
68 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
69 /* For the meaning of a, b in calls to my_callback(), see the
70 * documentation of the function that calls the callback.
71 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
72 * my_callback should return 1 if it wants BN_is_prime_ex()
73 * to continue, or 0 to stop.
78 *) Change the ZLIB compression method to be stateful, and make it
79 available to TLS with the number defined in
80 draft-ietf-tls-compression-04.txt.
83 *) Add the ASN.1 structures and functions for CertificatePair, which
84 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
86 CertificatePair ::= SEQUENCE {
87 forward [0] Certificate OPTIONAL,
88 reverse [1] Certificate OPTIONAL,
89 -- at least one of the pair shall be present -- }
91 Also implement the PEM functions to read and write certificate
92 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
94 This needed to be defined, mostly for the sake of the LDAP
95 attribute crossCertificatePair, but may prove useful elsewhere as
99 *) Make it possible to inhibit symlinking of shared libraries in
100 Makefile.shared, for Cygwin's sake.
103 *) Extend the BIGNUM API by creating new macros that behave like
106 void BN_set_sign(BIGNUM *a, int neg);
107 int BN_get_sign(const BIGNUM *a);
109 and avoid the need to access 'a->neg' directly in applications.
110 [Nils Larsch <nla@trustcenter.de>]
112 *) Implement fast modular reduction for pseudo-Mersenne primes
113 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
114 EC_GROUP_new_curve_GFp() will now automatically use this
116 [Nils Larsch <nla@trustcenter.de>]
118 *) Add new lock type (CRYPTO_LOCK_BN).
121 *) Change the ENGINE framework to automatically load engines
122 dynamically from specific directories unless they could be
123 found to already be built in or loaded. Move all the
124 current engines except for the cryptodev one to a new
126 The engines in engines/ are built as shared libraries if
127 the "shared" options was given to ./Configure or ./config.
128 Otherwise, they are inserted in libcrypto.a.
129 /usr/local/ssl/engines is the default directory for dynamic
130 engines, but that can be overriden at configure time through
131 the usual use of --prefix and/or --openssldir, and at run
132 time with the environment variable OPENSSL_ENGINES.
133 [Geoff Thorpe and Richard Levitte]
135 *) Add Makefile.shared, a helper makefile to build shared
136 libraries. Addapt Makefile.org.
139 *) Add version info to Win32 DLLs.
140 [Peter 'Luna' Runestig" <peter@runestig.com>]
142 *) Add new 'medium level' PKCS#12 API. Certificates and keys
143 can be added using this API to created arbitrary PKCS#12
144 files while avoiding the low level API.
146 New options to PKCS12_create(), key or cert can be NULL and
147 will then be omitted from the output file. The encryption
148 algorithm NIDs can be set to -1 for no encryption, the mac
149 iteration count can be set to 0 to omit the mac.
151 Enhance pkcs12 utility by making the -nokeys and -nocerts
152 options work when creating a PKCS#12 file. New option -nomac
153 to omit the mac, NONE can be set for an encryption algorithm.
154 New code is modified to use the enhanced PKCS12_create()
155 instead of the low level API.
158 *) Extend ASN1 encoder to support indefinite length constructed
159 encoding. This can output sequences tags and octet strings in
160 this form. Modify pk7_asn1.c to support indefinite length
161 encoding. This is experimental and needs additional code to
162 be useful, such as an ASN1 bio and some enhanced streaming
165 Extend template encode functionality so that tagging is passed
166 down to the template encoder.
169 *) Let 'openssl req' fail if an argument to '-newkey' is not
170 recognized instead of using RSA as a default.
173 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
174 As these are not official, they are not included in "ALL";
175 the "ECCdraft" ciphersuite group alias can be used to select them.
176 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
178 *) Add ECDH engine support.
179 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
181 *) Add ECDH in new directory crypto/ecdh/.
182 TODO: more general interface (return x coordinate, not its hash)
183 TODO: bug: pad x with leading zeros if necessary
184 [Douglas Stebila (Sun Microsystems Laboratories)]
186 *) Let BN_rand_range() abort with an error after 100 iterations
187 without success (which indicates a broken PRNG).
190 *) Change BN_mod_sqrt() so that it verifies that the input value
191 is really the square of the return value. (Previously,
192 BN_mod_sqrt would show GIGO behaviour.)
195 *) Add named elliptic curves over binary fields from X9.62, SECG,
196 and WAP/WTLS; add OIDs that were still missing.
198 [Sheueling Chang Shantz and Douglas Stebila
199 (Sun Microsystems Laboratories)]
201 *) Extend the EC library for elliptic curves over binary fields
202 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
205 EC_GF2m_simple_method
209 EC_GROUP_new_curve_GF2m
210 EC_GROUP_set_curve_GF2m
211 EC_GROUP_get_curve_GF2m
212 EC_POINT_set_affine_coordinates_GF2m
213 EC_POINT_get_affine_coordinates_GF2m
214 EC_POINT_set_compressed_coordinates_GF2m
216 Point compression for binary fields is disabled by default for
217 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
220 As binary polynomials are represented as BIGNUMs, various members
221 of the EC_GROUP and EC_POINT data structures can be shared
222 between the implementations for prime fields and binary fields;
223 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
224 are essentially identical to their ..._GFp counterparts.
225 (For simplicity, the '..._GFp' prefix has been dropped from
226 various internal method names.)
228 An internal 'field_div' method (similar to 'field_mul' and
229 'field_sqr') has been added; this is used only for binary fields.
231 [Sheueling Chang Shantz and Douglas Stebila
232 (Sun Microsystems Laboratories)]
234 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
235 through methods ('mul', 'precompute_mult').
237 The generic implementations (now internally called 'ec_wNAF_mul'
238 and 'ec_wNAF_precomputed_mult') remain the default if these
239 methods are undefined.
241 [Sheueling Chang Shantz and Douglas Stebila
242 (Sun Microsystems Laboratories)]
244 *) New function EC_GROUP_get_degree, which is defined through
245 EC_METHOD. For curves over prime fields, this returns the bit
246 length of the modulus.
248 [Sheueling Chang Shantz and Douglas Stebila
249 (Sun Microsystems Laboratories)]
251 *) New functions EC_GROUP_dup, EC_POINT_dup.
252 (These simply call ..._new and ..._copy).
254 [Sheueling Chang Shantz and Douglas Stebila
255 (Sun Microsystems Laboratories)]
257 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
258 Polynomials are represented as BIGNUMs (where the sign bit is not
259 used) in the following functions [macros]:
262 BN_GF2m_sub [= BN_GF2m_add]
263 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
264 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
265 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
267 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
268 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
269 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
270 BN_GF2m_cmp [= BN_ucmp]
272 (Note that only the 'mod' functions are actually for fields GF(2^m).
273 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
275 For some functions, an the irreducible polynomial defining a
276 field can be given as an 'unsigned int[]' with strictly
277 decreasing elements giving the indices of those bits that are set;
278 i.e., p[] represents the polynomial
279 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
281 p[0] > p[1] > ... > p[k] = 0.
282 This applies to the following functions:
287 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
288 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
291 BN_GF2m_mod_solve_quad_arr
295 Conversion can be performed by the following functions:
300 bntest.c has additional tests for binary polynomial arithmetic.
302 Two implementations for BN_GF2m_mod_div() are available.
303 The default algorithm simply uses BN_GF2m_mod_inv() and
304 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
305 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
306 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
308 [Sheueling Chang Shantz and Douglas Stebila
309 (Sun Microsystems Laboratories)]
311 *) Add new error code 'ERR_R_DISABLED' that can be used when some
312 functionality is disabled at compile-time.
313 [Douglas Stebila <douglas.stebila@sun.com>]
315 *) Change default behaviour of 'openssl asn1parse' so that more
316 information is visible when viewing, e.g., a certificate:
318 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
319 mode the content of non-printable OCTET STRINGs is output in a
320 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
321 avoid the appearance of a printable string.
322 [Nils Larsch <nla@trustcenter.de>]
324 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
326 EC_GROUP_set_asn1_flag()
327 EC_GROUP_get_asn1_flag()
328 EC_GROUP_set_point_conversion_form()
329 EC_GROUP_get_point_conversion_form()
330 These control ASN1 encoding details:
331 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
332 has been set to OPENSSL_EC_NAMED_CURVE.
333 - Points are encoded in uncompressed form by default; options for
334 asn1_for are as for point2oct, namely
335 POINT_CONVERSION_COMPRESSED
336 POINT_CONVERSION_UNCOMPRESSED
337 POINT_CONVERSION_HYBRID
339 Also add 'seed' and 'seed_len' members to EC_GROUP with access
343 EC_GROUP_get_seed_len()
344 This is used only for ASN1 purposes (so far).
345 [Nils Larsch <nla@trustcenter.de>]
347 *) Add 'field_type' member to EC_METHOD, which holds the NID
348 of the appropriate field type OID. The new function
349 EC_METHOD_get_field_type() returns this value.
350 [Nils Larsch <nla@trustcenter.de>]
357 providing useful interfaces to EC_POINT_point2oct() and
358 EC_POINT_oct2point().
359 [Nils Larsch <nla@trustcenter.de>]
361 *) Change internals of the EC library so that the functions
362 EC_GROUP_set_generator()
363 EC_GROUP_get_generator()
365 EC_GROUP_get_cofactor()
366 are implemented directly in crypto/ec/ec_lib.c and not dispatched
367 to methods, which would lead to unnecessary code duplication when
368 adding different types of curves.
369 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
371 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
372 arithmetic, and such that modified wNAFs are generated
373 (which avoid length expansion in many cases).
376 *) Add a function EC_GROUP_check_discriminant() (defined via
377 EC_METHOD) that verifies that the curve discriminant is non-zero.
379 Add a function EC_GROUP_check() that makes some sanity tests
380 on a EC_GROUP, its generator and order. This includes
381 EC_GROUP_check_discriminant().
382 [Nils Larsch <nla@trustcenter.de>]
384 *) Add ECDSA in new directory crypto/ecdsa/.
386 Add applications 'openssl ecparam' and 'openssl ecdsa'
387 (these are based on 'openssl dsaparam' and 'openssl dsa').
389 ECDSA support is also included in various other files across the
390 library. Most notably,
391 - 'openssl req' now has a '-newkey ecdsa:file' option;
392 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
393 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
394 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
395 them suitable for ECDSA where domain parameters must be
396 extracted before the specific public key;
397 - ECDSA engine support has been added.
398 [Nils Larsch <nla@trustcenter.de>]
400 *) Include some named elliptic curves, and add OIDs from X9.62,
401 SECG, and WAP/WTLS. Each curve can be obtained from the new
403 EC_GROUP_new_by_nid(),
404 and the list of available named curves can be obtained with
405 EC_get_builtin_curves().
406 Also add a 'curve_name' member to EC_GROUP objects, which can be
410 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
412 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
413 was actually never needed) and in BN_mul(). The removal in BN_mul()
414 required a small change in bn_mul_part_recursive() and the addition
415 of the functions bn_cmp_part_words(), bn_sub_part_words() and
416 bn_add_part_words(), which do the same thing as bn_cmp_words(),
417 bn_sub_words() and bn_add_words() except they take arrays with
421 Changes between 0.9.7 and 0.9.7a [XX xxx 2003]
423 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
424 Kerberos function mit_des_cbc_cksum(). Before this change,
425 the value returned by DES_cbc_cksum() was like the one from
426 mit_des_cbc_cksum(), except the bytes were swapped.
427 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
429 *) Allow an application to disable the automatic SSL chain building.
430 Before this a rather primitive chain build was always performed in
431 ssl3_output_cert_chain(): an application had no way to send the
432 correct chain if the automatic operation produced an incorrect result.
434 Now the chain builder is disabled if either:
436 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
438 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
440 The reasoning behind this is that an application would not want the
441 auto chain building to take place if extra chain certificates are
442 present and it might also want a means of sending no additional
443 certificates (for example the chain has two certificates and the
447 *) Add the possibility to build without the ENGINE framework.
448 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
450 *) Under Win32 gmtime() can return NULL: check return value in
451 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
454 *) DSA routines: under certain error conditions uninitialized BN objects
455 could be freed. Solution: make sure initialization is performed early
456 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
457 Nils Larsch <nla@trustcenter.de> via PR#459)
460 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
461 checked on reconnect on the client side, therefore session resumption
462 could still fail with a "ssl session id is different" error. This
463 behaviour is masked when SSL_OP_ALL is used due to
464 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
465 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
469 *) IA-32 assembler support enhancements: unified ELF targets, support
470 for SCO/Caldera platforms, fix for Cygwin shared build.
473 *) Add support for FreeBSD on sparc64. As a consequence, support for
474 FreeBSD on non-x86 processors is separate from x86 processors on
475 the config script, much like the NetBSD support.
476 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
478 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
480 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
481 code (06) was taken as the first octet of the session ID and the last
482 octet was ignored consequently. As a result SSLv2 client side session
483 caching could not have worked due to the session ID mismatch between
485 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
489 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
490 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
494 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
495 seems that in spite of existing for more than a year, many application
496 author have done nothing to provide the necessary callbacks, which
497 means that this particular engine will not work properly anywhere.
498 This is a very unfortunate situation which forces us, in the name
499 of usability, to give the hw_ncipher.c a static lock, which is part
501 NOTE: This is for the 0.9.7 series ONLY. This hack will never
502 appear in 0.9.8 or later. We EXPECT application authors to have
503 dealt properly with this when 0.9.8 is released (unless we actually
504 make such changes in the libcrypto locking code that changes will
505 have to be made anyway).
508 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
509 octets have been read, EOF or an error occurs. Without this change
510 some truncated ASN1 structures will not produce an error.
513 *) Disable Heimdal support, since it hasn't been fully implemented.
514 Still give the possibility to force the use of Heimdal, but with
515 warnings and a request that patches get sent to openssl-dev.
518 *) Add the VC-CE target, introduce the WINCE sysname, and add
519 INSTALL.WCE and appropriate conditionals to make it build.
520 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
522 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
523 cygssl-x.y.z.dll, where x, y and z are the major, minor and
524 edit numbers of the version.
525 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
527 *) Introduce safe string copy and catenation functions
528 (BUF_strlcpy() and BUF_strlcat()).
529 [Ben Laurie (CHATS) and Richard Levitte]
531 *) Avoid using fixed-size buffers for one-line DNs.
534 *) Add BUF_MEM_grow_clean() to avoid information leakage when
535 resizing buffers containing secrets, and use where appropriate.
538 *) Avoid using fixed size buffers for configuration file location.
541 *) Avoid filename truncation for various CA files.
544 *) Use sizeof in preference to magic numbers.
547 *) Avoid filename truncation in cert requests.
550 *) Add assertions to check for (supposedly impossible) buffer
554 *) Don't cache truncated DNS entries in the local cache (this could
555 potentially lead to a spoofing attack).
558 *) Fix various buffers to be large enough for hex/decimal
559 representations in a platform independent manner.
562 *) Add CRYPTO_realloc_clean() to avoid information leakage when
563 resizing buffers containing secrets, and use where appropriate.
566 *) Add BIO_indent() to avoid much slightly worrying code to do
570 *) Convert sprintf()/BIO_puts() to BIO_printf().
573 *) buffer_gets() could terminate with the buffer only half
577 *) Add assertions to prevent user-supplied crypto functions from
578 overflowing internal buffers by having large block sizes, etc.
581 *) New OPENSSL_assert() macro (similar to assert(), but enabled
585 *) Eliminate unused copy of key in RC4.
588 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
591 *) Fix off-by-one error in EGD path.
594 *) If RANDFILE path is too long, ignore instead of truncating.
597 *) Eliminate unused and incorrectly sized X.509 structure
601 *) Eliminate unused and dangerous function knumber().
604 *) Eliminate unused and dangerous structure, KSSL_ERR.
607 *) Protect against overlong session ID context length in an encoded
608 session object. Since these are local, this does not appear to be
612 *) Change from security patch (see 0.9.6e below) that did not affect
613 the 0.9.6 release series:
615 Remote buffer overflow in SSL3 protocol - an attacker could
616 supply an oversized master key in Kerberos-enabled versions.
620 *) Change the SSL kerb5 codes to match RFC 2712.
623 *) Make -nameopt work fully for req and add -reqopt switch.
624 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
626 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
627 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
629 *) Make sure tests can be performed even if the corresponding algorithms
630 have been removed entirely. This was also the last step to make
631 OpenSSL compilable with DJGPP under all reasonable conditions.
632 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
634 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
635 to allow version independent disabling of normally unselected ciphers,
636 which may be activated as a side-effect of selecting a single cipher.
638 (E.g., cipher list string "RSA" enables ciphersuites that are left
639 out of "ALL" because they do not provide symmetric encryption.
640 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
641 [Lutz Jaenicke, Bodo Moeller]
643 *) Add appropriate support for separate platform-dependent build
644 directories. The recommended way to make a platform-dependent
645 build directory is the following (tested on Linux), maybe with
648 # Place yourself outside of the OpenSSL source tree. In
649 # this example, the environment variable OPENSSL_SOURCE
650 # is assumed to contain the absolute OpenSSL source directory.
651 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
652 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
653 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
654 mkdir -p `dirname $F`
655 ln -s $OPENSSL_SOURCE/$F $F
658 To be absolutely sure not to disturb the source tree, a "make clean"
659 is a good thing. If it isn't successfull, don't worry about it,
660 it probably means the source directory is very clean.
663 *) Make sure any ENGINE control commands make local copies of string
664 pointers passed to them whenever necessary. Otherwise it is possible
665 the caller may have overwritten (or deallocated) the original string
666 data when a later ENGINE operation tries to use the stored values.
667 [Götz Babin-Ebell <babinebell@trustcenter.de>]
669 *) Improve diagnostics in file reading and command-line digests.
670 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
672 *) Add AES modes CFB and OFB to the object database. Correct an
673 error in AES-CFB decryption.
676 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
677 allows existing EVP_CIPHER_CTX structures to be reused after
678 calling EVP_*Final(). This behaviour is used by encryption
679 BIOs and some applications. This has the side effect that
680 applications must explicitly clean up cipher contexts with
681 EVP_CIPHER_CTX_cleanup() or they will leak memory.
684 *) Check the values of dna and dnb in bn_mul_recursive before calling
685 bn_mul_comba (a non zero value means the a or b arrays do not contain
686 n2 elements) and fallback to bn_mul_normal if either is not zero.
689 *) Fix escaping of non-ASCII characters when using the -subj option
690 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
693 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
694 form for "surname", serialNumber has no short form.
695 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
696 therefore remove "mail" short name for "internet 7".
697 The OID for unique identifiers in X509 certificates is
698 x500UniqueIdentifier, not uniqueIdentifier.
699 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
702 *) Add an "init" command to the ENGINE config module and auto initialize
703 ENGINEs. Without any "init" command the ENGINE will be initialized
704 after all ctrl commands have been executed on it. If init=1 the
705 ENGINE is initailized at that point (ctrls before that point are run
706 on the uninitialized ENGINE and after on the initialized one). If
707 init=0 then the ENGINE will not be iniatialized at all.
710 *) Fix the 'app_verify_callback' interface so that the user-defined
711 argument is actually passed to the callback: In the
712 SSL_CTX_set_cert_verify_callback() prototype, the callback
713 declaration has been changed from
716 int (*cb)(X509_STORE_CTX *,void *);
717 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
718 i=s->ctx->app_verify_callback(&ctx)
719 has been changed into
720 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
722 To update applications using SSL_CTX_set_cert_verify_callback(),
723 a dummy argument can be added to their callback functions.
724 [D. K. Smetters <smetters@parc.xerox.com>]
726 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
727 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
729 *) Add and OPENSSL_LOAD_CONF define which will cause
730 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
731 This allows older applications to transparently support certain
732 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
733 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
734 load the config file and OPENSSL_add_all_algorithms_conf() which will
735 always load it have also been added.
738 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
739 Adjust NIDs and EVP layer.
740 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
742 *) Config modules support in openssl utility.
744 Most commands now load modules from the config file,
745 though in a few (such as version) this isn't done
746 because it couldn't be used for anything.
748 In the case of ca and req the config file used is
749 the same as the utility itself: that is the -config
750 command line option can be used to specify an
754 *) Move default behaviour from OPENSSL_config(). If appname is NULL
755 use "openssl_conf" if filename is NULL use default openssl config file.
758 *) Add an argument to OPENSSL_config() to allow the use of an alternative
759 config section name. Add a new flag to tolerate a missing config file
760 and move code to CONF_modules_load_file().
763 *) Support for crypto accelerator cards from Accelerated Encryption
764 Processing, www.aep.ie. (Use engine 'aep')
765 The support was copied from 0.9.6c [engine] and adapted/corrected
766 to work with the new engine framework.
767 [AEP Inc. and Richard Levitte]
769 *) Support for SureWare crypto accelerator cards from Baltimore
770 Technologies. (Use engine 'sureware')
771 The support was copied from 0.9.6c [engine] and adapted
772 to work with the new engine framework.
775 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
776 make the newer ENGINE framework commands for the CHIL engine work.
777 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
779 *) Make it possible to produce shared libraries on ReliantUNIX.
780 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
782 *) Add the configuration target debug-linux-ppro.
783 Make 'openssl rsa' use the general key loading routines
784 implemented in apps.c, and make those routines able to
785 handle the key format FORMAT_NETSCAPE and the variant
787 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
789 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
790 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
792 *) Add -keyform to rsautl, and document -engine.
793 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
795 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
796 BIO_R_NO_SUCH_FILE error code rather than the generic
797 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
802 ERR_peek_last_error_line
803 ERR_peek_last_error_line_data.
807 ERR_peek_error_line_data,
808 but report on the latest error recorded rather than the first one
809 still in the error queue.
810 [Ben Laurie, Bodo Moeller]
812 *) default_algorithms option in ENGINE config module. This allows things
814 default_algorithms = ALL
815 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
818 *) Prelminary ENGINE config module.
821 *) New experimental application configuration code.
824 *) Change the AES code to follow the same name structure as all other
825 symmetric ciphers, and behave the same way. Move everything to
826 the directory crypto/aes, thereby obsoleting crypto/rijndael.
827 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
829 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
830 [Ben Laurie and Theo de Raadt]
832 *) Add option to output public keys in req command.
833 [Massimiliano Pala madwolf@openca.org]
835 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
836 (up to about 10% better than before for P-192 and P-224).
839 *) New functions/macros
841 SSL_CTX_set_msg_callback(ctx, cb)
842 SSL_CTX_set_msg_callback_arg(ctx, arg)
843 SSL_set_msg_callback(ssl, cb)
844 SSL_set_msg_callback_arg(ssl, arg)
846 to request calling a callback function
848 void cb(int write_p, int version, int content_type,
849 const void *buf, size_t len, SSL *ssl, void *arg)
851 whenever a protocol message has been completely received
852 (write_p == 0) or sent (write_p == 1). Here 'version' is the
853 protocol version according to which the SSL library interprets
854 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
855 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
856 the content type as defined in the SSL 3.0/TLS 1.0 protocol
857 specification (change_cipher_spec(20), alert(21), handshake(22)).
858 'buf' and 'len' point to the actual message, 'ssl' to the
859 SSL object, and 'arg' is the application-defined value set by
860 SSL[_CTX]_set_msg_callback_arg().
862 'openssl s_client' and 'openssl s_server' have new '-msg' options
863 to enable a callback that displays all protocol messages.
866 *) Change the shared library support so shared libraries are built as
867 soon as the corresponding static library is finished, and thereby get
868 openssl and the test programs linked against the shared library.
869 This still only happens when the keyword "shard" has been given to
870 the configuration scripts.
872 NOTE: shared library support is still an experimental thing, and
873 backward binary compatibility is still not guaranteed.
874 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
876 *) Add support for Subject Information Access extension.
877 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
879 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
880 additional bytes when new memory had to be allocated, not just
881 when reusing an existing buffer.
884 *) New command line and configuration option 'utf8' for the req command.
885 This allows field values to be specified as UTF8 strings.
888 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
889 runs for the former and machine-readable output for the latter.
892 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
893 of the e-mail address in the DN (i.e., it will go into a certificate
894 extension only). The new configuration file option 'email_in_dn = no'
896 [Massimiliano Pala madwolf@openca.org]
898 *) Change all functions with names starting with des_ to be starting
899 with DES_ instead. Add wrappers that are compatible with libdes,
900 but are named _ossl_old_des_*. Finally, add macros that map the
901 des_* symbols to the corresponding _ossl_old_des_* if libdes
902 compatibility is desired. If OpenSSL 0.9.6c compatibility is
903 desired, the des_* symbols will be mapped to DES_*, with one
906 Since we provide two compatibility mappings, the user needs to
907 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
908 compatibility is desired. The default (i.e., when that macro
909 isn't defined) is OpenSSL 0.9.6c compatibility.
911 There are also macros that enable and disable the support of old
912 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
913 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
914 are defined, the default will apply: to support the old des routines.
916 In either case, one must include openssl/des.h to get the correct
917 definitions. Do not try to just include openssl/des_old.h, that
920 NOTE: This is a major break of an old API into a new one. Software
921 authors are encouraged to switch to the DES_ style functions. Some
922 time in the future, des_old.h and the libdes compatibility functions
923 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
924 default), and then completely removed.
927 *) Test for certificates which contain unsupported critical extensions.
928 If such a certificate is found during a verify operation it is
929 rejected by default: this behaviour can be overridden by either
930 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
931 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
932 X509_supported_extension() has also been added which returns 1 if a
933 particular extension is supported.
936 *) Modify the behaviour of EVP cipher functions in similar way to digests
937 to retain compatibility with existing code.
940 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
941 compatibility with existing code. In particular the 'ctx' parameter does
942 not have to be to be initialized before the call to EVP_DigestInit() and
943 it is tidied up after a call to EVP_DigestFinal(). New function
944 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
945 EVP_MD_CTX_copy() changed to not require the destination to be
946 initialized valid and new function EVP_MD_CTX_copy_ex() added which
947 requires the destination to be valid.
949 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
950 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
953 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
954 so that complete 'Handshake' protocol structures are kept in memory
955 instead of overwriting 'msg_type' and 'length' with 'body' data.
958 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
959 [Massimo Santin via Richard Levitte]
961 *) Major restructuring to the underlying ENGINE code. This includes
962 reduction of linker bloat, separation of pure "ENGINE" manipulation
963 (initialisation, etc) from functionality dealing with implementations
964 of specific crypto iterfaces. This change also introduces integrated
965 support for symmetric ciphers and digest implementations - so ENGINEs
966 can now accelerate these by providing EVP_CIPHER and EVP_MD
967 implementations of their own. This is detailed in crypto/engine/README
968 as it couldn't be adequately described here. However, there are a few
969 API changes worth noting - some RSA, DSA, DH, and RAND functions that
970 were changed in the original introduction of ENGINE code have now
971 reverted back - the hooking from this code to ENGINE is now a good
972 deal more passive and at run-time, operations deal directly with
973 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
974 dereferencing through an ENGINE pointer any more. Also, the ENGINE
975 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
976 they were not being used by the framework as there is no concept of a
977 BIGNUM_METHOD and they could not be generalised to the new
978 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
979 ENGINE_cpy() has been removed as it cannot be consistently defined in
983 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
986 *) Change mkdef.pl to sort symbols that get the same entry number,
987 and make sure the automatically generated functions ERR_load_*
988 become part of libeay.num as well.
991 *) New function SSL_renegotiate_pending(). This returns true once
992 renegotiation has been requested (either SSL_renegotiate() call
993 or HelloRequest/ClientHello receveived from the peer) and becomes
994 false once a handshake has been completed.
995 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
996 sends a HelloRequest, but does not ensure that a handshake takes
997 place. SSL_renegotiate_pending() is useful for checking if the
998 client has followed the request.)
1001 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1002 By default, clients may request session resumption even during
1003 renegotiation (if session ID contexts permit); with this option,
1004 session resumption is possible only in the first handshake.
1006 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1007 more bits available for options that should not be part of
1008 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1011 *) Add some demos for certificate and certificate request creation.
1014 *) Make maximum certificate chain size accepted from the peer application
1015 settable (SSL*_get/set_max_cert_list()), as proposed by
1016 "Douglas E. Engert" <deengert@anl.gov>.
1019 *) Add support for shared libraries for Unixware-7
1020 (Boyd Lynn Gerber <gerberb@zenez.com>).
1023 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1024 be done prior to destruction. Use this to unload error strings from
1025 ENGINEs that load their own error strings. NB: This adds two new API
1026 functions to "get" and "set" this destroy handler in an ENGINE.
1029 *) Alter all existing ENGINE implementations (except "openssl" and
1030 "openbsd") to dynamically instantiate their own error strings. This
1031 makes them more flexible to be built both as statically-linked ENGINEs
1032 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1033 Also, add stub code to each that makes building them as self-contained
1034 shared-libraries easier (see README.ENGINE).
1037 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1038 implementations into applications that are completely implemented in
1039 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1040 commands that can be used to configure what shared-library to load and
1041 to control aspects of the way it is handled. Also, made an update to
1042 the README.ENGINE file that brings its information up-to-date and
1043 provides some information and instructions on the "dynamic" ENGINE
1044 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1047 *) Make it possible to unload ranges of ERR strings with a new
1048 "ERR_unload_strings" function.
1051 *) Add a copy() function to EVP_MD.
1054 *) Make EVP_MD routines take a context pointer instead of just the
1055 md_data void pointer.
1058 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1059 that the digest can only process a single chunk of data
1060 (typically because it is provided by a piece of
1061 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1062 is only going to provide a single chunk of data, and hence the
1063 framework needn't accumulate the data for oneshot drivers.
1066 *) As with "ERR", make it possible to replace the underlying "ex_data"
1067 functions. This change also alters the storage and management of global
1068 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1069 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1070 index counters. The API functions that use this state have been changed
1071 to take a "class_index" rather than pointers to the class's local STACK
1072 and counter, and there is now an API function to dynamically create new
1073 classes. This centralisation allows us to (a) plug a lot of the
1074 thread-safety problems that existed, and (b) makes it possible to clean
1075 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1076 such data would previously have always leaked in application code and
1077 workarounds were in place to make the memory debugging turn a blind eye
1078 to it. Application code that doesn't use this new function will still
1079 leak as before, but their memory debugging output will announce it now
1080 rather than letting it slide.
1082 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1083 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1084 has a return value to indicate success or failure.
1087 *) Make it possible to replace the underlying "ERR" functions such that the
1088 global state (2 LHASH tables and 2 locks) is only used by the "default"
1089 implementation. This change also adds two functions to "get" and "set"
1090 the implementation prior to it being automatically set the first time
1091 any other ERR function takes place. Ie. an application can call "get",
1092 pass the return value to a module it has just loaded, and that module
1093 can call its own "set" function using that value. This means the
1094 module's "ERR" operations will use (and modify) the error state in the
1095 application and not in its own statically linked copy of OpenSSL code.
1098 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1099 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1100 the operation, and provides a more encapsulated way for external code
1101 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1102 to use these functions rather than manually incrementing the counts.
1104 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1107 *) Add EVP test program.
1110 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1113 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1114 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1115 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1116 These allow a CRL to be built without having to access X509_CRL fields
1117 directly. Modify 'ca' application to use new functions.
1120 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1121 bug workarounds. Rollback attack detection is a security feature.
1122 The problem will only arise on OpenSSL servers when TLSv1 is not
1123 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1124 Software authors not wanting to support TLSv1 will have special reasons
1125 for their choice and can explicitly enable this option.
1126 [Bodo Moeller, Lutz Jaenicke]
1128 *) Rationalise EVP so it can be extended: don't include a union of
1129 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1130 (similar to those existing for EVP_CIPHER_CTX).
1135 EVP_MD_CTX_init(&md); /* new function call */
1136 EVP_DigestInit(&md, EVP_sha1());
1137 EVP_DigestUpdate(&md, in, len);
1138 EVP_DigestFinal(&md, out, NULL);
1139 EVP_MD_CTX_cleanup(&md); /* new function call */
1143 *) Make DES key schedule conform to the usual scheme, as well as
1144 correcting its structure. This means that calls to DES functions
1145 now have to pass a pointer to a des_key_schedule instead of a
1146 plain des_key_schedule (which was actually always a pointer
1149 des_key_schedule ks;
1151 des_set_key_checked(..., &ks);
1152 des_ncbc_encrypt(..., &ks, ...);
1154 (Note that a later change renames 'des_...' into 'DES_...'.)
1157 *) Initial reduction of linker bloat: the use of some functions, such as
1158 PEM causes large amounts of unused functions to be linked in due to
1159 poor organisation. For example pem_all.c contains every PEM function
1160 which has a knock on effect of linking in large amounts of (unused)
1161 ASN1 code. Grouping together similar functions and splitting unrelated
1162 functions prevents this.
1165 *) Cleanup of EVP macros.
1168 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1169 correct _ecb suffix.
1172 *) Add initial OCSP responder support to ocsp application. The
1173 revocation information is handled using the text based index
1174 use by the ca application. The responder can either handle
1175 requests generated internally, supplied in files (for example
1176 via a CGI script) or using an internal minimal server.
1179 *) Add configuration choices to get zlib compression for TLS.
1182 *) Changes to Kerberos SSL for RFC 2712 compliance:
1183 1. Implemented real KerberosWrapper, instead of just using
1184 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1185 2. Implemented optional authenticator field of KerberosWrapper.
1187 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1188 and authenticator structs; see crypto/krb5/.
1190 Generalized Kerberos calls to support multiple Kerberos libraries.
1191 [Vern Staats <staatsvr@asc.hpc.mil>,
1192 Jeffrey Altman <jaltman@columbia.edu>
1193 via Richard Levitte]
1195 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1196 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1197 values for each of the key sizes rather than having just
1198 parameters (and 'speed' generating keys each time).
1201 *) Speed up EVP routines.
1204 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1205 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1206 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1207 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1209 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1210 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1211 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1214 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1216 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1219 *) Added the OS2-EMX target.
1220 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1222 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1223 to support NCONF routines in extension code. New function CONF_set_nconf()
1224 to allow functions which take an NCONF to also handle the old LHASH
1225 structure: this means that the old CONF compatible routines can be
1226 retained (in particular wrt extensions) without having to duplicate the
1227 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1230 *) Enhance the general user interface with mechanisms for inner control
1231 and with possibilities to have yes/no kind of prompts.
1234 *) Change all calls to low level digest routines in the library and
1235 applications to use EVP. Add missing calls to HMAC_cleanup() and
1236 don't assume HMAC_CTX can be copied using memcpy().
1237 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1239 *) Add the possibility to control engines through control names but with
1240 arbitrary arguments instead of just a string.
1241 Change the key loaders to take a UI_METHOD instead of a callback
1242 function pointer. NOTE: this breaks binary compatibility with earlier
1243 versions of OpenSSL [engine].
1244 Adapt the nCipher code for these new conditions and add a card insertion
1248 *) Enhance the general user interface with mechanisms to better support
1249 dialog box interfaces, application-defined prompts, the possibility
1250 to use defaults (for example default passwords from somewhere else)
1251 and interrupts/cancellations.
1254 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1255 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1258 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1259 tidy up some unnecessarily weird code in 'sk_new()').
1260 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1262 *) Change the key loading routines for ENGINEs to use the same kind
1263 callback (pem_password_cb) as all other routines that need this
1267 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1268 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1269 than this minimum value is recommended.
1272 *) New random seeder for OpenVMS, using the system process statistics
1273 that are easily reachable.
1276 *) Windows apparently can't transparently handle global
1277 variables defined in DLLs. Initialisations such as:
1279 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1281 wont compile. This is used by the any applications that need to
1282 declare their own ASN1 modules. This was fixed by adding the option
1283 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1284 needed for static libraries under Win32.
1287 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1288 setting of purpose and trust fields. New X509_STORE trust and
1289 purpose functions and tidy up setting in other SSL functions.
1292 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1293 structure. These are inherited by X509_STORE_CTX when it is
1294 initialised. This allows various defaults to be set in the
1295 X509_STORE structure (such as flags for CRL checking and custom
1296 purpose or trust settings) for functions which only use X509_STORE_CTX
1297 internally such as S/MIME.
1299 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1300 trust settings if they are not set in X509_STORE. This allows X509_STORE
1301 purposes and trust (in S/MIME for example) to override any set by default.
1303 Add command line options for CRL checking to smime, s_client and s_server
1307 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1308 are set then the CRL is looked up in the X509_STORE structure and
1309 its validity and signature checked, then if the certificate is found
1310 in the CRL the verify fails with a revoked error.
1312 Various new CRL related callbacks added to X509_STORE_CTX structure.
1314 Command line options added to 'verify' application to support this.
1316 This needs some additional work, such as being able to handle multiple
1317 CRLs with different times, extension based lookup (rather than just
1318 by subject name) and ultimately more complete V2 CRL extension
1322 *) Add a general user interface API (crypto/ui/). This is designed
1323 to replace things like des_read_password and friends (backward
1324 compatibility functions using this new API are provided).
1325 The purpose is to remove prompting functions from the DES code
1326 section as well as provide for prompting through dialog boxes in
1327 a window system and the like.
1330 *) Add "ex_data" support to ENGINE so implementations can add state at a
1331 per-structure level rather than having to store it globally.
1334 *) Make it possible for ENGINE structures to be copied when retrieved by
1335 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1336 This causes the "original" ENGINE structure to act like a template,
1337 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1338 operational state can be localised to each ENGINE structure, despite the
1339 fact they all share the same "methods". New ENGINE structures returned in
1340 this case have no functional references and the return value is the single
1341 structural reference. This matches the single structural reference returned
1342 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1346 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1347 needs to match any other type at all we need to manually clear the
1351 *) Changes to the "openssl engine" utility to include;
1352 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1353 about an ENGINE's available control commands.
1354 - executing control commands from command line arguments using the
1355 '-pre' and '-post' switches. '-post' is only used if '-t' is
1356 specified and the ENGINE is successfully initialised. The syntax for
1357 the individual commands are colon-separated, for example;
1358 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1361 *) New dynamic control command support for ENGINEs. ENGINEs can now
1362 declare their own commands (numbers), names (strings), descriptions,
1363 and input types for run-time discovery by calling applications. A
1364 subset of these commands are implicitly classed as "executable"
1365 depending on their input type, and only these can be invoked through
1366 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1367 can be based on user input, config files, etc). The distinction is
1368 that "executable" commands cannot return anything other than a boolean
1369 result and can only support numeric or string input, whereas some
1370 discoverable commands may only be for direct use through
1371 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1372 pointers, or other custom uses. The "executable" commands are to
1373 support parameterisations of ENGINE behaviour that can be
1374 unambiguously defined by ENGINEs and used consistently across any
1375 OpenSSL-based application. Commands have been added to all the
1376 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1377 control over shared-library paths without source code alterations.
1380 *) Changed all ENGINE implementations to dynamically allocate their
1381 ENGINEs rather than declaring them statically. Apart from this being
1382 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1383 this also allows the implementations to compile without using the
1384 internal engine_int.h header.
1387 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1388 'const' value. Any code that should be able to modify a RAND_METHOD
1389 should already have non-const pointers to it (ie. they should only
1390 modify their own ones).
1393 *) Made a variety of little tweaks to the ENGINE code.
1394 - "atalla" and "ubsec" string definitions were moved from header files
1395 to C code. "nuron" string definitions were placed in variables
1396 rather than hard-coded - allowing parameterisation of these values
1397 later on via ctrl() commands.
1398 - Removed unused "#if 0"'d code.
1399 - Fixed engine list iteration code so it uses ENGINE_free() to release
1400 structural references.
1401 - Constified the RAND_METHOD element of ENGINE structures.
1402 - Constified various get/set functions as appropriate and added
1403 missing functions (including a catch-all ENGINE_cpy that duplicates
1404 all ENGINE values onto a new ENGINE except reference counts/state).
1405 - Removed NULL parameter checks in get/set functions. Setting a method
1406 or function to NULL is a way of cancelling out a previously set
1407 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1408 and doesn't justify the extra error symbols and code.
1409 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1410 flags from engine_int.h to engine.h.
1411 - Changed prototypes for ENGINE handler functions (init(), finish(),
1412 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1415 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1416 to the algorithm using long division. The binary algorithm can be
1417 used only if the modulus is odd. On 32-bit systems, it is faster
1418 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1419 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1420 up to 450 bits. In 64-bit environments, the binary algorithm
1421 appears to be advantageous for much longer moduli; here we use it
1422 for moduli up to 2048 bits.
1425 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1426 could not support the combine flag in choice fields.
1429 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1430 extensions from a certificate request to the certificate.
1433 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1434 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1435 file: this allows the display of the certificate about to be
1436 signed to be customised, to allow certain fields to be included
1437 or excluded and extension details. The old system didn't display
1438 multicharacter strings properly, omitted fields not in the policy
1439 and couldn't display additional details such as extensions.
1442 *) Function EC_POINTs_mul for multiple scalar multiplication
1443 of an arbitrary number of elliptic curve points
1444 \sum scalars[i]*points[i],
1445 optionally including the generator defined for the EC_GROUP:
1446 scalar*generator + \sum scalars[i]*points[i].
1448 EC_POINT_mul is a simple wrapper function for the typical case
1449 that the point list has just one item (besides the optional
1453 *) First EC_METHODs for curves over GF(p):
1455 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1456 operations and provides various method functions that can also
1457 operate with faster implementations of modular arithmetic.
1459 EC_GFp_mont_method() reuses most functions that are part of
1460 EC_GFp_simple_method, but uses Montgomery arithmetic.
1462 [Bodo Moeller; point addition and point doubling
1463 implementation directly derived from source code provided by
1464 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1466 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1467 crypto/ec/ec_lib.c):
1469 Curves are EC_GROUP objects (with an optional group generator)
1470 based on EC_METHODs that are built into the library.
1472 Points are EC_POINT objects based on EC_GROUP objects.
1474 Most of the framework would be able to handle curves over arbitrary
1475 finite fields, but as there are no obvious types for fields other
1476 than GF(p), some functions are limited to that for now.
1479 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1480 that the file contains a complete HTTP response.
1483 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1484 change the def and num file printf format specifier from "%-40sXXX"
1485 to "%-39s XXX". The latter will always guarantee a space after the
1486 field while the former will cause them to run together if the field
1487 is 40 of more characters long.
1490 *) Constify the cipher and digest 'method' functions and structures
1491 and modify related functions to take constant EVP_MD and EVP_CIPHER
1495 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1496 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1499 *) Modify EVP_Digest*() routines so they now return values. Although the
1500 internal software routines can never fail additional hardware versions
1504 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1506 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1507 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1510 ERR_R_NESTED_ASN1_ERROR
1512 ERR_R_MISSING_ASN1_EOS
1513 were 4 .. 9, conflicting with
1514 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1516 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1517 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1519 Add new error code 'ERR_R_INTERNAL_ERROR'.
1522 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1526 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1527 sets the subject name for a new request or supersedes the
1528 subject name in a given request. Formats that can be parsed are
1529 'CN=Some Name, OU=myOU, C=IT'
1531 'CN=Some Name/OU=myOU/C=IT'.
1533 Add options '-batch' and '-verbose' to 'openssl req'.
1534 [Massimiliano Pala <madwolf@hackmasters.net>]
1536 *) Introduce the possibility to access global variables through
1537 functions on platform were that's the best way to handle exporting
1538 global variables in shared libraries. To enable this functionality,
1539 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1540 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1541 is normally done by Configure or something similar).
1543 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1544 in the source file (foo.c) like this:
1546 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1547 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1549 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1550 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1552 OPENSSL_DECLARE_GLOBAL(int,foo);
1553 #define foo OPENSSL_GLOBAL_REF(foo)
1554 OPENSSL_DECLARE_GLOBAL(double,bar);
1555 #define bar OPENSSL_GLOBAL_REF(bar)
1557 The #defines are very important, and therefore so is including the
1558 header file everywhere where the defined globals are used.
1560 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1561 of ASN.1 items, but that structure is a bit different.
1563 The largest change is in util/mkdef.pl which has been enhanced with
1564 better and easier to understand logic to choose which symbols should
1565 go into the Windows .def files as well as a number of fixes and code
1566 cleanup (among others, algorithm keywords are now sorted
1567 lexicographically to avoid constant rewrites).
1570 *) In BN_div() keep a copy of the sign of 'num' before writing the
1571 result to 'rm' because if rm==num the value will be overwritten
1572 and produce the wrong result if 'num' is negative: this caused
1573 problems with BN_mod() and BN_nnmod().
1576 *) Function OCSP_request_verify(). This checks the signature on an
1577 OCSP request and verifies the signer certificate. The signer
1578 certificate is just checked for a generic purpose and OCSP request
1582 *) Add OCSP_check_validity() function to check the validity of OCSP
1583 responses. OCSP responses are prepared in real time and may only
1584 be a few seconds old. Simply checking that the current time lies
1585 between thisUpdate and nextUpdate max reject otherwise valid responses
1586 caused by either OCSP responder or client clock inaccuracy. Instead
1587 we allow thisUpdate and nextUpdate to fall within a certain period of
1588 the current time. The age of the response can also optionally be
1589 checked. Two new options -validity_period and -status_age added to
1593 *) If signature or public key algorithm is unrecognized print out its
1594 OID rather that just UNKNOWN.
1597 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1598 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1599 ID to be generated from the issuer certificate alone which can then be
1600 passed to OCSP_id_issuer_cmp().
1603 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1604 ASN1 modules to export functions returning ASN1_ITEM pointers
1605 instead of the ASN1_ITEM structures themselves. This adds several
1606 new macros which allow the underlying ASN1 function/structure to
1607 be accessed transparently. As a result code should not use ASN1_ITEM
1608 references directly (such as &X509_it) but instead use the relevant
1609 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1610 use of the new ASN1 code on platforms where exporting structures
1611 is problematical (for example in shared libraries) but exporting
1612 functions returning pointers to structures is not.
1615 *) Add support for overriding the generation of SSL/TLS session IDs.
1616 These callbacks can be registered either in an SSL_CTX or per SSL.
1617 The purpose of this is to allow applications to control, if they wish,
1618 the arbitrary values chosen for use as session IDs, particularly as it
1619 can be useful for session caching in multiple-server environments. A
1620 command-line switch for testing this (and any client code that wishes
1621 to use such a feature) has been added to "s_server".
1622 [Geoff Thorpe, Lutz Jaenicke]
1624 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1625 of the form '#if defined(...) || defined(...) || ...' and
1626 '#if !defined(...) && !defined(...) && ...'. This also avoids
1627 the growing number of special cases it was previously handling.
1630 *) Make all configuration macros available for application by making
1631 sure they are available in opensslconf.h, by giving them names starting
1632 with "OPENSSL_" to avoid conflicts with other packages and by making
1633 sure e_os2.h will cover all platform-specific cases together with
1635 Additionally, it is now possible to define configuration/platform-
1636 specific names (called "system identities"). In the C code, these
1637 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1638 macro with the name beginning with "OPENSSL_SYS_", which is determined
1639 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1643 *) New option -set_serial to 'req' and 'x509' this allows the serial
1644 number to use to be specified on the command line. Previously self
1645 signed certificates were hard coded with serial number 0 and the
1646 CA options of 'x509' had to use a serial number in a file which was
1650 *) New options to 'ca' utility to support V2 CRL entry extensions.
1651 Currently CRL reason, invalidity date and hold instruction are
1652 supported. Add new CRL extensions to V3 code and some new objects.
1655 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1656 disable standard block padding (aka PKCS#5 padding) in the EVP
1657 API, which was previously mandatory. This means that the data is
1658 not padded in any way and so the total length much be a multiple
1659 of the block size, otherwise an error occurs.
1662 *) Initial (incomplete) OCSP SSL support.
1665 *) New function OCSP_parse_url(). This splits up a URL into its host,
1666 port and path components: primarily to parse OCSP URLs. New -url
1667 option to ocsp utility.
1670 *) New nonce behavior. The return value of OCSP_check_nonce() now
1671 reflects the various checks performed. Applications can decide
1672 whether to tolerate certain situations such as an absent nonce
1673 in a response when one was present in a request: the ocsp application
1674 just prints out a warning. New function OCSP_add1_basic_nonce()
1675 this is to allow responders to include a nonce in a response even if
1676 the request is nonce-less.
1679 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1680 skipped when using openssl x509 multiple times on a single input file,
1681 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1684 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1685 set string type: to handle setting ASN1_TIME structures. Fix ca
1686 utility to correctly initialize revocation date of CRLs.
1689 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1690 the clients preferred ciphersuites and rather use its own preferences.
1691 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1692 Internet Explorer by ensuring unchanged hash method during stepup.
1693 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1696 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1697 to aes and add a new 'exist' option to print out symbols that don't
1701 *) Additional options to ocsp utility to allow flags to be set and
1702 additional certificates supplied.
1705 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1706 OCSP client a number of certificate to only verify the response
1710 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1711 handle the new API. Currently only ECB, CBC modes supported. Add new
1714 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1715 Encryption Standard (AES) Ciphersuites for Transport Layer
1716 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1717 not enabled by default and were not part of the "ALL" ciphersuite
1718 alias because they were not yet official; they could be
1719 explicitly requested by specifying the "AESdraft" ciphersuite
1720 group alias. In the final release of OpenSSL 0.9.7, the group
1721 alias is called "AES" and is part of "ALL".)
1722 [Ben Laurie, Steve Henson, Bodo Moeller]
1724 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1725 request to response.
1728 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1729 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1730 extract information from a certificate request. OCSP_response_create()
1731 creates a response and optionally adds a basic response structure.
1732 OCSP_basic_add1_status() adds a complete single response to a basic
1733 response and returns the OCSP_SINGLERESP structure just added (to allow
1734 extensions to be included for example). OCSP_basic_add1_cert() adds a
1735 certificate to a basic response and OCSP_basic_sign() signs a basic
1736 response with various flags. New helper functions ASN1_TIME_check()
1737 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1738 (converts ASN1_TIME to GeneralizedTime).
1741 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1742 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1743 structure from a certificate. X509_pubkey_digest() digests the public_key
1744 contents: this is used in various key identifiers.
1747 *) Make sk_sort() tolerate a NULL argument.
1748 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1750 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1751 passed by the function are trusted implicitly. If any of them signed the
1752 response then it is assumed to be valid and is not verified.
1755 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1756 to data. This was previously part of the PKCS7 ASN1 code. This
1757 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1758 [Steve Henson, reported by Kenneth R. Robinette
1759 <support@securenetterm.com>]
1761 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1762 routines: without these tracing memory leaks is very painful.
1763 Fix leaks in PKCS12 and PKCS7 routines.
1766 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1767 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1768 effectively meant GeneralizedTime would never be used. Now it
1769 is initialised to -1 but X509_time_adj() now has to check the value
1770 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1771 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1772 [Steve Henson, reported by Kenneth R. Robinette
1773 <support@securenetterm.com>]
1775 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1776 result in a zero length in the ASN1_INTEGER structure which was
1777 not consistent with the structure when d2i_ASN1_INTEGER() was used
1778 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1779 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1780 where it did not print out a minus for negative ASN1_INTEGER.
1783 *) Add summary printout to ocsp utility. The various functions which
1784 convert status values to strings have been renamed to:
1785 OCSP_response_status_str(), OCSP_cert_status_str() and
1786 OCSP_crl_reason_str() and are no longer static. New options
1787 to verify nonce values and to disable verification. OCSP response
1788 printout format cleaned up.
1791 *) Add additional OCSP certificate checks. These are those specified
1792 in RFC2560. This consists of two separate checks: the CA of the
1793 certificate being checked must either be the OCSP signer certificate
1794 or the issuer of the OCSP signer certificate. In the latter case the
1795 OCSP signer certificate must contain the OCSP signing extended key
1796 usage. This check is performed by attempting to match the OCSP
1797 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1798 in the OCSP_CERTID structures of the response.
1801 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1802 and related routines. This uses the standard OpenSSL certificate
1803 verify routines to perform initial checks (just CA validity) and
1804 to obtain the certificate chain. Then additional checks will be
1805 performed on the chain. Currently the root CA is checked to see
1806 if it is explicitly trusted for OCSP signing. This is used to set
1807 a root CA as a global signing root: that is any certificate that
1808 chains to that CA is an acceptable OCSP signing certificate.
1811 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1812 extensions from a separate configuration file.
1813 As when reading extensions from the main configuration file,
1814 the '-extensions ...' option may be used for specifying the
1816 [Massimiliano Pala <madwolf@comune.modena.it>]
1818 *) New OCSP utility. Allows OCSP requests to be generated or
1819 read. The request can be sent to a responder and the output
1820 parsed, outputed or printed in text form. Not complete yet:
1821 still needs to check the OCSP response validity.
1824 *) New subcommands for 'openssl ca':
1825 'openssl ca -status <serial>' prints the status of the cert with
1826 the given serial number (according to the index file).
1827 'openssl ca -updatedb' updates the expiry status of certificates
1829 [Massimiliano Pala <madwolf@comune.modena.it>]
1831 *) New '-newreq-nodes' command option to CA.pl. This is like
1832 '-newreq', but calls 'openssl req' with the '-nodes' option
1833 so that the resulting key is not encrypted.
1834 [Damien Miller <djm@mindrot.org>]
1836 *) New configuration for the GNU Hurd.
1837 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1839 *) Initial code to implement OCSP basic response verify. This
1840 is currently incomplete. Currently just finds the signer's
1841 certificate and verifies the signature on the response.
1844 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1845 value of OPENSSLDIR. This is available via the new '-d' option
1846 to 'openssl version', and is also included in 'openssl version -a'.
1849 *) Allowing defining memory allocation callbacks that will be given
1850 file name and line number information in additional arguments
1851 (a const char* and an int). The basic functionality remains, as
1852 well as the original possibility to just replace malloc(),
1853 realloc() and free() by functions that do not know about these
1854 additional arguments. To register and find out the current
1855 settings for extended allocation functions, the following
1856 functions are provided:
1858 CRYPTO_set_mem_ex_functions
1859 CRYPTO_set_locked_mem_ex_functions
1860 CRYPTO_get_mem_ex_functions
1861 CRYPTO_get_locked_mem_ex_functions
1863 These work the same way as CRYPTO_set_mem_functions and friends.
1864 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1865 extended allocation function is enabled.
1866 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1867 a conventional allocation function is enabled.
1868 [Richard Levitte, Bodo Moeller]
1870 *) Finish off removing the remaining LHASH function pointer casts.
1871 There should no longer be any prototype-casting required when using
1872 the LHASH abstraction, and any casts that remain are "bugs". See
1873 the callback types and macros at the head of lhash.h for details
1874 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1877 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1878 If /dev/[u]random devices are not available or do not return enough
1879 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1881 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1882 /etc/entropy will be queried once each in this sequence, quering stops
1883 when enough entropy was collected without querying more sockets.
1886 *) Change the Unix RAND_poll() variant to be able to poll several
1887 random devices, as specified by DEVRANDOM, until a sufficient amount
1888 of data has been collected. We spend at most 10 ms on each file
1889 (select timeout) and read in non-blocking mode. DEVRANDOM now
1890 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1891 (previously it was just the string "/dev/urandom"), so on typical
1892 platforms the 10 ms delay will never occur.
1893 Also separate out the Unix variant to its own file, rand_unix.c.
1894 For VMS, there's a currently-empty rand_vms.c.
1897 *) Move OCSP client related routines to ocsp_cl.c. These
1898 provide utility functions which an application needing
1899 to issue a request to an OCSP responder and analyse the
1900 response will typically need: as opposed to those which an
1901 OCSP responder itself would need which will be added later.
1903 OCSP_request_sign() signs an OCSP request with an API similar
1904 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1905 response. OCSP_response_get1_basic() extracts basic response
1906 from response. OCSP_resp_find_status(): finds and extracts status
1907 information from an OCSP_CERTID structure (which will be created
1908 when the request structure is built). These are built from lower
1909 level functions which work on OCSP_SINGLERESP structures but
1910 wont normally be used unless the application wishes to examine
1911 extensions in the OCSP response for example.
1913 Replace nonce routines with a pair of functions.
1914 OCSP_request_add1_nonce() adds a nonce value and optionally
1915 generates a random value. OCSP_check_nonce() checks the
1916 validity of the nonce in an OCSP response.
1919 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1920 This doesn't copy the supplied OCSP_CERTID and avoids the
1921 need to free up the newly created id. Change return type
1922 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1923 This can then be used to add extensions to the request.
1924 Deleted OCSP_request_new(), since most of its functionality
1925 is now in OCSP_REQUEST_new() (and the case insensitive name
1926 clash) apart from the ability to set the request name which
1927 will be added elsewhere.
1930 *) Update OCSP API. Remove obsolete extensions argument from
1931 various functions. Extensions are now handled using the new
1932 OCSP extension code. New simple OCSP HTTP function which
1933 can be used to send requests and parse the response.
1936 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1937 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1938 uses the special reorder version of SET OF to sort the attributes
1939 and reorder them to match the encoded order. This resolves a long
1940 standing problem: a verify on a PKCS7 structure just after signing
1941 it used to fail because the attribute order did not match the
1942 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1943 it uses the received order. This is necessary to tolerate some broken
1944 software that does not order SET OF. This is handled by encoding
1945 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1946 to produce the required SET OF.
1949 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1950 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1951 files to get correct declarations of the ASN.1 item variables.
1954 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1955 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1956 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1957 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1958 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1959 ASN1_ITEM and no wrapper functions.
1962 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1963 replace the old function pointer based I/O routines. Change most of
1964 the *_d2i_bio() and *_d2i_fp() functions to use these.
1967 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1968 lines, recognice more "algorithms" that can be deselected, and make
1969 it complain about algorithm deselection that isn't recognised.
1972 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1973 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1974 to use new functions. Add NO_ASN1_OLD which can be set to remove
1975 some old style ASN1 functions: this can be used to determine if old
1976 code will still work when these eventually go away.
1979 *) New extension functions for OCSP structures, these follow the
1980 same conventions as certificates and CRLs.
1983 *) New function X509V3_add1_i2d(). This automatically encodes and
1984 adds an extension. Its behaviour can be customised with various
1985 flags to append, replace or delete. Various wrappers added for
1986 certifcates and CRLs.
1989 *) Fix to avoid calling the underlying ASN1 print routine when
1990 an extension cannot be parsed. Correct a typo in the
1991 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1994 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1995 entries for variables.
1998 *) Add functionality to apps/openssl.c for detecting locking
1999 problems: As the program is single-threaded, all we have
2000 to do is register a locking callback using an array for
2001 storing which locks are currently held by the program.
2004 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2005 SSL_get_ex_data_X509_STORE_idx(), which is used in
2006 ssl_verify_cert_chain() and thus can be called at any time
2007 during TLS/SSL handshakes so that thread-safety is essential.
2008 Unfortunately, the ex_data design is not at all suited
2009 for multi-threaded use, so it probably should be abolished.
2012 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2013 [Broadcom, tweaked and integrated by Geoff Thorpe]
2015 *) Move common extension printing code to new function
2016 X509V3_print_extensions(). Reorganise OCSP print routines and
2017 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2020 *) New function X509_signature_print() to remove duplication in some
2024 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2025 set (this was treated exactly the same as SET OF previously). This
2026 is used to reorder the STACK representing the structure to match the
2027 encoding. This will be used to get round a problem where a PKCS7
2028 structure which was signed could not be verified because the STACK
2029 order did not reflect the encoded order.
2032 *) Reimplement the OCSP ASN1 module using the new code.
2035 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2036 for its ASN1 operations. The old style function pointers still exist
2037 for now but they will eventually go away.
2040 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2041 completely replaces the old ASN1 functionality with a table driven
2042 encoder and decoder which interprets an ASN1_ITEM structure describing
2043 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2044 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2045 has also been converted to the new form.
2048 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2049 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2050 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2051 for negative moduli.
2054 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2055 of not touching the result's sign bit.
2058 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2062 *) Changed the LHASH code to use prototypes for callbacks, and created
2063 macros to declare and implement thin (optionally static) functions
2064 that provide type-safety and avoid function pointer casting for the
2065 type-specific callbacks.
2068 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2070 [Veers Staats <staatsvr@asc.hpc.mil>,
2071 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2073 *) Reformat the FAQ so the different questions and answers can be divided
2074 in sections depending on the subject.
2077 *) Have the zlib compression code load ZLIB.DLL dynamically under
2081 *) New function BN_mod_sqrt for computing square roots modulo a prime
2082 (using the probabilistic Tonelli-Shanks algorithm unless
2083 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2084 be handled deterministically).
2085 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2087 *) Make BN_mod_inverse faster by explicitly handling small quotients
2088 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2089 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2092 *) New function BN_kronecker.
2095 *) Fix BN_gcd so that it works on negative inputs; the result is
2096 positive unless both parameters are zero.
2097 Previously something reasonably close to an infinite loop was
2098 possible because numbers could be growing instead of shrinking
2099 in the implementation of Euclid's algorithm.
2102 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2103 sign of the number in question.
2105 Fix BN_is_word(a,w) to work correctly for w == 0.
2107 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2108 because its test if the absolute value of 'a' equals 'w'.
2109 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2110 it exists mostly for use in the implementations of BN_is_zero(),
2111 BN_is_one(), and BN_is_word().
2114 *) New function BN_swap.
2117 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2118 the exponentiation functions are more likely to produce reasonable
2119 results on negative inputs.
2122 *) Change BN_mod_mul so that the result is always non-negative.
2123 Previously, it could be negative if one of the factors was negative;
2124 I don't think anyone really wanted that behaviour.
2127 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2128 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2129 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2130 and add new functions:
2139 BN_mod_lshift1_quick
2143 These functions always generate non-negative results.
2145 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2146 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2148 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2149 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2150 be reduced modulo m.
2151 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2154 The following entry accidentily appeared in the CHANGES file
2155 distributed with OpenSSL 0.9.7. The modifications described in
2156 it do *not* apply to OpenSSL 0.9.7.
2158 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2159 was actually never needed) and in BN_mul(). The removal in BN_mul()
2160 required a small change in bn_mul_part_recursive() and the addition
2161 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2162 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2163 bn_sub_words() and bn_add_words() except they take arrays with
2168 *) In 'openssl passwd', verify passwords read from the terminal
2169 unless the '-salt' option is used (which usually means that
2170 verification would just waste user's time since the resulting
2171 hash is going to be compared with some given password hash)
2172 or the new '-noverify' option is used.
2174 This is an incompatible change, but it does not affect
2175 non-interactive use of 'openssl passwd' (passwords on the command
2176 line, '-stdin' option, '-in ...' option) and thus should not
2180 *) Remove all references to RSAref, since there's no more need for it.
2183 *) Make DSO load along a path given through an environment variable
2184 (SHLIB_PATH) with shl_load().
2187 *) Constify the ENGINE code as a result of BIGNUM constification.
2188 Also constify the RSA code and most things related to it. In a
2189 few places, most notable in the depth of the ASN.1 code, ugly
2190 casts back to non-const were required (to be solved at a later
2194 *) Make it so the openssl application has all engines loaded by default.
2197 *) Constify the BIGNUM routines a little more.
2200 *) Add the following functions:
2202 ENGINE_load_cswift()
2204 ENGINE_load_atalla()
2206 ENGINE_load_builtin_engines()
2208 That way, an application can itself choose if external engines that
2209 are built-in in OpenSSL shall ever be used or not. The benefit is
2210 that applications won't have to be linked with libdl or other dso
2211 libraries unless it's really needed.
2213 Changed 'openssl engine' to load all engines on demand.
2214 Changed the engine header files to avoid the duplication of some
2215 declarations (they differed!).
2218 *) 'openssl engine' can now list capabilities.
2221 *) Better error reporting in 'openssl engine'.
2224 *) Never call load_dh_param(NULL) in s_server.
2227 *) Add engine application. It can currently list engines by name and
2228 identity, and test if they are actually available.
2231 *) Improve RPM specification file by forcing symbolic linking and making
2232 sure the installed documentation is also owned by root.root.
2233 [Damien Miller <djm@mindrot.org>]
2235 *) Give the OpenSSL applications more possibilities to make use of
2236 keys (public as well as private) handled by engines.
2239 *) Add OCSP code that comes from CertCo.
2242 *) Add VMS support for the Rijndael code.
2245 *) Added untested support for Nuron crypto accelerator.
2248 *) Add support for external cryptographic devices. This code was
2249 previously distributed separately as the "engine" branch.
2250 [Geoff Thorpe, Richard Levitte]
2252 *) Rework the filename-translation in the DSO code. It is now possible to
2253 have far greater control over how a "name" is turned into a filename
2254 depending on the operating environment and any oddities about the
2255 different shared library filenames on each system.
2258 *) Support threads on FreeBSD-elf in Configure.
2261 *) Fix for SHA1 assembly problem with MASM: it produces
2262 warnings about corrupt line number information when assembling
2263 with debugging information. This is caused by the overlapping
2265 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2268 NCONF_get_number() has no error checking at all. As a replacement,
2269 NCONF_get_number_e() is defined (_e for "error checking") and is
2270 promoted strongly. The old NCONF_get_number is kept around for
2271 binary backward compatibility.
2272 Make it possible for methods to load from something other than a BIO,
2273 by providing a function pointer that is given a name instead of a BIO.
2274 For example, this could be used to load configuration data from an
2278 *) Fix for non blocking accept BIOs. Added new I/O special reason
2279 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2280 with non blocking I/O was not possible because no retry code was
2281 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2285 *) Added the beginnings of Rijndael support.
2288 *) Fix for bug in DirectoryString mask setting. Add support for
2289 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2290 to allow certificate printing to more controllable, additional
2291 'certopt' option to 'x509' to allow new printing options to be
2295 *) Clean old EAY MD5 hack from e_os.h.
2298 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2300 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2301 memory from it's contents. This is done with a counter that will
2302 place alternating values in each byte. This can be used to solve
2303 two issues: 1) the removal of calls to memset() by highly optimizing
2304 compilers, and 2) cleansing with other values than 0, since those can
2305 be read through on certain media, for example a swap space on disk.
2308 *) Bugfix: client side session caching did not work with external caching,
2309 because the session->cipher setting was not restored when reloading
2310 from the external cache. This problem was masked, when
2311 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2312 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2315 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2316 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2317 [Zeev Lieber <zeev-l@yahoo.com>]
2319 *) Undo an undocumented change introduced in 0.9.6e which caused
2320 repeated calls to OpenSSL_add_all_ciphers() and
2321 OpenSSL_add_all_digests() to be ignored, even after calling
2325 *) Change the default configuration reader to deal with last line not
2326 being properly terminated.
2329 *) Change X509_NAME_cmp() so it applies the special rules on handling
2330 DN values that are of type PrintableString, as well as RDNs of type
2331 emailAddress where the value has the type ia5String.
2332 [stefank@valicert.com via Richard Levitte]
2334 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2335 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2336 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2337 the bitwise-OR of the two for use by the majority of applications
2338 wanting this behaviour, and update the docs. The documented
2339 behaviour and actual behaviour were inconsistent and had been
2340 changing anyway, so this is more a bug-fix than a behavioural
2342 [Geoff Thorpe, diagnosed by Nadav Har'El]
2344 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2345 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2348 *) Fix initialization code race conditions in
2349 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2350 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2351 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2352 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2353 ssl2_get_cipher_by_char(),
2354 ssl3_get_cipher_by_char().
2355 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2357 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2358 the cached sessions are flushed, as the remove_cb() might use ex_data
2359 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2360 (see [openssl.org #212]).
2361 [Geoff Thorpe, Lutz Jaenicke]
2363 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2364 length, instead of the encoding length to d2i_ASN1_OBJECT.
2367 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2369 *) [In 0.9.6g-engine release:]
2370 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2371 [Lynn Gazis <lgazis@rainbow.com>]
2373 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2375 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2376 and get fix the header length calculation.
2377 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2378 Alon Kantor <alonk@checkpoint.com> (and others),
2381 *) Use proper error handling instead of 'assertions' in buffer
2382 overflow checks added in 0.9.6e. This prevents DoS (the
2383 assertions could call abort()).
2384 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2386 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2388 *) Add various sanity checks to asn1_get_length() to reject
2389 the ASN1 length bytes if they exceed sizeof(long), will appear
2390 negative or the content length exceeds the length of the
2392 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2394 *) Fix cipher selection routines: ciphers without encryption had no flags
2395 for the cipher strength set and where therefore not handled correctly
2396 by the selection routines (PR #130).
2399 *) Fix EVP_dsa_sha macro.
2403 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2404 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2405 that was added in OpenSSL 0.9.6d.
2407 As the countermeasure turned out to be incompatible with some
2408 broken SSL implementations, the new option is part of SSL_OP_ALL.
2409 SSL_OP_ALL is usually employed when compatibility with weird SSL
2410 implementations is desired (e.g. '-bugs' option to 's_client' and
2411 's_server'), so the new option is automatically set in many
2415 *) Changes in security patch:
2417 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2418 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2419 Air Force Materiel Command, USAF, under agreement number
2422 *) Add various sanity checks to asn1_get_length() to reject
2423 the ASN1 length bytes if they exceed sizeof(long), will appear
2424 negative or the content length exceeds the length of the
2425 supplied buffer. (CAN-2002-0659)
2426 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2428 *) Assertions for various potential buffer overflows, not known to
2430 [Ben Laurie (CHATS)]
2432 *) Various temporary buffers to hold ASCII versions of integers were
2433 too small for 64 bit platforms. (CAN-2002-0655)
2434 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2436 *) Remote buffer overflow in SSL3 protocol - an attacker could
2437 supply an oversized session ID to a client. (CAN-2002-0656)
2438 [Ben Laurie (CHATS)]
2440 *) Remote buffer overflow in SSL2 protocol - an attacker could
2441 supply an oversized client master key. (CAN-2002-0656)
2442 [Ben Laurie (CHATS)]
2444 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2446 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2447 encoded as NULL) with id-dsa-with-sha1.
2448 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2450 *) Check various X509_...() return values in apps/req.c.
2451 [Nils Larsch <nla@trustcenter.de>]
2453 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2454 an end-of-file condition would erronously be flagged, when the CRLF
2455 was just at the end of a processed block. The bug was discovered when
2456 processing data through a buffering memory BIO handing the data to a
2457 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2458 <ptsekov@syntrex.com> and Nedelcho Stanev.
2461 *) Implement a countermeasure against a vulnerability recently found
2462 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2463 before application data chunks to avoid the use of known IVs
2464 with data potentially chosen by the attacker.
2467 *) Fix length checks in ssl3_get_client_hello().
2470 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2471 to prevent ssl3_read_internal() from incorrectly assuming that
2472 ssl3_read_bytes() found application data while handshake
2473 processing was enabled when in fact s->s3->in_read_app_data was
2474 merely automatically cleared during the initial handshake.
2475 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2477 *) Fix object definitions for Private and Enterprise: they were not
2478 recognized in their shortname (=lowercase) representation. Extend
2479 obj_dat.pl to issue an error when using undefined keywords instead
2480 of silently ignoring the problem (Svenning Sorensen
2481 <sss@sss.dnsalias.net>).
2484 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2485 generators, i.e. generators other than 2 and 5. (Previously, the
2486 code did not properly initialise the 'add' and 'rem' values to
2487 BN_generate_prime().)
2489 In the new general case, we do not insist that 'generator' is
2490 actually a primitive root: This requirement is rather pointless;
2491 a generator of the order-q subgroup is just as good, if not
2495 *) Map new X509 verification errors to alerts. Discovered and submitted by
2496 Tom Wu <tom@arcot.com>.
2499 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2500 returning non-zero before the data has been completely received
2501 when using non-blocking I/O.
2502 [Bodo Moeller; problem pointed out by John Hughes]
2504 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2505 [Ben Laurie, Lutz Jaenicke]
2507 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2508 Yoram Zahavi <YoramZ@gilian.com>).
2511 *) Add information about CygWin 1.3 and on, and preserve proper
2512 configuration for the versions before that.
2513 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2515 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2516 check whether we deal with a copy of a session and do not delete from
2517 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2518 <izhar@checkpoint.com>.
2521 *) Do not store session data into the internal session cache, if it
2522 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2523 flag is set). Proposed by Aslam <aslam@funk.com>.
2526 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2530 *) [In 0.9.6d-engine release:]
2531 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2532 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2534 *) Add the configuration target linux-s390x.
2535 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2537 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2538 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2539 variable as an indication that a ClientHello message has been
2540 received. As the flag value will be lost between multiple
2541 invocations of ssl3_accept when using non-blocking I/O, the
2542 function may not be aware that a handshake has actually taken
2543 place, thus preventing a new session from being added to the
2546 To avoid this problem, we now set s->new_session to 2 instead of
2547 using a local variable.
2548 [Lutz Jaenicke, Bodo Moeller]
2550 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2551 if the SSL_R_LENGTH_MISMATCH error is detected.
2552 [Geoff Thorpe, Bodo Moeller]
2554 *) New 'shared_ldflag' column in Configure platform table.
2557 *) Fix EVP_CIPHER_mode macro.
2558 ["Dan S. Camper" <dan@bti.net>]
2560 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2561 type, we must throw them away by setting rr->length to 0.
2562 [D P Chang <dpc@qualys.com>]
2564 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2566 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2567 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2568 worked incorrectly for those cases where range = 10..._2 and
2569 3*range is two bits longer than range.)
2572 *) Only add signing time to PKCS7 structures if it is not already
2576 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2577 OBJ_ld_ce should be OBJ_id_ce.
2578 Also some ip-pda OIDs in crypto/objects/objects.txt were
2579 incorrect (cf. RFC 3039).
2580 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2582 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2583 returns early because it has nothing to do.
2584 [Andy Schneider <andy.schneider@bjss.co.uk>]
2586 *) [In 0.9.6c-engine release:]
2587 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2588 [Andy Schneider <andy.schneider@bjss.co.uk>]
2590 *) [In 0.9.6c-engine release:]
2591 Add support for Cryptographic Appliance's keyserver technology.
2592 (Use engine 'keyclient')
2593 [Cryptographic Appliances and Geoff Thorpe]
2595 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2596 is called via tools/c89.sh because arguments have to be
2597 rearranged (all '-L' options must appear before the first object
2599 [Richard Shapiro <rshapiro@abinitio.com>]
2601 *) [In 0.9.6c-engine release:]
2602 Add support for Broadcom crypto accelerator cards, backported
2604 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2606 *) [In 0.9.6c-engine release:]
2607 Add support for SureWare crypto accelerator cards from
2608 Baltimore Technologies. (Use engine 'sureware')
2609 [Baltimore Technologies and Mark Cox]
2611 *) [In 0.9.6c-engine release:]
2612 Add support for crypto accelerator cards from Accelerated
2613 Encryption Processing, www.aep.ie. (Use engine 'aep')
2614 [AEP Inc. and Mark Cox]
2616 *) Add a configuration entry for gcc on UnixWare.
2617 [Gary Benson <gbenson@redhat.com>]
2619 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2620 messages are stored in a single piece (fixed-length part and
2621 variable-length part combined) and fix various bugs found on the way.
2624 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2625 instead. BIO_gethostbyname() does not know what timeouts are
2626 appropriate, so entries would stay in cache even when they have
2628 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2630 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2631 faced with a pathologically small ClientHello fragment that does
2632 not contain client_version: Instead of aborting with an error,
2633 simply choose the highest available protocol version (i.e.,
2634 TLS 1.0 unless it is disabled). In practice, ClientHello
2635 messages are never sent like this, but this change gives us
2636 strictly correct behaviour at least for TLS.
2639 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2640 never resets s->method to s->ctx->method when called from within
2641 one of the SSL handshake functions.
2642 [Bodo Moeller; problem pointed out by Niko Baric]
2644 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2645 (sent using the client's version number) if client_version is
2646 smaller than the protocol version in use. Also change
2647 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2648 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2649 the client will at least see that alert.
2652 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2656 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2657 client receives HelloRequest while in a handshake.
2658 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2660 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2661 should end in 'break', not 'goto end' which circuments various
2662 cleanups done in state SSL_ST_OK. But session related stuff
2663 must be disabled for SSL_ST_OK in the case that we just sent a
2666 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2667 before just sending a HelloRequest.
2668 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2670 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2671 reveal whether illegal block cipher padding was found or a MAC
2672 verification error occured. (Neither SSLerr() codes nor alerts
2673 are directly visible to potential attackers, but the information
2674 may leak via logfiles.)
2676 Similar changes are not required for the SSL 2.0 implementation
2677 because the number of padding bytes is sent in clear for SSL 2.0,
2678 and the extra bytes are just ignored. However ssl/s2_pkt.c
2679 failed to verify that the purported number of padding bytes is in
2683 *) Add OpenUNIX-8 support including shared libraries
2684 (Boyd Lynn Gerber <gerberb@zenez.com>).
2687 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2688 'wristwatch attack' using huge encoding parameters (cf.
2689 James H. Manger's CRYPTO 2001 paper). Note that the
2690 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2691 encoding parameters and hence was not vulnerable.
2694 *) BN_sqr() bug fix.
2695 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2697 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2698 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2699 followed by modular reduction.
2700 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2702 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2703 equivalent based on BN_pseudo_rand() instead of BN_rand().
2706 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2707 This function was broken, as the check for a new client hello message
2708 to handle SGC did not allow these large messages.
2709 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2712 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2715 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2716 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2719 *) Rework the configuration and shared library support for Tru64 Unix.
2720 The configuration part makes use of modern compiler features and
2721 still retains old compiler behavior for those that run older versions
2722 of the OS. The shared library support part includes a variant that
2723 uses the RPATH feature, and is available through the special
2724 configuration target "alpha-cc-rpath", which will never be selected
2726 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2728 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2729 with the same message size as in ssl3_get_certificate_request().
2730 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2731 messages might inadvertently be reject as too long.
2732 [Petr Lampa <lampa@fee.vutbr.cz>]
2734 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2737 *) Modified SSL library such that the verify_callback that has been set
2738 specificly for an SSL object with SSL_set_verify() is actually being
2739 used. Before the change, a verify_callback set with this function was
2740 ignored and the verify_callback() set in the SSL_CTX at the time of
2741 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2742 to allow the necessary settings.
2745 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2746 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2747 done automatically (in contradiction to the requirements of the C
2748 standard). This made problems when used from OpenSSH.
2751 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2752 dh->length and always used
2754 BN_rand_range(priv_key, dh->p).
2756 BN_rand_range() is not necessary for Diffie-Hellman, and this
2757 specific range makes Diffie-Hellman unnecessarily inefficient if
2758 dh->length (recommended exponent length) is much smaller than the
2759 length of dh->p. We could use BN_rand_range() if the order of
2760 the subgroup was stored in the DH structure, but we only have
2765 BN_rand(priv_key, l, ...)
2767 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2773 RSA_eay_public_encrypt
2774 RSA_eay_private_decrypt
2775 RSA_eay_private_encrypt (signing)
2776 RSA_eay_public_decrypt (signature verification)
2778 (default implementations for RSA_public_encrypt,
2779 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2780 always reject numbers >= n.
2783 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2784 to synchronize access to 'locking_thread'. This is necessary on
2785 systems where access to 'locking_thread' (an 'unsigned long'
2786 variable) is not atomic.
2789 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2790 *before* setting the 'crypto_lock_rand' flag. The previous code had
2791 a race condition if 0 is a valid thread ID.
2792 [Travis Vitek <vitek@roguewave.com>]
2794 *) Add support for shared libraries under Irix.
2795 [Albert Chin-A-Young <china@thewrittenword.com>]
2797 *) Add configuration option to build on Linux on both big-endian and
2799 [Ralf Baechle <ralf@uni-koblenz.de>]
2801 *) Add the possibility to create shared libraries on HP-UX.
2804 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2806 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2807 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2808 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2809 PRNG state recovery was possible based on the output of
2810 one PRNG request appropriately sized to gain knowledge on
2811 'md' followed by enough consecutive 1-byte PRNG requests
2812 to traverse all of 'state'.
2814 1. When updating 'md_local' (the current thread's copy of 'md')
2815 during PRNG output generation, hash all of the previous
2816 'md_local' value, not just the half used for PRNG output.
2818 2. Make the number of bytes from 'state' included into the hash
2819 independent from the number of PRNG bytes requested.
2821 The first measure alone would be sufficient to avoid
2822 Markku-Juhani's attack. (Actually it had never occurred
2823 to me that the half of 'md_local' used for chaining was the
2824 half from which PRNG output bytes were taken -- I had always
2825 assumed that the secret half would be used.) The second
2826 measure makes sure that additional data from 'state' is never
2827 mixed into 'md_local' in small portions; this heuristically
2828 further strengthens the PRNG.
2831 *) Fix crypto/bn/asm/mips3.s.
2834 *) When only the key is given to "enc", the IV is undefined. Print out
2835 an error message in this case.
2838 *) Handle special case when X509_NAME is empty in X509 printing routines.
2841 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2842 positive and less than q.
2845 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2846 used: it isn't thread safe and the add_lock_callback should handle
2848 [Paul Rose <Paul.Rose@bridge.com>]
2850 *) Verify that incoming data obeys the block size in
2851 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2855 [Ulf Möller, Bodo Möller]
2857 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2858 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2859 when fixing the server behaviour for backwards-compatible 'client
2860 hello' messages. (Note that the attack is impractical against
2861 SSL 3.0 and TLS 1.0 anyway because length and version checking
2862 means that the probability of guessing a valid ciphertext is
2863 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2866 Before 0.9.5, the countermeasure (hide the error by generating a
2867 random 'decryption result') did not work properly because
2868 ERR_clear_error() was missing, meaning that SSL_get_error() would
2869 detect the supposedly ignored error.
2871 Both problems are now fixed.
2874 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2875 (previously it was 1024).
2878 *) Fix for compatibility mode trust settings: ignore trust settings
2879 unless some valid trust or reject settings are present.
2882 *) Fix for blowfish EVP: its a variable length cipher.
2885 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2886 parameters in DSA public key structures and return an error in the
2887 DSA routines if parameters are absent.
2890 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2891 in the current directory if neither $RANDFILE nor $HOME was set.
2892 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2893 caused some confusion to Windows users who haven't defined $HOME.
2894 Thus RAND_file_name() is changed again: e_os.h can define a
2895 DEFAULT_HOME, which will be used if $HOME is not set.
2896 For Windows, we use "C:"; on other platforms, we still require
2897 environment variables.
2899 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2900 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2901 having multiple threads call RAND_poll() concurrently.
2904 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2905 combination of a flag and a thread ID variable.
2906 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2907 flag), *other* threads can enter ssleay_add_bytes without obeying
2908 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2909 that they do not hold after the first thread unsets add_do_not_lock).
2912 *) Change bctest again: '-x' expressions are not available in all
2916 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2918 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2919 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2921 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2922 the default extension for executables, if any. Also, make the perl
2923 scripts that use symlink() to test if it really exists and use "cp"
2924 if it doesn't. All this made OpenSSL compilable and installable in
2928 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2929 If SEQUENCE is length is indefinite just set c->slen to the total
2930 amount of data available.
2931 [Steve Henson, reported by shige@FreeBSD.org]
2932 [This change does not apply to 0.9.7.]
2934 *) Change bctest to avoid here-documents inside command substitution
2935 (workaround for FreeBSD /bin/sh bug).
2936 For compatibility with Ultrix, avoid shell functions (introduced
2937 in the bctest version that searches along $PATH).
2940 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2941 with des_encrypt() defined on some operating systems, like Solaris
2945 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2946 On the Importance of Eliminating Errors in Cryptographic
2947 Computations, J. Cryptology 14 (2001) 2, 101-119,
2948 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2951 *) MIPS assembler BIGNUM division bug fix.
2954 *) Disabled incorrect Alpha assembler code.
2957 *) Fix PKCS#7 decode routines so they correctly update the length
2958 after reading an EOC for the EXPLICIT tag.
2960 [This change does not apply to 0.9.7.]
2962 *) Fix bug in PKCS#12 key generation routines. This was triggered
2963 if a 3DES key was generated with a 0 initial byte. Include
2964 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2965 (but broken) behaviour.
2968 *) Enhance bctest to search for a working bc along $PATH and print
2970 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2972 *) Fix memory leaks in err.c: free err_data string if necessary;
2973 don't write to the wrong index in ERR_set_error_data.
2976 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2980 *) Replace rdtsc with _emit statements for VC++ version 5.
2981 [Jeremy Cooper <jeremy@baymoo.org>]
2983 *) Make it possible to reuse SSLv2 sessions.
2986 *) In copy_email() check for >= 0 as a return value for
2987 X509_NAME_get_index_by_NID() since 0 is a valid index.
2988 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2990 *) Avoid coredump with unsupported or invalid public keys by checking if
2991 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2992 PKCS7_verify() fails with non detached data.
2995 *) Don't use getenv in library functions when run as setuid/setgid.
2996 New function OPENSSL_issetugid().
2999 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3000 due to incorrect handling of multi-threading:
3002 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3004 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3006 3. Count how many times MemCheck_off() has been called so that
3007 nested use can be treated correctly. This also avoids
3008 inband-signalling in the previous code (which relied on the
3009 assumption that thread ID 0 is impossible).
3012 *) Add "-rand" option also to s_client and s_server.
3015 *) Fix CPU detection on Irix 6.x.
3016 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3017 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3019 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3022 [This change does not apply to 0.9.7.]
3024 *) Use the cached encoding of an X509_NAME structure rather than
3025 copying it. This is apparently the reason for the libsafe "errors"
3026 but the code is actually correct.
3029 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3030 Bleichenbacher's DSA attack.
3031 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3032 to be set and top=0 forces the highest bit to be set; top=-1 is new
3033 and leaves the highest bit random.
3034 [Ulf Moeller, Bodo Moeller]
3036 *) In the NCONF_...-based implementations for CONF_... queries
3037 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3038 a temporary CONF structure with the data component set to NULL
3039 (which gives segmentation faults in lh_retrieve).
3040 Instead, use NULL for the CONF pointer in CONF_get_string and
3041 CONF_get_number (which may use environment variables) and directly
3042 return NULL from CONF_get_section.
3045 *) Fix potential buffer overrun for EBCDIC.
3048 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3049 keyUsage if basicConstraints absent for a CA.
3052 *) Make SMIME_write_PKCS7() write mail header values with a format that
3053 is more generally accepted (no spaces before the semicolon), since
3054 some programs can't parse those values properly otherwise. Also make
3055 sure BIO's that break lines after each write do not create invalid
3059 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3060 macros previously used would not encode an empty SEQUENCE OF
3061 and break the signature.
3063 [This change does not apply to 0.9.7.]
3065 *) Zero the premaster secret after deriving the master secret in
3069 *) Add some EVP_add_digest_alias registrations (as found in
3070 OpenSSL_add_all_digests()) to SSL_library_init()
3071 aka OpenSSL_add_ssl_algorithms(). This provides improved
3072 compatibility with peers using X.509 certificates
3073 with unconventional AlgorithmIdentifier OIDs.
3076 *) Fix for Irix with NO_ASM.
3077 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3079 *) ./config script fixes.
3080 [Ulf Moeller, Richard Levitte]
3082 *) Fix 'openssl passwd -1'.
3085 *) Change PKCS12_key_gen_asc() so it can cope with non null
3086 terminated strings whose length is passed in the passlen
3087 parameter, for example from PEM callbacks. This was done
3088 by adding an extra length parameter to asc2uni().
3089 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3091 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3092 call failed, free the DSA structure.
3095 *) Fix to uni2asc() to cope with zero length Unicode strings.
3096 These are present in some PKCS#12 files.
3099 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3100 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3101 when writing a 32767 byte record.
3102 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3104 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3105 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3107 (RSA objects have a reference count access to which is protected
3108 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3109 so they are meant to be shared between threads.)
3110 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3111 "Reddie, Steven" <Steven.Reddie@ca.com>]
3113 *) Fix a deadlock in CRYPTO_mem_leaks().
3116 *) Use better test patterns in bntest.
3119 *) rand_win.c fix for Borland C.
3122 *) BN_rshift bugfix for n == 0.
3125 *) Add a 'bctest' script that checks for some known 'bc' bugs
3126 so that 'make test' does not abort just because 'bc' is broken.
3129 *) Store verify_result within SSL_SESSION also for client side to
3130 avoid potential security hole. (Re-used sessions on the client side
3131 always resulted in verify_result==X509_V_OK, not using the original
3132 result of the server certificate verification.)
3135 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3136 SSL3_RT_APPLICATION_DATA, return 0.
3137 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3141 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3142 releases, have been re-implemented by renaming the previous
3143 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3144 and ssl3_read_internal, respectively, and adding 'peek' parameters
3145 to them. The new ssl[23]_{read,peek} functions are calls to
3146 ssl[23]_read_internal with the 'peek' flag set appropriately.
3147 A 'peek' parameter has also been added to ssl3_read_bytes, which
3148 does the actual work for ssl3_read_internal.
3151 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3152 the method-specific "init()" handler. Also clean up ex_data after
3153 calling the method-specific "finish()" handler. Previously, this was
3154 happening the other way round.
3157 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3158 The previous value, 12, was not always sufficient for BN_mod_exp().
3161 *) Make sure that shared libraries get the internal name engine with
3162 the full version number and not just 0. This should mark the
3163 shared libraries as not backward compatible. Of course, this should
3164 be changed again when we can guarantee backward binary compatibility.
3167 *) Fix typo in get_cert_by_subject() in by_dir.c
3168 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3170 *) Rework the system to generate shared libraries:
3172 - Make note of the expected extension for the shared libraries and
3173 if there is a need for symbolic links from for example libcrypto.so.0
3174 to libcrypto.so.0.9.7. There is extended info in Configure for
3177 - Make as few rebuilds of the shared libraries as possible.
3179 - Still avoid linking the OpenSSL programs with the shared libraries.
3181 - When installing, install the shared libraries separately from the
3185 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3187 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3188 and not in SSL_clear because the latter is also used by the
3189 accept/connect functions; previously, the settings made by
3190 SSL_set_read_ahead would be lost during the handshake.
3191 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3193 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3194 Previously, it would create entries for disableed algorithms no
3198 *) Added several new manual pages for SSL_* function.
3201 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3203 *) In ssl23_get_client_hello, generate an error message when faced
3204 with an initial SSL 3.0/TLS record that is too small to contain the
3205 first two bytes of the ClientHello message, i.e. client_version.
3206 (Note that this is a pathologic case that probably has never happened
3207 in real life.) The previous approach was to use the version number
3208 from the record header as a substitute; but our protocol choice
3209 should not depend on that one because it is not authenticated
3210 by the Finished messages.
3213 *) More robust randomness gathering functions for Windows.
3214 [Jeffrey Altman <jaltman@columbia.edu>]
3216 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3217 not set then we don't setup the error code for issuer check errors
3218 to avoid possibly overwriting other errors which the callback does
3219 handle. If an application does set the flag then we assume it knows
3220 what it is doing and can handle the new informational codes
3224 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3225 a general "ANY" type, as such it should be able to decode anything
3226 including tagged types. However it didn't check the class so it would
3227 wrongly interpret tagged types in the same way as their universal
3228 counterpart and unknown types were just rejected. Changed so that the
3229 tagged and unknown types are handled in the same way as a SEQUENCE:
3230 that is the encoding is stored intact. There is also a new type
3231 "V_ASN1_OTHER" which is used when the class is not universal, in this
3232 case we have no idea what the actual type is so we just lump them all
3236 *) On VMS, stdout may very well lead to a file that is written to
3237 in a record-oriented fashion. That means that every write() will
3238 write a separate record, which will be read separately by the
3239 programs trying to read from it. This can be very confusing.
3241 The solution is to put a BIO filter in the way that will buffer
3242 text until a linefeed is reached, and then write everything a
3243 line at a time, so every record written will be an actual line,
3244 not chunks of lines and not (usually doesn't happen, but I've
3245 seen it once) several lines in one record. BIO_f_linebuffer() is
3248 Currently, it's a VMS-only method, because that's where it has
3249 been tested well enough.
3252 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3253 it can return incorrect results.
3254 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3255 but it was in 0.9.6-beta[12].)
3258 *) Disable the check for content being present when verifying detached
3259 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3260 include zero length content when signing messages.
3263 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3264 BIO_ctrl (for BIO pairs).
3267 *) Add DSO method for VMS.
3270 *) Bug fix: Montgomery multiplication could produce results with the
3274 *) Add RPM specification openssl.spec and modify it to build three
3275 packages. The default package contains applications, application
3276 documentation and run-time libraries. The devel package contains
3277 include files, static libraries and function documentation. The
3278 doc package contains the contents of the doc directory. The original
3279 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3282 *) Add a large number of documentation files for many SSL routines.
3283 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3285 *) Add a configuration entry for Sony News 4.
3286 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3288 *) Don't set the two most significant bits to one when generating a
3289 random number < q in the DSA library.
3292 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3293 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3294 the underlying transport is blocking) if a handshake took place.
3295 (The default behaviour is needed by applications such as s_client
3296 and s_server that use select() to determine when to use SSL_read;
3297 but for applications that know in advance when to expect data, it
3298 just makes things more complicated.)
3301 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3305 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3306 work better on such systems.
3307 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3309 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3310 Update PKCS12_parse() so it copies the friendlyName and the
3311 keyid to the certificates aux info.
3314 *) Fix bug in PKCS7_verify() which caused an infinite loop
3315 if there was more than one signature.
3316 [Sven Uszpelkat <su@celocom.de>]
3318 *) Major change in util/mkdef.pl to include extra information
3319 about each symbol, as well as presentig variables as well
3320 as functions. This change means that there's n more need
3321 to rebuild the .num files when some algorithms are excluded.
3324 *) Allow the verify time to be set by an application,
3325 rather than always using the current time.
3328 *) Phase 2 verify code reorganisation. The certificate
3329 verify code now looks up an issuer certificate by a
3330 number of criteria: subject name, authority key id
3331 and key usage. It also verifies self signed certificates
3332 by the same criteria. The main comparison function is
3333 X509_check_issued() which performs these checks.
3335 Lot of changes were necessary in order to support this
3336 without completely rewriting the lookup code.
3338 Authority and subject key identifier are now cached.
3340 The LHASH 'certs' is X509_STORE has now been replaced
3341 by a STACK_OF(X509_OBJECT). This is mainly because an
3342 LHASH can't store or retrieve multiple objects with
3343 the same hash value.
3345 As a result various functions (which were all internal
3346 use only) have changed to handle the new X509_STORE
3347 structure. This will break anything that messed round
3348 with X509_STORE internally.
3350 The functions X509_STORE_add_cert() now checks for an
3351 exact match, rather than just subject name.
3353 The X509_STORE API doesn't directly support the retrieval
3354 of multiple certificates matching a given criteria, however
3355 this can be worked round by performing a lookup first
3356 (which will fill the cache with candidate certificates)
3357 and then examining the cache for matches. This is probably
3358 the best we can do without throwing out X509_LOOKUP
3359 entirely (maybe later...).
3361 The X509_VERIFY_CTX structure has been enhanced considerably.
3363 All certificate lookup operations now go via a get_issuer()
3364 callback. Although this currently uses an X509_STORE it
3365 can be replaced by custom lookups. This is a simple way
3366 to bypass the X509_STORE hackery necessary to make this
3367 work and makes it possible to use more efficient techniques
3368 in future. A very simple version which uses a simple
3369 STACK for its trusted certificate store is also provided
3370 using X509_STORE_CTX_trusted_stack().
3372 The verify_cb() and verify() callbacks now have equivalents
3373 in the X509_STORE_CTX structure.
3375 X509_STORE_CTX also has a 'flags' field which can be used
3376 to customise the verify behaviour.
3379 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3380 excludes S/MIME capabilities.
3383 *) When a certificate request is read in keep a copy of the
3384 original encoding of the signed data and use it when outputing
3385 again. Signatures then use the original encoding rather than
3386 a decoded, encoded version which may cause problems if the
3387 request is improperly encoded.
3390 *) For consistency with other BIO_puts implementations, call
3391 buffer_write(b, ...) directly in buffer_puts instead of calling
3394 In BIO_puts, increment b->num_write as in BIO_write.
3395 [Peter.Sylvester@EdelWeb.fr]
3397 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3398 BN_zero, we may not return a BIGNUM with an array consisting of
3402 *) Avoid calling abort() from within the library when problems are
3403 detected, except if preprocessor symbols have been defined
3404 (such as REF_CHECK, BN_DEBUG etc.).
3407 *) New openssl application 'rsautl'. This utility can be
3408 used for low level RSA operations. DER public key
3409 BIO/fp routines also added.
3412 *) New Configure entry and patches for compiling on QNX 4.
3413 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3415 *) A demo state-machine implementation was sponsored by
3416 Nuron (http://www.nuron.com/) and is now available in
3417 demos/state_machine.
3420 *) New options added to the 'dgst' utility for signature
3421 generation and verification.
3424 *) Unrecognized PKCS#7 content types are now handled via a
3425 catch all ASN1_TYPE structure. This allows unsupported
3426 types to be stored as a "blob" and an application can
3427 encode and decode it manually.
3430 *) Fix various signed/unsigned issues to make a_strex.c
3432 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3434 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3435 length if passed a buffer. ASN1_INTEGER_to_BN failed
3436 if passed a NULL BN and its argument was negative.
3437 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3439 *) Modification to PKCS#7 encoding routines to output definite
3440 length encoding. Since currently the whole structures are in
3441 memory there's not real point in using indefinite length
3442 constructed encoding. However if OpenSSL is compiled with
3443 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3446 *) Added BIO_vprintf() and BIO_vsnprintf().
3449 *) Added more prefixes to parse for in the the strings written
3450 through a logging bio, to cover all the levels that are available
3451 through syslog. The prefixes are now:
3453 PANIC, EMERG, EMR => LOG_EMERG
3454 ALERT, ALR => LOG_ALERT
3455 CRIT, CRI => LOG_CRIT
3456 ERROR, ERR => LOG_ERR
3457 WARNING, WARN, WAR => LOG_WARNING
3458 NOTICE, NOTE, NOT => LOG_NOTICE
3459 INFO, INF => LOG_INFO
3460 DEBUG, DBG => LOG_DEBUG
3462 and as before, if none of those prefixes are present at the
3463 beginning of the string, LOG_ERR is chosen.
3465 On Win32, the LOG_* levels are mapped according to this:
3467 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3468 LOG_WARNING => EVENTLOG_WARNING_TYPE
3469 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3473 *) Made it possible to reconfigure with just the configuration
3474 argument "reconf" or "reconfigure". The command line arguments
3475 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3476 and are retrieved from there when reconfiguring.
3480 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3482 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3485 *) The obj_dat.pl script was messing up the sorting of object
3486 names. The reason was that it compared the quoted version
3487 of strings as a result "OCSP" > "OCSP Signing" because
3488 " > SPACE. Changed script to store unquoted versions of
3489 names and add quotes on output. It was also omitting some
3490 names from the lookup table if they were given a default
3491 value (that is if SN is missing it is given the same
3492 value as LN and vice versa), these are now added on the
3493 grounds that if an object has a name we should be able to
3494 look it up. Finally added warning output when duplicate
3495 short or long names are found.
3498 *) Changes needed for Tandem NSK.
3499 [Scott Uroff <scott@xypro.com>]
3501 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3502 RSA_padding_check_SSLv23(), special padding was never detected
3503 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3504 version rollback attacks was not effective.
3506 In s23_clnt.c, don't use special rollback-attack detection padding
3507 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3508 client; similarly, in s23_srvr.c, don't do the rollback check if
3509 SSL 2.0 is the only protocol enabled in the server.
3512 *) Make it possible to get hexdumps of unprintable data with 'openssl
3513 asn1parse'. By implication, the functions ASN1_parse_dump() and
3514 BIO_dump_indent() are added.
3517 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3518 these print out strings and name structures based on various
3519 flags including RFC2253 support and proper handling of
3520 multibyte characters. Added options to the 'x509' utility
3521 to allow the various flags to be set.
3524 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3525 Also change the functions X509_cmp_current_time() and
3526 X509_gmtime_adj() work with an ASN1_TIME structure,
3527 this will enable certificates using GeneralizedTime in validity
3528 dates to be checked.
3531 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3532 negative public key encodings) on by default,
3533 NO_NEG_PUBKEY_BUG can be set to disable it.
3536 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3537 content octets. An i2c_ASN1_OBJECT is unnecessary because
3538 the encoding can be trivially obtained from the structure.
3541 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3542 not read locks (CRYPTO_r_[un]lock).
3545 *) A first attempt at creating official support for shared
3546 libraries through configuration. I've kept it so the
3547 default is static libraries only, and the OpenSSL programs
3548 are always statically linked for now, but there are
3549 preparations for dynamic linking in place.
3550 This has been tested on Linux and Tru64.
3553 *) Randomness polling function for Win9x, as described in:
3554 Peter Gutmann, Software Generation of Practically Strong
3558 *) Fix so PRNG is seeded in req if using an already existing
3562 *) New options to smime application. -inform and -outform
3563 allow alternative formats for the S/MIME message including
3564 PEM and DER. The -content option allows the content to be
3565 specified separately. This should allow things like Netscape
3566 form signing output easier to verify.
3569 *) Fix the ASN1 encoding of tags using the 'long form'.
3572 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3573 STRING types. These convert content octets to and from the
3574 underlying type. The actual tag and length octets are
3575 already assumed to have been read in and checked. These
3576 are needed because all other string types have virtually
3577 identical handling apart from the tag. By having versions
3578 of the ASN1 functions that just operate on content octets
3579 IMPLICIT tagging can be handled properly. It also allows
3580 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3581 and ASN1_INTEGER are identical apart from the tag.
3584 *) Change the handling of OID objects as follows:
3586 - New object identifiers are inserted in objects.txt, following
3587 the syntax given in objects.README.
3588 - objects.pl is used to process obj_mac.num and create a new
3590 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3593 This is currently kind of a hack, and the perl code in objects.pl
3594 isn't very elegant, but it works as I intended. The simplest way
3595 to check that it worked correctly is to look in obj_dat.h and
3596 check the array nid_objs and make sure the objects haven't moved
3597 around (this is important!). Additions are OK, as well as
3598 consistent name changes.
3601 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3604 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3605 The given file adds to whatever has already been seeded into the
3606 random pool through the RANDFILE configuration file option or
3607 environment variable, or the default random state file.
3610 *) mkstack.pl now sorts each macro group into lexical order.
3611 Previously the output order depended on the order the files
3612 appeared in the directory, resulting in needless rewriting
3616 *) Patches to make OpenSSL compile under Win32 again. Mostly
3617 work arounds for the VC++ problem that it treats func() as
3618 func(void). Also stripped out the parts of mkdef.pl that
3619 added extra typesafe functions: these no longer exist.
3622 *) Reorganisation of the stack code. The macros are now all
3623 collected in safestack.h . Each macro is defined in terms of
3624 a "stack macro" of the form SKM_<name>(type, a, b). The
3625 DEBUG_SAFESTACK is now handled in terms of function casts,
3626 this has the advantage of retaining type safety without the
3627 use of additional functions. If DEBUG_SAFESTACK is not defined
3628 then the non typesafe macros are used instead. Also modified the
3629 mkstack.pl script to handle the new form. Needs testing to see
3630 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3631 the default if no major problems. Similar behaviour for ASN1_SET_OF
3632 and PKCS12_STACK_OF.
3635 *) When some versions of IIS use the 'NET' form of private key the
3636 key derivation algorithm is different. Normally MD5(password) is
3637 used as a 128 bit RC4 key. In the modified case
3638 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3639 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3640 as the old Netscape_RSA functions except they have an additional
3641 'sgckey' parameter which uses the modified algorithm. Also added
3642 an -sgckey command line option to the rsa utility. Thanks to
3643 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3644 algorithm to openssl-dev.
3647 *) The evp_local.h macros were using 'c.##kname' which resulted in
3648 invalid expansion on some systems (SCO 5.0.5 for example).
3649 Corrected to 'c.kname'.
3650 [Phillip Porch <root@theporch.com>]
3652 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3653 a STACK of email addresses from a certificate or request, these look
3654 in the subject name and the subject alternative name extensions and
3655 omit any duplicate addresses.
3658 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3659 This makes DSA verification about 2 % faster.
3662 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3663 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3664 plus overhead for 1024 bit moduli).
3665 This makes exponentiations about 0.5 % faster for 1024 bit
3666 exponents (as measured by "openssl speed rsa2048").
3669 *) Rename memory handling macros to avoid conflicts with other
3671 Malloc => OPENSSL_malloc
3672 Malloc_locked => OPENSSL_malloc_locked
3673 Realloc => OPENSSL_realloc
3674 Free => OPENSSL_free
3677 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3678 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3681 *) CygWin32 support.
3682 [John Jarvie <jjarvie@newsguy.com>]
3684 *) The type-safe stack code has been rejigged. It is now only compiled
3685 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3686 by default all type-specific stack functions are "#define"d back to
3687 standard stack functions. This results in more streamlined output
3688 but retains the type-safety checking possibilities of the original
3692 *) The STACK code has been cleaned up, and certain type declarations
3693 that didn't make a lot of sense have been brought in line. This has
3694 also involved a cleanup of sorts in safestack.h to more correctly
3695 map type-safe stack functions onto their plain stack counterparts.
3696 This work has also resulted in a variety of "const"ifications of
3697 lots of the code, especially "_cmp" operations which should normally
3698 be prototyped with "const" parameters anyway.
3701 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3702 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3703 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3704 where all of 'md' is used each time the PRNG is used, but 'state'
3705 is used only indexed by a cyclic counter. As entropy may not be
3706 well distributed from the beginning, 'md' is important as a
3707 chaining variable. However, the output function chains only half
3708 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3709 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3710 in all of 'state' being rewritten, with the new values depending
3711 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3714 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3715 the handshake is continued after ssl_verify_cert_chain();
3716 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3717 can lead to 'unexplainable' connection aborts later.
3718 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3720 *) Major EVP API cipher revision.
3721 Add hooks for extra EVP features. This allows various cipher
3722 parameters to be set in the EVP interface. Support added for variable
3723 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3724 setting of RC2 and RC5 parameters.
3726 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3729 Remove lots of duplicated code from the EVP library. For example *every*
3730 cipher init() function handles the 'iv' in the same way according to the
3731 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3732 for CFB and OFB modes they zero ctx->num.
3734 New functionality allows removal of S/MIME code RC2 hack.
3736 Most of the routines have the same form and so can be declared in terms
3739 By shifting this to the top level EVP_CipherInit() it can be removed from
3740 all individual ciphers. If the cipher wants to handle IVs or keys
3741 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3744 Change lots of functions like EVP_EncryptUpdate() to now return a
3745 value: although software versions of the algorithms cannot fail
3746 any installed hardware versions can.
3749 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3750 this option is set, tolerate broken clients that send the negotiated
3751 protocol version number instead of the requested protocol version
3755 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3756 i.e. non-zero for export ciphersuites, zero otherwise.
3757 Previous versions had this flag inverted, inconsistent with
3758 rsa_tmp_cb (..._TMP_RSA_CB).
3759 [Bodo Moeller; problem reported by Amit Chopra]
3761 *) Add missing DSA library text string. Work around for some IIS
3762 key files with invalid SEQUENCE encoding.
3765 *) Add a document (doc/standards.txt) that list all kinds of standards
3766 and so on that are implemented in OpenSSL.
3769 *) Enhance c_rehash script. Old version would mishandle certificates
3770 with the same subject name hash and wouldn't handle CRLs at all.
3771 Added -fingerprint option to crl utility, to support new c_rehash
3775 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3778 *) Fix for SSL server purpose checking. Server checking was
3779 rejecting certificates which had extended key usage present
3780 but no ssl client purpose.
3781 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3783 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3784 is a little unclear about how a blank password is handled.
3785 Since the password in encoded as a BMPString with terminating
3786 double NULL a zero length password would end up as just the
3787 double NULL. However no password at all is different and is
3788 handled differently in the PKCS#12 key generation code. NS
3789 treats a blank password as zero length. MSIE treats it as no
3790 password on export: but it will try both on import. We now do
3791 the same: PKCS12_parse() tries zero length and no password if
3792 the password is set to "" or NULL (NULL is now a valid password:
3793 it wasn't before) as does the pkcs12 application.
3796 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3797 perror when PEM_read_bio_X509_REQ fails, the error message must
3798 be obtained from the error queue.
3801 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3802 it in ERR_remove_state if appropriate, and change ERR_get_state
3803 accordingly to avoid race conditions (this is necessary because
3804 thread_hash is no longer constant once set).
3807 *) Bugfix for linux-elf makefile.one.
3810 *) RSA_get_default_method() will now cause a default
3811 RSA_METHOD to be chosen if one doesn't exist already.
3812 Previously this was only set during a call to RSA_new()
3813 or RSA_new_method(NULL) meaning it was possible for
3814 RSA_get_default_method() to return NULL.
3817 *) Added native name translation to the existing DSO code
3818 that will convert (if the flag to do so is set) filenames
3819 that are sufficiently small and have no path information
3820 into a canonical native form. Eg. "blah" converted to
3821 "libblah.so" or "blah.dll" etc.
3824 *) New function ERR_error_string_n(e, buf, len) which is like
3825 ERR_error_string(e, buf), but writes at most 'len' bytes
3826 including the 0 terminator. For ERR_error_string_n, 'buf'
3828 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3830 *) CONF library reworked to become more general. A new CONF
3831 configuration file reader "class" is implemented as well as a
3832 new functions (NCONF_*, for "New CONF") to handle it. The now
3833 old CONF_* functions are still there, but are reimplemented to
3834 work in terms of the new functions. Also, a set of functions
3835 to handle the internal storage of the configuration data is
3836 provided to make it easier to write new configuration file
3837 reader "classes" (I can definitely see something reading a
3838 configuration file in XML format, for example), called _CONF_*,
3839 or "the configuration storage API"...
3841 The new configuration file reading functions are:
3843 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3844 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3846 NCONF_default, NCONF_WIN32
3848 NCONF_dump_fp, NCONF_dump_bio
3850 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3851 NCONF_new creates a new CONF object. This works in the same way
3852 as other interfaces in OpenSSL, like the BIO interface.
3853 NCONF_dump_* dump the internal storage of the configuration file,
3854 which is useful for debugging. All other functions take the same
3855 arguments as the old CONF_* functions wth the exception of the
3856 first that must be a `CONF *' instead of a `LHASH *'.
3858 To make it easer to use the new classes with the old CONF_* functions,
3859 the function CONF_set_default_method is provided.
3862 *) Add '-tls1' option to 'openssl ciphers', which was already
3863 mentioned in the documentation but had not been implemented.
3864 (This option is not yet really useful because even the additional
3865 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3868 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3869 OpenSSL-based applications) load shared libraries and bind to
3870 them in a portable way.
3871 [Geoff Thorpe, with contributions from Richard Levitte]
3873 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3875 *) Make sure _lrotl and _lrotr are only used with MSVC.
3877 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3878 (the default implementation of RAND_status).
3880 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3881 to '-clrext' (= clear extensions), as intended and documented.
3882 [Bodo Moeller; inconsistency pointed out by Michael Attili
3885 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3886 was larger than the MD block size.
3887 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3889 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3890 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3891 using the passed key: if the passed key was a private key the result
3892 of X509_print(), for example, would be to print out all the private key
3896 *) des_quad_cksum() byte order bug fix.
3897 [Ulf Möller, using the problem description in krb4-0.9.7, where
3898 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3900 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3902 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3904 *) For easily testing in shell scripts whether some command
3905 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3906 returns with exit code 0 iff no command of the given name is available.
3907 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3908 the output goes to stdout and nothing is printed to stderr.
3909 Additional arguments are always ignored.
3911 Since for each cipher there is a command of the same name,
3912 the 'no-cipher' compilation switches can be tested this way.
3914 ('openssl no-XXX' is not able to detect pseudo-commands such
3915 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3918 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3921 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3922 is set; it will be thrown away anyway because each handshake creates
3924 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3925 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3926 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3927 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3930 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3931 'Q' and 'R' lose their special meanings (quit/renegotiate).
3932 This is part of what -quiet does; unlike -quiet, -ign_eof
3933 does not suppress any output.
3936 *) Add compatibility options to the purpose and trust code. The
3937 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3938 accepts a certificate or CA, this was the previous behaviour,
3939 with all the associated security issues.
3941 X509_TRUST_COMPAT is the old trust behaviour: only and
3942 automatically trust self signed roots in certificate store. A
3943 new trust setting X509_TRUST_DEFAULT is used to specify that
3944 a purpose has no associated trust setting and it should instead
3945 use the value in the default purpose.
3948 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3949 and fix a memory leak.
3952 *) In util/mkerr.pl (which implements 'make errors'), preserve
3953 reason strings from the previous version of the .c file, as
3954 the default to have only downcase letters (and digits) in
3955 automatically generated reasons codes is not always appropriate.
3958 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3959 using strerror. Previously, ERR_reason_error_string() returned
3960 library names as reason strings for SYSerr; but SYSerr is a special
3961 case where small numbers are errno values, not library numbers.
3964 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3965 converts DSA parameters into DH parameters. (When creating parameters,
3966 DSA_generate_parameters is used.)
3969 *) Include 'length' (recommended exponent length) in C code generated
3970 by 'openssl dhparam -C'.
3973 *) The second argument to set_label in perlasm was already being used
3974 so couldn't be used as a "file scope" flag. Moved to third argument
3978 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3979 instead of RAND_bytes for encryption IVs and salts.
3982 *) Include RAND_status() into RAND_METHOD instead of implementing
3983 it only for md_rand.c Otherwise replacing the PRNG by calling
3984 RAND_set_rand_method would be impossible.
3987 *) Don't let DSA_generate_key() enter an infinite loop if the random
3988 number generation fails.
3991 *) New 'rand' application for creating pseudo-random output.
3994 *) Added configuration support for Linux/IA64
3995 [Rolf Haberrecker <rolf@suse.de>]
3997 *) Assembler module support for Mingw32.
4000 *) Shared library support for HPUX (in shlib/).
4001 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4003 *) Shared library support for Solaris gcc.
4004 [Lutz Behnke <behnke@trustcenter.de>]
4006 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4008 *) PKCS7_encrypt() was adding text MIME headers twice because they
4009 were added manually and by SMIME_crlf_copy().
4012 *) In bntest.c don't call BN_rand with zero bits argument.
4013 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4015 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4016 case was implemented. This caused BN_div_recp() to fail occasionally.
4019 *) Add an optional second argument to the set_label() in the perl
4020 assembly language builder. If this argument exists and is set
4021 to 1 it signals that the assembler should use a symbol whose
4022 scope is the entire file, not just the current function. This
4023 is needed with MASM which uses the format label:: for this scope.
4024 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4026 *) Change the ASN1 types so they are typedefs by default. Before
4027 almost all types were #define'd to ASN1_STRING which was causing
4028 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4032 *) Change names of new functions to the new get1/get0 naming
4033 convention: After 'get1', the caller owns a reference count
4034 and has to call ..._free; 'get0' returns a pointer to some
4035 data structure without incrementing reference counters.
4036 (Some of the existing 'get' functions increment a reference
4037 counter, some don't.)
4038 Similarly, 'set1' and 'add1' functions increase reference
4039 counters or duplicate objects.
4042 *) Allow for the possibility of temp RSA key generation failure:
4043 the code used to assume it always worked and crashed on failure.
4046 *) Fix potential buffer overrun problem in BIO_printf().
4047 [Ulf Möller, using public domain code by Patrick Powell; problem
4048 pointed out by David Sacerdote <das33@cornell.edu>]
4050 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4051 RAND_egd() and RAND_status(). In the command line application,
4052 the EGD socket can be specified like a seed file using RANDFILE
4056 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4057 Some CAs (e.g. Verisign) distribute certificates in this form.
4060 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4061 list to exclude them. This means that no special compilation option
4062 is needed to use anonymous DH: it just needs to be included in the
4066 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4067 EVP_MD_type. The old functionality is available in a new macro called
4068 EVP_MD_md(). Change code that uses it and update docs.
4071 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4072 where the 'void *' argument is replaced by a function pointer argument.
4073 Previously 'void *' was abused to point to functions, which works on
4074 many platforms, but is not correct. As these functions are usually
4075 called by macros defined in OpenSSL header files, most source code
4076 should work without changes.
4079 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4080 sections with information on -D... compiler switches used for
4081 compiling the library so that applications can see them. To enable
4082 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4083 must be defined. E.g.,
4084 #define OPENSSL_ALGORITHM_DEFINES
4085 #include <openssl/opensslconf.h>
4086 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4087 [Richard Levitte, Ulf and Bodo Möller]
4089 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4093 *) Change the 'other' type in certificate aux info to a STACK_OF
4094 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4095 the required ASN1 format: arbitrary types determined by an OID.
4098 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4099 argument to 'req'. This is not because the function is newer or
4100 better than others it just uses the work 'NEW' in the certificate
4101 request header lines. Some software needs this.
4104 *) Reorganise password command line arguments: now passwords can be
4105 obtained from various sources. Delete the PEM_cb function and make
4106 it the default behaviour: i.e. if the callback is NULL and the
4107 usrdata argument is not NULL interpret it as a null terminated pass
4108 phrase. If usrdata and the callback are NULL then the pass phrase
4109 is prompted for as usual.
4112 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4113 the support is automatically enabled. The resulting binaries will
4114 autodetect the card and use it if present.
4115 [Ben Laurie and Compaq Inc.]
4117 *) Work around for Netscape hang bug. This sends certificate request
4118 and server done in one record. Since this is perfectly legal in the
4119 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4120 the bugs/SSLv3 entry for more info.
4123 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4126 *) Add -rand argument to smime and pkcs12 applications and read/write
4130 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4133 *) Add command line password options to the remaining applications.
4136 *) Bug fix for BN_div_recp() for numerators with an even number of
4140 *) More tests in bntest.c, and changed test_bn output.
4143 *) ./config recognizes MacOS X now.
4146 *) Bug fix for BN_div() when the first words of num and divsor are
4147 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4150 *) Add support for various broken PKCS#8 formats, and command line
4151 options to produce them.
4154 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4155 get temporary BIGNUMs from a BN_CTX.
4158 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4162 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4163 include a #define from the old name to the new. The original intent
4164 was that statically linked binaries could for example just call
4165 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4166 link with digests. This never worked becayse SSLeay_add_all_digests()
4167 and SSLeay_add_all_ciphers() were in the same source file so calling
4168 one would link with the other. They are now in separate source files.
4171 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4174 *) Use a less unusual form of the Miller-Rabin primality test (it used
4175 a binary algorithm for exponentiation integrated into the Miller-Rabin
4176 loop, our standard modexp algorithms are faster).
4179 *) Support for the EBCDIC character set completed.
4180 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4182 *) Source code cleanups: use const where appropriate, eliminate casts,
4183 use void * instead of char * in lhash.
4186 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4187 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4188 this the server could overwrite ephemeral keys that the client
4192 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4193 using 50 iterations of the Rabin-Miller test.
4195 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4196 iterations of the Rabin-Miller test as required by the appendix
4197 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4198 As BN_is_prime_fasttest includes trial division, DSA parameter
4199 generation becomes much faster.
4201 This implies a change for the callback functions in DSA_is_prime
4202 and DSA_generate_parameters: The callback function is called once
4203 for each positive witness in the Rabin-Miller test, not just
4204 occasionally in the inner loop; and the parameters to the
4205 callback function now provide an iteration count for the outer
4206 loop rather than for the current invocation of the inner loop.
4207 DSA_generate_parameters additionally can call the callback
4208 function with an 'iteration count' of -1, meaning that a
4209 candidate has passed the trial division test (when q is generated
4210 from an application-provided seed, trial division is skipped).
4213 *) New function BN_is_prime_fasttest that optionally does trial
4214 division before starting the Rabin-Miller test and has
4215 an additional BN_CTX * argument (whereas BN_is_prime always
4216 has to allocate at least one BN_CTX).
4217 'callback(1, -1, cb_arg)' is called when a number has passed the
4218 trial division stage.
4221 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4225 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4228 *) New function BN_pseudo_rand().
4231 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4232 bignum version of BN_from_montgomery() with the working code from
4233 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4237 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4238 made it impossible to use the same SSL_SESSION data structure in
4239 SSL2 clients in multiple threads.
4242 *) The return value of RAND_load_file() no longer counts bytes obtained
4243 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4244 to seed the PRNG (previously an explicit byte count was required).
4245 [Ulf Möller, Bodo Möller]
4247 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4248 used (char *) instead of (void *) and had casts all over the place.
4251 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4254 *) Retain source code compatibility for BN_prime_checks macro:
4255 BN_is_prime(..., BN_prime_checks, ...) now uses
4256 BN_prime_checks_for_size to determine the appropriate number of
4257 Rabin-Miller iterations.
4260 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4261 DH_CHECK_P_NOT_SAFE_PRIME.
4262 (Check if this is true? OpenPGP calls them "strong".)
4265 *) Merge the functionality of "dh" and "gendh" programs into a new program
4266 "dhparam". The old programs are retained for now but will handle DH keys
4267 (instead of parameters) in future.
4270 *) Make the ciphers, s_server and s_client programs check the return values
4271 when a new cipher list is set.
4274 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4275 ciphers. Before when the 56bit ciphers were enabled the sorting was
4278 The syntax for the cipher sorting has been extended to support sorting by
4279 cipher-strength (using the strength_bits hard coded in the tables).
4280 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4282 Fix a bug in the cipher-command parser: when supplying a cipher command
4283 string with an "undefined" symbol (neither command nor alphanumeric
4284 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4285 an error is flagged.
4287 Due to the strength-sorting extension, the code of the
4288 ssl_create_cipher_list() function was completely rearranged. I hope that
4289 the readability was also increased :-)
4290 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4292 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4293 for the first serial number and places 2 in the serial number file. This
4294 avoids problems when the root CA is created with serial number zero and
4295 the first user certificate has the same issuer name and serial number
4299 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4300 the new code. Add documentation for this stuff.
4303 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4304 X509_*() to X509at_*() on the grounds that they don't handle X509
4305 structures and behave in an analagous way to the X509v3 functions:
4306 they shouldn't be called directly but wrapper functions should be used
4309 So we also now have some wrapper functions that call the X509at functions
4310 when passed certificate requests. (TO DO: similar things can be done with
4311 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4312 things. Some of these need some d2i or i2d and print functionality
4313 because they handle more complex structures.)
4316 *) Add missing #ifndefs that caused missing symbols when building libssl
4317 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4318 NO_RSA in ssl/s2*.c.
4319 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4321 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4322 has a return value which indicates the quality of the random data
4323 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4324 error queue. New function RAND_pseudo_bytes() generates output that is
4325 guaranteed to be unique but not unpredictable. RAND_add is like
4326 RAND_seed, but takes an extra argument for an entropy estimate
4327 (RAND_seed always assumes full entropy).
4330 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4331 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4332 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4333 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4334 false-positive rate of at most 2^-80 for random input.
4337 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4340 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4341 in the 0.9.5 release), this returns the chain
4342 from an X509_CTX structure with a dup of the stack and all
4343 the X509 reference counts upped: so the stack will exist
4344 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4347 Also make SSL_SESSION_print() print out the verify return
4351 *) Add manpage for the pkcs12 command. Also change the default
4352 behaviour so MAC iteration counts are used unless the new
4353 -nomaciter option is used. This improves file security and
4354 only older versions of MSIE (4.0 for example) need it.
4357 *) Honor the no-xxx Configure options when creating .DEF files.
4360 *) Add PKCS#10 attributes to field table: challengePassword,
4361 unstructuredName and unstructuredAddress. These are taken from
4362 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4363 international characters are used.
4365 More changes to X509_ATTRIBUTE code: allow the setting of types
4366 based on strings. Remove the 'loc' parameter when adding
4367 attributes because these will be a SET OF encoding which is sorted
4371 *) Initial changes to the 'req' utility to allow request generation
4372 automation. This will allow an application to just generate a template
4373 file containing all the field values and have req construct the
4376 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4377 used all over the place including certificate requests and PKCS#7
4378 structures. They are currently handled manually where necessary with
4379 some primitive wrappers for PKCS#7. The new functions behave in a
4380 manner analogous to the X509 extension functions: they allow
4381 attributes to be looked up by NID and added.
4383 Later something similar to the X509V3 code would be desirable to
4384 automatically handle the encoding, decoding and printing of the
4385 more complex types. The string types like challengePassword can
4386 be handled by the string table functions.
4388 Also modified the multi byte string table handling. Now there is
4389 a 'global mask' which masks out certain types. The table itself
4390 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4391 is useful when for example there is only one permissible type
4392 (as in countryName) and using the mask might result in no valid
4396 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4397 SSL_get_peer_finished to allow applications to obtain the latest
4398 Finished messages sent to the peer or expected from the peer,
4399 respectively. (SSL_get_peer_finished is usually the Finished message
4400 actually received from the peer, otherwise the protocol will be aborted.)
4402 As the Finished message are message digests of the complete handshake
4403 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4404 be used for external authentication procedures when the authentication
4405 provided by SSL/TLS is not desired or is not enough.
4408 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4409 the host supports BWX extension and if Compaq C is present on the
4410 $PATH. Just exploiting of the BWX extension results in 20-30%
4411 performance kick for some algorithms, e.g. DES and RC4 to mention
4412 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4416 *) Add support for MS "fast SGC". This is arguably a violation of the
4417 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4418 weak crypto and after checking the certificate is SGC a second one
4419 with strong crypto. MS SGC stops the first handshake after receiving
4420 the server certificate message and sends a second client hello. Since
4421 a server will typically do all the time consuming operations before
4422 expecting any further messages from the client (server key exchange
4423 is the most expensive) there is little difference between the two.
4425 To get OpenSSL to support MS SGC we have to permit a second client
4426 hello message after we have sent server done. In addition we have to
4427 reset the MAC if we do get this second client hello.
4430 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4431 if a DER encoded private key is RSA or DSA traditional format. Changed
4432 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4433 format DER encoded private key. Newer code should use PKCS#8 format which
4434 has the key type encoded in the ASN1 structure. Added DER private key
4435 support to pkcs8 application.
4438 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4439 ciphersuites has been selected (as required by the SSL 3/TLS 1
4440 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4441 is set, we interpret this as a request to violate the specification
4442 (the worst that can happen is a handshake failure, and 'correct'
4443 behaviour would result in a handshake failure anyway).
4446 *) In SSL_CTX_add_session, take into account that there might be multiple
4447 SSL_SESSION structures with the same session ID (e.g. when two threads
4448 concurrently obtain them from an external cache).
4449 The internal cache can handle only one SSL_SESSION with a given ID,
4450 so if there's a conflict, we now throw out the old one to achieve
4454 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4455 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4456 some routines that use cipher OIDs: some ciphers do not have OIDs
4457 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4461 *) Simplify the trust setting structure and code. Now we just have
4462 two sequences of OIDs for trusted and rejected settings. These will
4463 typically have values the same as the extended key usage extension
4464 and any application specific purposes.
4466 The trust checking code now has a default behaviour: it will just
4467 check for an object with the same NID as the passed id. Functions can
4468 be provided to override either the default behaviour or the behaviour
4469 for a given id. SSL client, server and email already have functions
4470 in place for compatibility: they check the NID and also return "trusted"
4471 if the certificate is self signed.
4474 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4475 traditional format into an EVP_PKEY structure.
4478 *) Add a password callback function PEM_cb() which either prompts for
4479 a password if usr_data is NULL or otherwise assumes it is a null
4480 terminated password. Allow passwords to be passed on command line
4481 environment or config files in a few more utilities.
4484 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4485 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4486 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4487 Update documentation.
4490 *) Support for ASN1 "NULL" type. This could be handled before by using
4491 ASN1_TYPE but there wasn't any function that would try to read a NULL
4492 and produce an error if it couldn't. For compatibility we also have
4493 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4494 don't allocate anything because they don't need to.
4497 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4499 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4501 *) Rebuild of the memory allocation routines used by OpenSSL code and
4502 possibly others as well. The purpose is to make an interface that
4503 provide hooks so anyone can build a separate set of allocation and
4504 deallocation routines to be used by OpenSSL, for example memory
4505 pool implementations, or something else, which was previously hard
4506 since Malloc(), Realloc() and Free() were defined as macros having
4507 the values malloc, realloc and free, respectively (except for Win32
4508 compilations). The same is provided for memory debugging code.
4509 OpenSSL already comes with functionality to find memory leaks, but
4510 this gives people a chance to debug other memory problems.
4512 With these changes, a new set of functions and macros have appeared:
4514 CRYPTO_set_mem_debug_functions() [F]
4515 CRYPTO_get_mem_debug_functions() [F]
4516 CRYPTO_dbg_set_options() [F]
4517 CRYPTO_dbg_get_options() [F]
4518 CRYPTO_malloc_debug_init() [M]
4520 The memory debug functions are NULL by default, unless the library
4521 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4522 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4523 gives the standard debugging functions that come with OpenSSL) or
4524 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4525 provided by the library user) must be used. When the standard
4526 debugging functions are used, CRYPTO_dbg_set_options can be used to
4527 request additional information:
4528 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4529 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4531 Also, things like CRYPTO_set_mem_functions will always give the
4532 expected result (the new set of functions is used for allocation
4533 and deallocation) at all times, regardless of platform and compiler
4536 To finish it up, some functions that were never use in any other
4537 way than through macros have a new API and new semantic:
4540 CRYPTO_dbg_realloc()
4543 All macros of value have retained their old syntax.
4544 [Richard Levitte and Bodo Moeller]
4546 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4547 ordering of SMIMECapabilities wasn't in "strength order" and there
4548 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4552 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4553 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4554 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4556 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4557 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4558 functionality to handle multipart/signed properly) and a utility
4559 called 'smime' to call all this stuff. This is based on code I
4560 originally wrote for Celo who have kindly allowed it to be
4561 included in OpenSSL.
4564 *) Add variants des_set_key_checked and des_set_key_unchecked of
4565 des_set_key (aka des_key_sched). Global variable des_check_key
4566 decides which of these is called by des_set_key; this way
4567 des_check_key behaves as it always did, but applications and
4568 the library itself, which was buggy for des_check_key == 1,
4569 have a cleaner way to pick the version they need.
4572 *) New function PKCS12_newpass() which changes the password of a
4576 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4577 dynamic mix. In both cases the ids can be used as an index into the
4578 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4579 functions so they accept a list of the field values and the
4580 application doesn't need to directly manipulate the X509_TRUST
4584 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4588 *) Modify the way the V3 extension code looks up extensions. This now
4589 works in a similar way to the object code: we have some "standard"
4590 extensions in a static table which is searched with OBJ_bsearch()
4591 and the application can add dynamic ones if needed. The file
4592 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4593 updated whenever a new extension is added to the core code and kept
4594 in ext_nid order. There is a simple program 'tabtest.c' which checks
4595 this. New extensions are not added too often so this file can readily
4596 be maintained manually.
4598 There are two big advantages in doing things this way. The extensions
4599 can be looked up immediately and no longer need to be "added" using
4600 X509V3_add_standard_extensions(): this function now does nothing.
4601 [Side note: I get *lots* of email saying the extension code doesn't
4602 work because people forget to call this function]
4603 Also no dynamic allocation is done unless new extensions are added:
4604 so if we don't add custom extensions there is no need to call
4605 X509V3_EXT_cleanup().
4608 *) Modify enc utility's salting as follows: make salting the default. Add a
4609 magic header, so unsalted files fail gracefully instead of just decrypting
4610 to garbage. This is because not salting is a big security hole, so people
4611 should be discouraged from doing it.
4614 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4615 digest to be passed on the command line but it only used this
4616 parameter when signing a certificate. Modified so all relevant
4617 operations are affected by the digest parameter including the
4618 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4619 DSA key was used because it didn't fix the digest.
4622 *) Initial certificate chain verify code. Currently tests the untrusted
4623 certificates for consistency with the verify purpose (which is set
4624 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4626 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4627 this is because it will reject chains with invalid extensions whereas
4628 every previous version of OpenSSL and SSLeay made no checks at all.
4630 Trust code: checks the root CA for the relevant trust settings. Trust
4631 settings have an initial value consistent with the verify purpose: e.g.
4632 if the verify purpose is for SSL client use it expects the CA to be
4633 trusted for SSL client use. However the default value can be changed to
4634 permit custom trust settings: one example of this would be to only trust
4635 certificates from a specific "secure" set of CAs.
4637 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4638 which should be used for version portability: especially since the
4639 verify structure is likely to change more often now.
4641 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4642 to set them. If not set then assume SSL clients will verify SSL servers
4645 Two new options to the verify program: -untrusted allows a set of
4646 untrusted certificates to be passed in and -purpose which sets the
4647 intended purpose of the certificate. If a purpose is set then the
4648 new chain verify code is used to check extension consistency.
4651 *) Support for the authority information access extension.
4654 *) Modify RSA and DSA PEM read routines to transparently handle
4655 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4656 public keys in a format compatible with certificate
4657 SubjectPublicKeyInfo structures. Unfortunately there were already
4658 functions called *_PublicKey_* which used various odd formats so
4659 these are retained for compatibility: however the DSA variants were
4660 never in a public release so they have been deleted. Changed dsa/rsa
4661 utilities to handle the new format: note no releases ever handled public
4662 keys so we should be OK.
4664 The primary motivation for this change is to avoid the same fiasco
4665 that dogs private keys: there are several incompatible private key
4666 formats some of which are standard and some OpenSSL specific and
4667 require various evil hacks to allow partial transparent handling and
4668 even then it doesn't work with DER formats. Given the option anything
4669 other than PKCS#8 should be dumped: but the other formats have to
4670 stay in the name of compatibility.
4672 With public keys and the benefit of hindsight one standard format
4673 is used which works with EVP_PKEY, RSA or DSA structures: though
4674 it clearly returns an error if you try to read the wrong kind of key.
4676 Added a -pubkey option to the 'x509' utility to output the public key.
4677 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4678 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4679 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4680 that do the same as the EVP_PKEY_assign_*() except they up the
4681 reference count of the added key (they don't "swallow" the
4685 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4686 CRLs would fail if the file contained no certificates or no CRLs:
4687 added a new function to read in both types and return the number
4688 read: this means that if none are read it will be an error. The
4689 DER versions of the certificate and CRL reader would always fail
4690 because it isn't possible to mix certificates and CRLs in DER format
4691 without choking one or the other routine. Changed this to just read
4692 a certificate: this is the best we can do. Also modified the code
4693 in apps/verify.c to take notice of return codes: it was previously
4694 attempting to read in certificates from NULL pointers and ignoring
4695 any errors: this is one reason why the cert and CRL reader seemed
4696 to work. It doesn't check return codes from the default certificate
4697 routines: these may well fail if the certificates aren't installed.
4700 *) Code to support otherName option in GeneralName.
4703 *) First update to verify code. Change the verify utility
4704 so it warns if it is passed a self signed certificate:
4705 for consistency with the normal behaviour. X509_verify
4706 has been modified to it will now verify a self signed
4707 certificate if *exactly* the same certificate appears
4708 in the store: it was previously impossible to trust a
4709 single self signed certificate. This means that:
4710 openssl verify ss.pem
4711 now gives a warning about a self signed certificate but
4712 openssl verify -CAfile ss.pem ss.pem
4716 *) For servers, store verify_result in SSL_SESSION data structure
4717 (and add it to external session representation).
4718 This is needed when client certificate verifications fails,
4719 but an application-provided verification callback (set by
4720 SSL_CTX_set_cert_verify_callback) allows accepting the session
4721 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4722 but returns 1): When the session is reused, we have to set
4723 ssl->verify_result to the appropriate error code to avoid
4725 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4727 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4728 case in PKCS7_dataInit() where the signed PKCS7 structure
4729 didn't contain any existing data because it was being created.
4730 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4732 *) Add a salt to the key derivation routines in enc.c. This
4733 forms the first 8 bytes of the encrypted file. Also add a
4734 -S option to allow a salt to be input on the command line.
4737 *) New function X509_cmp(). Oddly enough there wasn't a function
4738 to compare two certificates. We do this by working out the SHA1
4739 hash and comparing that. X509_cmp() will be needed by the trust
4743 *) SSL_get1_session() is like SSL_get_session(), but increments
4744 the reference count in the SSL_SESSION returned.
4745 [Geoff Thorpe <geoff@eu.c2.net>]
4747 *) Fix for 'req': it was adding a null to request attributes.
4748 Also change the X509_LOOKUP and X509_INFO code to handle
4749 certificate auxiliary information.
4752 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4756 *) Add the possibility to add extra information to the memory leak
4757 detecting output, to form tracebacks, showing from where each
4758 allocation was originated: CRYPTO_push_info("constant string") adds
4759 the string plus current file name and line number to a per-thread
4760 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4761 is like calling CYRPTO_pop_info() until the stack is empty.
4762 Also updated memory leak detection code to be multi-thread-safe.
4765 *) Add options -text and -noout to pkcs7 utility and delete the
4766 encryption options which never did anything. Update docs.
4769 *) Add options to some of the utilities to allow the pass phrase
4770 to be included on either the command line (not recommended on
4771 OSes like Unix) or read from the environment. Update the
4772 manpages and fix a few bugs.
4775 *) Add a few manpages for some of the openssl commands.
4778 *) Fix the -revoke option in ca. It was freeing up memory twice,
4779 leaking and not finding already revoked certificates.
4782 *) Extensive changes to support certificate auxiliary information.
4783 This involves the use of X509_CERT_AUX structure and X509_AUX
4784 functions. An X509_AUX function such as PEM_read_X509_AUX()
4785 can still read in a certificate file in the usual way but it
4786 will also read in any additional "auxiliary information". By
4787 doing things this way a fair degree of compatibility can be
4788 retained: existing certificates can have this information added
4789 using the new 'x509' options.
4791 Current auxiliary information includes an "alias" and some trust
4792 settings. The trust settings will ultimately be used in enhanced
4793 certificate chain verification routines: currently a certificate
4794 can only be trusted if it is self signed and then it is trusted
4798 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4799 The problem was that one of the replacement routines had not been working
4800 since SSLeay releases. For now the offending routine has been replaced
4801 with non-optimised assembler. Even so, this now gives around 95%
4802 performance improvement for 1024 bit RSA signs.
4805 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4806 handling. Most clients have the effective key size in bits equal to
4807 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4808 A few however don't do this and instead use the size of the decrypted key
4809 to determine the RC2 key length and the AlgorithmIdentifier to determine
4810 the effective key length. In this case the effective key length can still
4811 be 40 bits but the key length can be 168 bits for example. This is fixed
4812 by manually forcing an RC2 key into the EVP_PKEY structure because the
4813 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4814 the key length and effective key length are equal.
4817 *) Add a bunch of functions that should simplify the creation of
4818 X509_NAME structures. Now you should be able to do:
4819 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4820 and have it automatically work out the correct field type and fill in
4821 the structures. The more adventurous can try:
4822 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4823 and it will (hopefully) work out the correct multibyte encoding.
4826 *) Change the 'req' utility to use the new field handling and multibyte
4827 copy routines. Before the DN field creation was handled in an ad hoc
4828 way in req, ca, and x509 which was rather broken and didn't support
4829 BMPStrings or UTF8Strings. Since some software doesn't implement
4830 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4831 using the dirstring_type option. See the new comment in the default
4832 openssl.cnf for more info.
4835 *) Make crypto/rand/md_rand.c more robust:
4836 - Assure unique random numbers after fork().
4837 - Make sure that concurrent threads access the global counter and
4838 md serializably so that we never lose entropy in them
4839 or use exactly the same state in multiple threads.
4840 Access to the large state is not always serializable because
4841 the additional locking could be a performance killer, and
4842 md should be large enough anyway.
4845 *) New file apps/app_rand.c with commonly needed functionality
4846 for handling the random seed file.
4848 Use the random seed file in some applications that previously did not:
4850 dsaparam -genkey (which also ignored its '-rand' option),
4853 x509 (when signing).
4854 Except on systems with /dev/urandom, it is crucial to have a random
4855 seed file at least for key creation, DSA signing, and for DH exchanges;
4856 for RSA signatures we could do without one.
4858 gendh and gendsa (unlike genrsa) used to read only the first byte
4859 of each file listed in the '-rand' option. The function as previously
4860 found in genrsa is now in app_rand.c and is used by all programs
4861 that support '-rand'.
4864 *) In RAND_write_file, use mode 0600 for creating files;
4865 don't just chmod when it may be too late.
4868 *) Report an error from X509_STORE_load_locations
4869 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4872 *) New function ASN1_mbstring_copy() this copies a string in either
4873 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4874 into an ASN1_STRING type. A mask of permissible types is passed
4875 and it chooses the "minimal" type to use or an error if not type
4879 *) Add function equivalents to the various macros in asn1.h. The old
4880 macros are retained with an M_ prefix. Code inside the library can
4881 use the M_ macros. External code (including the openssl utility)
4882 should *NOT* in order to be "shared library friendly".
4885 *) Add various functions that can check a certificate's extensions
4886 to see if it usable for various purposes such as SSL client,
4887 server or S/MIME and CAs of these types. This is currently
4888 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4889 verification. Also added a -purpose flag to x509 utility to
4890 print out all the purposes.
4893 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4897 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4898 for, obtain and decode and extension and obtain its critical flag.
4899 This allows all the necessary extension code to be handled in a
4900 single function call.
4903 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4904 platforms. See crypto/rc4/rc4_enc.c for further details.
4907 *) New -noout option to asn1parse. This causes no output to be produced
4908 its main use is when combined with -strparse and -out to extract data
4909 from a file (which may not be in ASN.1 format).
4912 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4913 when producing the local key id.
4914 [Richard Levitte <levitte@stacken.kth.se>]
4916 *) New option -dhparam in s_server. This allows a DH parameter file to be
4917 stated explicitly. If it is not stated then it tries the first server
4918 certificate file. The previous behaviour hard coded the filename
4922 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4923 a public key to be input or output. For example:
4924 openssl rsa -in key.pem -pubout -out pubkey.pem
4925 Also added necessary DSA public key functions to handle this.
4928 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4929 in the message. This was handled by allowing
4930 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4931 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4933 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4934 to the end of the strings whereas this didn't. This would cause problems
4935 if strings read with d2i_ASN1_bytes() were later modified.
4936 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4938 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4939 data and it contains EOF it will end up returning an error. This is
4940 caused by input 46 bytes long. The cause is due to the way base64
4941 BIOs find the start of base64 encoded data. They do this by trying a
4942 trial decode on each line until they find one that works. When they
4943 do a flag is set and it starts again knowing it can pass all the
4944 data directly through the decoder. Unfortunately it doesn't reset
4945 the context it uses. This means that if EOF is reached an attempt
4946 is made to pass two EOFs through the context and this causes the
4947 resulting error. This can also cause other problems as well. As is
4948 usual with these problems it takes *ages* to find and the fix is
4949 trivial: move one line.
4950 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4952 *) Ugly workaround to get s_client and s_server working under Windows. The
4953 old code wouldn't work because it needed to select() on sockets and the
4954 tty (for keypresses and to see if data could be written). Win32 only
4955 supports select() on sockets so we select() with a 1s timeout on the
4956 sockets and then see if any characters are waiting to be read, if none
4957 are present then we retry, we also assume we can always write data to
4958 the tty. This isn't nice because the code then blocks until we've
4959 received a complete line of data and it is effectively polling the
4960 keyboard at 1s intervals: however it's quite a bit better than not
4961 working at all :-) A dedicated Windows application might handle this
4962 with an event loop for example.
4965 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4966 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4967 will be called when RSA_sign() and RSA_verify() are used. This is useful
4968 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4969 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4970 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4971 This necessitated the support of an extra signature type NID_md5_sha1
4972 for SSL signatures and modifications to the SSL library to use it instead
4973 of calling RSA_public_decrypt() and RSA_private_encrypt().
4976 *) Add new -verify -CAfile and -CApath options to the crl program, these
4977 will lookup a CRL issuers certificate and verify the signature in a
4978 similar way to the verify program. Tidy up the crl program so it
4979 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4980 less strict. It will now permit CRL extensions even if it is not
4981 a V2 CRL: this will allow it to tolerate some broken CRLs.
4984 *) Initialize all non-automatic variables each time one of the openssl
4985 sub-programs is started (this is necessary as they may be started
4986 multiple times from the "OpenSSL>" prompt).
4987 [Lennart Bang, Bodo Moeller]
4989 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4990 removing all other RSA functionality (this is what NO_RSA does). This
4991 is so (for example) those in the US can disable those operations covered
4992 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4996 *) Non-copying interface to BIO pairs.
4997 (still largely untested)
5000 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5001 ASCII string. This was handled independently in various places before.
5004 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5005 UTF8 strings a character at a time.
5008 *) Use client_version from client hello to select the protocol
5009 (s23_srvr.c) and for RSA client key exchange verification
5010 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5013 *) Add various utility functions to handle SPKACs, these were previously
5014 handled by poking round in the structure internals. Added new function
5015 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5016 print, verify and generate SPKACs. Based on an original idea from
5017 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5020 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5023 *) Allow the config file extension section to be overwritten on the
5024 command line. Based on an original idea from Massimiliano Pala
5025 <madwolf@comune.modena.it>. The new option is called -extensions
5026 and can be applied to ca, req and x509. Also -reqexts to override
5027 the request extensions in req and -crlexts to override the crl extensions
5031 *) Add new feature to the SPKAC handling in ca. Now you can include
5032 the same field multiple times by preceding it by "XXXX." for example:
5035 this is the same syntax as used in the req config file.
5038 *) Allow certificate extensions to be added to certificate requests. These
5039 are specified in a 'req_extensions' option of the req section of the
5040 config file. They can be printed out with the -text option to req but
5041 are otherwise ignored at present.
5044 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5045 data read consists of only the final block it would not decrypted because
5046 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5047 A misplaced 'break' also meant the decrypted final block might not be
5048 copied until the next read.
5051 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5052 a few extra parameters to the DH structure: these will be useful if
5053 for example we want the value of 'q' or implement X9.42 DH.
5056 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5057 provides hooks that allow the default DSA functions or functions on a
5058 "per key" basis to be replaced. This allows hardware acceleration and
5059 hardware key storage to be handled without major modification to the
5060 library. Also added low level modexp hooks and CRYPTO_EX structure and
5061 associated functions.
5064 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5065 as "read only": it can't be written to and the buffer it points to will
5066 not be freed. Reading from a read only BIO is much more efficient than
5067 a normal memory BIO. This was added because there are several times when
5068 an area of memory needs to be read from a BIO. The previous method was
5069 to create a memory BIO and write the data to it, this results in two
5070 copies of the data and an O(n^2) reading algorithm. There is a new
5071 function BIO_new_mem_buf() which creates a read only memory BIO from
5072 an area of memory. Also modified the PKCS#7 routines to use read only
5076 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5077 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5078 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5079 but a retry condition occured while trying to read the rest.
5082 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5083 NID_pkcs7_encrypted by default: this was wrong since this should almost
5084 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5085 the encrypted data type: this is a more sensible place to put it and it
5086 allows the PKCS#12 code to be tidied up that duplicated this
5090 *) Changed obj_dat.pl script so it takes its input and output files on
5091 the command line. This should avoid shell escape redirection problems
5095 *) Initial support for certificate extension requests, these are included
5096 in things like Xenroll certificate requests. Included functions to allow
5097 extensions to be obtained and added.
5100 *) -crlf option to s_client and s_server for sending newlines as
5101 CRLF (as required by many protocols).
5104 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5106 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5107 [Ralf S. Engelschall]
5109 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5110 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5112 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5116 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5117 DH parameters/keys (q is lost during that conversion, but the resulting
5118 DH parameters contain its length).
5120 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5121 much faster than DH_generate_parameters (which creates parameters
5122 where p = 2*q + 1), and also the smaller q makes DH computations
5123 much more efficient (160-bit exponentiation instead of 1024-bit
5124 exponentiation); so this provides a convenient way to support DHE
5125 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5126 utter importance to use
5127 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5129 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5130 when such DH parameters are used, because otherwise small subgroup
5131 attacks may become possible!
5134 *) Avoid memory leak in i2d_DHparams.
5137 *) Allow the -k option to be used more than once in the enc program:
5138 this allows the same encrypted message to be read by multiple recipients.
5141 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5142 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5143 it will always use the numerical form of the OID, even if it has a short
5147 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5148 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5149 otherwise bn_mod_exp was called. In the case of hardware keys for example
5150 no private key components need be present and it might store extra data
5151 in the RSA structure, which cannot be accessed from bn_mod_exp.
5152 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5153 private key operations.
5156 *) Added support for SPARC Linux.
5159 *) pem_password_cb function type incompatibly changed from
5160 typedef int pem_password_cb(char *buf, int size, int rwflag);
5162 ....(char *buf, int size, int rwflag, void *userdata);
5163 so that applications can pass data to their callbacks:
5164 The PEM[_ASN1]_{read,write}... functions and macros now take an
5165 additional void * argument, which is just handed through whenever
5166 the password callback is called.
5167 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5169 New function SSL_CTX_set_default_passwd_cb_userdata.
5171 Compatibility note: As many C implementations push function arguments
5172 onto the stack in reverse order, the new library version is likely to
5173 interoperate with programs that have been compiled with the old
5174 pem_password_cb definition (PEM_whatever takes some data that
5175 happens to be on the stack as its last argument, and the callback
5176 just ignores this garbage); but there is no guarantee whatsoever that
5179 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5180 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5181 problems not only on Windows, but also on some Unix platforms.
5182 To avoid problematic command lines, these definitions are now in an
5183 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5184 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5187 *) MIPS III/IV assembler module is reimplemented.
5190 *) More DES library cleanups: remove references to srand/rand and
5191 delete an unused file.
5194 *) Add support for the the free Netwide assembler (NASM) under Win32,
5195 since not many people have MASM (ml) and it can be hard to obtain.
5196 This is currently experimental but it seems to work OK and pass all
5197 the tests. Check out INSTALL.W32 for info.
5200 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5201 without temporary keys kept an extra copy of the server key,
5202 and connections with temporary keys did not free everything in case
5206 *) New function RSA_check_key and new openssl rsa option -check
5207 for verifying the consistency of RSA keys.
5208 [Ulf Moeller, Bodo Moeller]
5210 *) Various changes to make Win32 compile work:
5211 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5212 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5213 comparison" warnings.
5214 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5217 *) Add a debugging option to PKCS#5 v2 key generation function: when
5218 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5219 derived keys are printed to stderr.
5222 *) Copy the flags in ASN1_STRING_dup().
5223 [Roman E. Pavlov <pre@mo.msk.ru>]
5225 *) The x509 application mishandled signing requests containing DSA
5226 keys when the signing key was also DSA and the parameters didn't match.
5228 It was supposed to omit the parameters when they matched the signing key:
5229 the verifying software was then supposed to automatically use the CA's
5230 parameters if they were absent from the end user certificate.
5232 Omitting parameters is no longer recommended. The test was also
5233 the wrong way round! This was probably due to unusual behaviour in
5234 EVP_cmp_parameters() which returns 1 if the parameters match.
5235 This meant that parameters were omitted when they *didn't* match and
5236 the certificate was useless. Certificates signed with 'ca' didn't have
5238 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5240 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5241 The interface is as follows:
5242 Applications can use
5243 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5244 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5245 "off" is now the default.
5246 The library internally uses
5247 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5248 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5249 to disable memory-checking temporarily.
5251 Some inconsistent states that previously were possible (and were
5252 even the default) are now avoided.
5254 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5255 with each memory chunk allocated; this is occasionally more helpful
5256 than just having a counter.
5258 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5260 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5264 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5265 which largely parallels "options", but is for changing API behaviour,
5266 whereas "options" are about protocol behaviour.
5267 Initial "mode" flags are:
5269 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5270 a single record has been written.
5271 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5272 retries use the same buffer location.
5273 (But all of the contents must be
5277 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5280 *) Fix problems with no-hmac etc.
5281 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5283 *) New functions RSA_get_default_method(), RSA_set_method() and
5284 RSA_get_method(). These allows replacement of RSA_METHODs without having
5285 to mess around with the internals of an RSA structure.
5288 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5289 Also really enable memory leak checks in openssl.c and in some
5291 [Chad C. Mulligan, Bodo Moeller]
5293 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5294 up the length of negative integers. This has now been simplified to just
5295 store the length when it is first determined and use it later, rather
5296 than trying to keep track of where data is copied and updating it to
5298 [Steve Henson, reported by Brien Wheeler
5299 <bwheeler@authentica-security.com>]
5301 *) Add a new function PKCS7_signatureVerify. This allows the verification
5302 of a PKCS#7 signature but with the signing certificate passed to the
5303 function itself. This contrasts with PKCS7_dataVerify which assumes the
5304 certificate is present in the PKCS#7 structure. This isn't always the
5305 case: certificates can be omitted from a PKCS#7 structure and be
5306 distributed by "out of band" means (such as a certificate database).
5309 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5310 function prototypes in pem.h, also change util/mkdef.pl to add the
5311 necessary function names.
5314 *) mk1mf.pl (used by Windows builds) did not properly read the
5315 options set by Configure in the top level Makefile, and Configure
5316 was not even able to write more than one option correctly.
5317 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5320 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5321 file to be loaded from a BIO or FILE pointer. The BIO version will
5322 for example allow memory BIOs to contain config info.
5325 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5326 Whoever hopes to achieve shared-library compatibility across versions
5327 must use this, not the compile-time macro.
5328 (Exercise 0.9.4: Which is the minimum library version required by
5330 Note: All this applies only to multi-threaded programs, others don't
5334 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5335 through a BIO pair triggered the default case, i.e.
5336 SSLerr(...,SSL_R_UNKNOWN_STATE).
5339 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5340 can use the SSL library even if none of the specific BIOs is
5344 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5345 for the encoded length.
5346 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5348 *) Add initial documentation of the X509V3 functions.
5351 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5352 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5353 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5354 secure PKCS#8 private key format with a high iteration count.
5357 *) Fix determination of Perl interpreter: A perl or perl5
5358 _directory_ in $PATH was also accepted as the interpreter.
5359 [Ralf S. Engelschall]
5361 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5362 wrong with it but it was very old and did things like calling
5363 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5367 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5368 to use the new extension code.
5371 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5372 with macros. This should make it easier to change their form, add extra
5373 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5377 *) Add to configuration table a new entry that can specify an alternative
5378 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5379 according to Mark Crispin <MRC@Panda.COM>.
5383 *) DES CBC did not update the IV. Weird.
5386 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5387 Changing the behaviour of the former might break existing programs --
5388 where IV updating is needed, des_ncbc_encrypt can be used.
5391 *) When bntest is run from "make test" it drives bc to check its
5392 calculations, as well as internally checking them. If an internal check
5393 fails, it needs to cause bc to give a non-zero result or make test carries
5394 on without noticing the failure. Fixed.
5397 *) DES library cleanups.
5400 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5401 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5402 ciphers. NOTE: although the key derivation function has been verified
5403 against some published test vectors it has not been extensively tested
5404 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5408 *) Instead of "mkdir -p", which is not fully portable, use new
5409 Perl script "util/mkdir-p.pl".
5412 *) Rewrite the way password based encryption (PBE) is handled. It used to
5413 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5414 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5415 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5416 the 'parameter' field of the AlgorithmIdentifier is passed to the
5417 underlying key generation function so it must do its own ASN1 parsing.
5418 This has also changed the EVP_PBE_CipherInit() function which now has a
5419 'parameter' argument instead of literal salt and iteration count values
5420 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5423 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5424 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5425 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5426 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5427 value was just used as a "magic string" and not used directly its
5428 value doesn't matter.
5431 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5435 *) "linux-sparc64" configuration (ultrapenguin).
5436 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5437 "linux-sparc" configuration.
5438 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5440 *) config now generates no-xxx options for missing ciphers.
5443 *) Support the EBCDIC character set (work in progress).
5444 File ebcdic.c not yet included because it has a different license.
5445 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5447 *) Support BS2000/OSD-POSIX.
5448 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5450 *) Make callbacks for key generation use void * instead of char *.
5453 *) Make S/MIME samples compile (not yet tested).
5456 *) Additional typesafe stacks.
5459 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5463 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5465 *) New configuration variant "sco5-gcc".
5467 *) Updated some demos.
5468 [Sean O Riordain, Wade Scholine]
5470 *) Add missing BIO_free at exit of pkcs12 application.
5473 *) Fix memory leak in conf.c.
5476 *) Updates for Win32 to assembler version of MD5.
5479 *) Set #! path to perl in apps/der_chop to where we found it
5480 instead of using a fixed path.
5483 *) SHA library changes for irix64-mips4-cc.
5486 *) Improvements for VMS support.
5490 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5492 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5493 This also avoids the problems with SC4.2 and unpatched SC5.
5494 [Andy Polyakov <appro@fy.chalmers.se>]
5496 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5497 These are required because of the typesafe stack would otherwise break
5498 existing code. If old code used a structure member which used to be STACK
5499 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5500 sk_num or sk_value it would produce an error because the num, data members
5501 are not present in STACK_OF. Now it just produces a warning. sk_set
5502 replaces the old method of assigning a value to sk_value
5503 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5504 that does this will no longer work (and should use sk_set instead) but
5505 this could be regarded as a "questionable" behaviour anyway.
5508 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5509 correctly handle encrypted S/MIME data.
5512 *) Change type of various DES function arguments from des_cblock
5513 (which means, in function argument declarations, pointer to char)
5514 to des_cblock * (meaning pointer to array with 8 char elements),
5515 which allows the compiler to do more typechecking; it was like
5516 that back in SSLeay, but with lots of ugly casts.
5518 Introduce new type const_des_cblock.
5521 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5522 problems: find RecipientInfo structure that matches recipient certificate
5523 and initialise the ASN1 structures properly based on passed cipher.
5526 *) Belatedly make the BN tests actually check the results.
5529 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5530 to and from BNs: it was completely broken. New compilation option
5531 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5532 key elements as negative integers.
5535 *) Reorganize and speed up MD5.
5536 [Andy Polyakov <appro@fy.chalmers.se>]
5539 [Richard Levitte <richard@levitte.org>]
5541 *) New option -out to asn1parse to allow the parsed structure to be
5542 output to a file. This is most useful when combined with the -strparse
5543 option to examine the output of things like OCTET STRINGS.
5546 *) Make SSL library a little more fool-proof by not requiring any longer
5547 that SSL_set_{accept,connect}_state be called before
5548 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5549 in many applications because usually everything *appeared* to work as
5550 intended anyway -- now it really works as intended).
5553 *) Move openssl.cnf out of lib/.
5556 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5557 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5558 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5559 [Ralf S. Engelschall]
5561 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5562 handle PKCS#7 enveloped data properly.
5563 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5565 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5566 copying pointers. The cert_st handling is changed by this in
5567 various ways (and thus what used to be known as ctx->default_cert
5568 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5569 any longer when s->cert does not give us what we need).
5570 ssl_cert_instantiate becomes obsolete by this change.
5571 As soon as we've got the new code right (possibly it already is?),
5572 we have solved a couple of bugs of the earlier code where s->cert
5573 was used as if it could not have been shared with other SSL structures.
5575 Note that using the SSL API in certain dirty ways now will result
5576 in different behaviour than observed with earlier library versions:
5577 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5578 does not influence s as it used to.
5580 In order to clean up things more thoroughly, inside SSL_SESSION
5581 we don't use CERT any longer, but a new structure SESS_CERT
5582 that holds per-session data (if available); currently, this is
5583 the peer's certificate chain and, for clients, the server's certificate
5584 and temporary key. CERT holds only those values that can have
5585 meaningful defaults in an SSL_CTX.
5588 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5589 from the internal representation. Various PKCS#7 fixes: remove some
5590 evil casts and set the enc_dig_alg field properly based on the signing
5594 *) Allow PKCS#12 password to be set from the command line or the
5595 environment. Let 'ca' get its config file name from the environment
5596 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5600 *) Allow certificate policies extension to use an IA5STRING for the
5601 organization field. This is contrary to the PKIX definition but
5602 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5606 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5607 without disallowing inline assembler and the like for non-pedantic builds.
5610 *) Support Borland C++ builder.
5611 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5616 *) SHA-1 cleanups and performance enhancements.
5617 [Andy Polyakov <appro@fy.chalmers.se>]
5619 *) Sparc v8plus assembler for the bignum library.
5620 [Andy Polyakov <appro@fy.chalmers.se>]
5622 *) Accept any -xxx and +xxx compiler options in Configure.
5625 *) Update HPUX configuration.
5628 *) Add missing sk_<type>_unshift() function to safestack.h
5629 [Ralf S. Engelschall]
5631 *) New function SSL_CTX_use_certificate_chain_file that sets the
5632 "extra_cert"s in addition to the certificate. (This makes sense
5633 only for "PEM" format files, as chains as a whole are not
5637 *) Support verify_depth from the SSL API.
5638 x509_vfy.c had what can be considered an off-by-one-error:
5639 Its depth (which was not part of the external interface)
5640 was actually counting the number of certificates in a chain;
5641 now it really counts the depth.
5644 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5645 instead of X509err, which often resulted in confusing error
5646 messages since the error codes are not globally unique
5647 (e.g. an alleged error in ssl3_accept when a certificate
5648 didn't match the private key).
5650 *) New function SSL_CTX_set_session_id_context that allows to set a default
5651 value (so that you don't need SSL_set_session_id_context for each
5652 connection using the SSL_CTX).
5655 *) OAEP decoding bug fix.
5658 *) Support INSTALL_PREFIX for package builders, as proposed by
5662 *) New Configure options "threads" and "no-threads". For systems
5663 where the proper compiler options are known (currently Solaris
5664 and Linux), "threads" is the default.
5667 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5670 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5671 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5672 such as /usr/local/bin.
5675 *) "make linux-shared" to build shared libraries.
5676 [Niels Poppe <niels@netbox.org>]
5678 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5681 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5682 extension adding in x509 utility.
5685 *) Remove NOPROTO sections and error code comments.
5688 *) Partial rewrite of the DEF file generator to now parse the ANSI
5692 *) New Configure options --prefix=DIR and --openssldir=DIR.
5695 *) Complete rewrite of the error code script(s). It is all now handled
5696 by one script at the top level which handles error code gathering,
5697 header rewriting and C source file generation. It should be much better
5698 than the old method: it now uses a modified version of Ulf's parser to
5699 read the ANSI prototypes in all header files (thus the old K&R definitions
5700 aren't needed for error creation any more) and do a better job of
5701 translating function codes into names. The old 'ASN1 error code imbedded
5702 in a comment' is no longer necessary and it doesn't use .err files which
5703 have now been deleted. Also the error code call doesn't have to appear all
5704 on one line (which resulted in some large lines...).
5707 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5710 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5711 0 (which usually indicates a closed connection), but continue reading.
5714 *) Fix some race conditions.
5717 *) Add support for CRL distribution points extension. Add Certificate
5718 Policies and CRL distribution points documentation.
5721 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5724 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5725 8 of keying material. Merlin has also confirmed interop with this fix
5726 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5727 [Merlin Hughes <merlin@baltimore.ie>]
5729 *) Fix lots of warnings.
5730 [Richard Levitte <levitte@stacken.kth.se>]
5732 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5733 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5734 [Richard Levitte <levitte@stacken.kth.se>]
5736 *) Fix problems with sizeof(long) == 8.
5737 [Andy Polyakov <appro@fy.chalmers.se>]
5739 *) Change functions to ANSI C.
5742 *) Fix typos in error codes.
5743 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5745 *) Remove defunct assembler files from Configure.
5748 *) SPARC v8 assembler BIGNUM implementation.
5749 [Andy Polyakov <appro@fy.chalmers.se>]
5751 *) Support for Certificate Policies extension: both print and set.
5752 Various additions to support the r2i method this uses.
5755 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5756 return a const string when you are expecting an allocated buffer.
5759 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5760 types DirectoryString and DisplayText.
5763 *) Add code to allow r2i extensions to access the configuration database,
5764 add an LHASH database driver and add several ctx helper functions.
5767 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5768 fail when they extended the size of a BIGNUM.
5771 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5772 support typesafe stack.
5775 *) Fix typo in SSL_[gs]et_options().
5776 [Nils Frostberg <nils@medcom.se>]
5778 *) Delete various functions and files that belonged to the (now obsolete)
5779 old X509V3 handling code.
5782 *) New Configure option "rsaref".
5785 *) Don't auto-generate pem.h.
5788 *) Introduce type-safe ASN.1 SETs.
5791 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5792 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5794 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5795 that links with OpenSSL (well at least cause lots of warnings), but fear
5796 not: the conversion is trivial, and it eliminates loads of evil casts. A
5797 few STACKed things have been converted already. Feel free to convert more.
5798 In the fullness of time, I'll do away with the STACK type altogether.
5801 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5802 specified in <certfile> by updating the entry in the index.txt file.
5803 This way one no longer has to edit the index.txt file manually for
5804 revoking a certificate. The -revoke option does the gory details now.
5805 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5807 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5808 `-text' option at all and this way the `-noout -text' combination was
5809 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5810 [Ralf S. Engelschall]
5812 *) Make sure a corresponding plain text error message exists for the
5813 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5814 verify callback function determined that a certificate was revoked.
5815 [Ralf S. Engelschall]
5817 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5818 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5819 all available cipers including rc5, which was forgotten until now.
5820 In order to let the testing shell script know which algorithms
5821 are available, a new (up to now undocumented) command
5822 "openssl list-cipher-commands" is used.
5825 *) Bugfix: s_client occasionally would sleep in select() when
5826 it should have checked SSL_pending() first.
5829 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5830 the raw DSA values prior to ASN.1 encoding.
5833 *) Tweaks to Configure
5834 [Niels Poppe <niels@netbox.org>]
5836 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5840 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5843 *) New config option to avoid instructions that are illegal on the 80386.
5844 The default code is faster, but requires at least a 486.
5847 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5848 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5849 same as SSL2_VERSION anyway.
5852 *) New "-showcerts" option for s_client.
5855 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5856 application. Various cleanups and fixes.
5859 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5860 modify error routines to work internally. Add error codes and PBE init
5861 to library startup routines.
5864 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5865 packing functions to asn1 and evp. Changed function names and error
5866 codes along the way.
5869 *) PKCS12 integration: and so it begins... First of several patches to
5870 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5871 objects to objects.h
5874 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5875 and display support for Thawte strong extranet extension.
5878 *) Add LinuxPPC support.
5879 [Jeff Dubrule <igor@pobox.org>]
5881 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5882 bn_div_words in alpha.s.
5883 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5885 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5886 OAEP isn't supported when OpenSSL is built with RSAref.
5887 [Ulf Moeller <ulf@fitug.de>]
5889 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5890 so they no longer are missing under -DNOPROTO.
5891 [Soren S. Jorvang <soren@t.dk>]
5894 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5896 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5897 doesn't work when the session is reused. Coming soon!
5900 *) Fix a security hole, that allows sessions to be reused in the wrong
5901 context thus bypassing client cert protection! All software that uses
5902 client certs and session caches in multiple contexts NEEDS PATCHING to
5903 allow session reuse! A fuller solution is in the works.
5904 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5906 *) Some more source tree cleanups (removed obsolete files
5907 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5908 permission on "config" script to be executable) and a fix for the INSTALL
5910 [Ulf Moeller <ulf@fitug.de>]
5912 *) Remove some legacy and erroneous uses of malloc, free instead of
5914 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5916 *) Make rsa_oaep_test return non-zero on error.
5917 [Ulf Moeller <ulf@fitug.de>]
5919 *) Add support for native Solaris shared libraries. Configure
5920 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5921 if someone would make that last step automatic.
5922 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5924 *) ctx_size was not built with the right compiler during "make links". Fixed.
5927 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5928 except NULL ciphers". This means the default cipher list will no longer
5929 enable NULL ciphers. They need to be specifically enabled e.g. with
5930 the string "DEFAULT:eNULL".
5933 *) Fix to RSA private encryption routines: if p < q then it would
5934 occasionally produce an invalid result. This will only happen with
5935 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5938 *) Be less restrictive and allow also `perl util/perlpath.pl
5939 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5940 because this way one can also use an interpreter named `perl5' (which is
5941 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5942 installed as `perl').
5943 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5945 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5946 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5948 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5949 advapi32.lib to Win32 build and change the pem test comparision
5950 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5951 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5952 and crypto/des/ede_cbcm_enc.c.
5955 *) DES quad checksum was broken on big-endian architectures. Fixed.
5958 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5959 Win32 test batch file so it (might) work again. The Win32 test batch file
5960 is horrible: I feel ill....
5963 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5964 in e_os.h. Audit of header files to check ANSI and non ANSI
5965 sections: 10 functions were absent from non ANSI section and not exported
5966 from Windows DLLs. Fixed up libeay.num for new functions.
5969 *) Make `openssl version' output lines consistent.
5970 [Ralf S. Engelschall]
5972 *) Fix Win32 symbol export lists for BIO functions: Added
5973 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5974 to ms/libeay{16,32}.def.
5975 [Ralf S. Engelschall]
5977 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5978 fine under Unix and passes some trivial tests I've now added. But the
5979 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5980 added to make sure no one expects that this stuff really works in the
5981 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5982 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5984 [Ralf S. Engelschall]
5986 *) Fix the generation of two part addresses in perl.
5987 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5989 *) Add config entry for Linux on MIPS.
5990 [John Tobey <jtobey@channel1.com>]
5992 *) Make links whenever Configure is run, unless we are on Windoze.
5995 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5996 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6000 *) Add a useful kludge to allow package maintainers to specify compiler and
6001 other platforms details on the command line without having to patch the
6002 Configure script everytime: One now can use ``perl Configure
6003 <id>:<details>'', i.e. platform ids are allowed to have details appended
6004 to them (seperated by colons). This is treated as there would be a static
6005 pre-configured entry in Configure's %table under key <id> with value
6006 <details> and ``perl Configure <id>'' is called. So, when you want to
6007 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6008 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6009 now, which overrides the FreeBSD-elf entry on-the-fly.
6010 [Ralf S. Engelschall]
6012 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6015 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6016 on the `perl Configure ...' command line. This way one can compile
6017 OpenSSL libraries with Position Independent Code (PIC) which is needed
6018 for linking it into DSOs.
6019 [Ralf S. Engelschall]
6021 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6025 *) Cleaned up the LICENSE document: The official contact for any license
6026 questions now is the OpenSSL core team under openssl-core@openssl.org.
6027 And add a paragraph about the dual-license situation to make sure people
6028 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6029 to the OpenSSL toolkit.
6030 [Ralf S. Engelschall]
6032 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6033 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6034 Additonally cleaned up the `make links' target: Remove unnecessary
6035 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6036 to speed processing and no longer clutter the display with confusing
6037 stuff. Instead only the actually done links are displayed.
6038 [Ralf S. Engelschall]
6040 *) Permit null encryption ciphersuites, used for authentication only. It used
6041 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6042 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6046 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6047 signed attributes when verifying signatures (this would break them),
6048 the detached data encoding was wrong and public keys obtained using
6049 X509_get_pubkey() weren't freed.
6052 *) Add text documentation for the BUFFER functions. Also added a work around
6053 to a Win95 console bug. This was triggered by the password read stuff: the
6054 last character typed gets carried over to the next fread(). If you were
6055 generating a new cert request using 'req' for example then the last
6056 character of the passphrase would be CR which would then enter the first
6060 *) Added the new `Includes OpenSSL Cryptography Software' button as
6061 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6062 button and can be used by applications based on OpenSSL to show the
6063 relationship to the OpenSSL project.
6064 [Ralf S. Engelschall]
6066 *) Remove confusing variables in function signatures in files
6067 ssl/ssl_lib.c and ssl/ssl.h.
6068 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6070 *) Don't install bss_file.c under PREFIX/include/
6071 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6073 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6074 functions that return function pointers and has support for NT specific
6075 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6076 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6077 unsigned to signed types: this was killing the Win32 compile.
6080 *) Add new certificate file to stack functions,
6081 SSL_add_dir_cert_subjects_to_stack() and
6082 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6083 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6084 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6085 This means that Apache-SSL and similar packages don't have to mess around
6086 to add as many CAs as they want to the preferred list.
6089 *) Experiment with doxygen documentation. Currently only partially applied to
6091 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6092 openssl.doxy as the configuration file.
6095 *) Get rid of remaining C++-style comments which strict C compilers hate.
6096 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6098 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6099 compiled in by default: it has problems with large keys.
6102 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6103 DH private keys and/or callback functions which directly correspond to
6104 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6105 is needed for applications which have to configure certificates on a
6106 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6108 For the RSA certificate situation is makes no difference, but
6109 for the DSA certificate situation this fixes the "no shared cipher"
6110 problem where the OpenSSL cipher selection procedure failed because the
6111 temporary keys were not overtaken from the context and the API provided
6112 no way to reconfigure them.
6113 The new functions now let applications reconfigure the stuff and they
6114 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6115 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6116 non-public-API function ssl_cert_instantiate() is used as a helper
6117 function and also to reduce code redundancy inside ssl_rsa.c.
6118 [Ralf S. Engelschall]
6120 *) Move s_server -dcert and -dkey options out of the undocumented feature
6121 area because they are useful for the DSA situation and should be
6122 recognized by the users.
6123 [Ralf S. Engelschall]
6125 *) Fix the cipher decision scheme for export ciphers: the export bits are
6126 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6127 SSL_EXP_MASK. So, the original variable has to be used instead of the
6128 already masked variable.
6129 [Richard Levitte <levitte@stacken.kth.se>]
6131 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6132 [Richard Levitte <levitte@stacken.kth.se>]
6134 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6135 from `int' to `unsigned int' because it's a length and initialized by
6136 EVP_DigestFinal() which expects an `unsigned int *'.
6137 [Richard Levitte <levitte@stacken.kth.se>]
6139 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6140 script. Instead use the usual Shell->Perl transition trick.
6141 [Ralf S. Engelschall]
6143 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6144 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6145 -noout -modulus' as it's already the case for `openssl rsa -noout
6146 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6147 currently the public key is printed (a decision which was already done by
6148 `openssl dsa -modulus' in the past) which serves a similar purpose.
6149 Additionally the NO_RSA no longer completely removes the whole -modulus
6150 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6152 [Ralf S. Engelschall]
6154 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6155 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6156 [Arne Ansper <arne@ats.cyber.ee>]
6158 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6159 to be added. Now both 'req' and 'ca' can use new objects defined in the
6163 *) Add cool BIO that does syslog (or event log on NT).
6164 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6166 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6167 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6168 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6169 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6172 *) Add preliminary config info for new extension code.
6175 *) Make RSA_NO_PADDING really use no padding.
6176 [Ulf Moeller <ulf@fitug.de>]
6178 *) Generate errors when private/public key check is done.
6181 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6182 for some CRL extensions and new objects added.
6185 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6186 key usage extension and fuller support for authority key id.
6189 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6190 padding method for RSA, which is recommended for new applications in PKCS
6191 #1 v2.0 (RFC 2437, October 1998).
6192 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6193 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6194 against Bleichbacher's attack on RSA.
6195 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6198 *) Updates to the new SSL compression code
6199 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6201 *) Fix so that the version number in the master secret, when passed
6202 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6203 (because the server will not accept higher), that the version number
6204 is 0x03,0x01, not 0x03,0x00
6205 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6207 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6208 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6209 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6212 *) Support for RAW extensions where an arbitrary extension can be
6213 created by including its DER encoding. See apps/openssl.cnf for
6217 *) Make sure latest Perl versions don't interpret some generated C array
6218 code as Perl array code in the crypto/err/err_genc.pl script.
6219 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6221 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6222 not many people have the assembler. Various Win32 compilation fixes and
6223 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6227 *) Modify configure script 'Configure' to automatically create crypto/date.h
6228 file under Win32 and also build pem.h from pem.org. New script
6229 util/mkfiles.pl to create the MINFO file on environments that can't do a
6230 'make files': perl util/mkfiles.pl >MINFO should work.
6233 *) Major rework of DES function declarations, in the pursuit of correctness
6234 and purity. As a result, many evil casts evaporated, and some weirdness,
6235 too. You may find this causes warnings in your code. Zapping your evil
6236 casts will probably fix them. Mostly.
6239 *) Fix for a typo in asn1.h. Bug fix to object creation script
6240 obj_dat.pl. It considered a zero in an object definition to mean
6241 "end of object": none of the objects in objects.h have any zeros
6242 so it wasn't spotted.
6243 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6245 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6246 Masking (CBCM). In the absence of test vectors, the best I have been able
6247 to do is check that the decrypt undoes the encrypt, so far. Send me test
6248 vectors if you have them.
6251 *) Correct calculation of key length for export ciphers (too much space was
6252 allocated for null ciphers). This has not been tested!
6255 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6256 message is now correct (it understands "crypto" and "ssl" on its
6257 command line). There is also now an "update" option. This will update
6258 the util/ssleay.num and util/libeay.num files with any new functions.
6260 perl util/mkdef.pl crypto ssl update
6261 it will update them.
6264 *) Overhauled the Perl interface (perl/*):
6265 - ported BN stuff to OpenSSL's different BN library
6266 - made the perl/ source tree CVS-aware
6267 - renamed the package from SSLeay to OpenSSL (the files still contain
6268 their history because I've copied them in the repository)
6269 - removed obsolete files (the test scripts will be replaced
6270 by better Test::Harness variants in the future)
6271 [Ralf S. Engelschall]
6273 *) First cut for a very conservative source tree cleanup:
6274 1. merge various obsolete readme texts into doc/ssleay.txt
6275 where we collect the old documents and readme texts.
6276 2. remove the first part of files where I'm already sure that we no
6277 longer need them because of three reasons: either they are just temporary
6278 files which were left by Eric or they are preserved original files where
6279 I've verified that the diff is also available in the CVS via "cvs diff
6280 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6281 the crypto/md/ stuff).
6282 [Ralf S. Engelschall]
6284 *) More extension code. Incomplete support for subject and issuer alt
6285 name, issuer and authority key id. Change the i2v function parameters
6286 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6287 what that's for :-) Fix to ASN1 macro which messed up
6288 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6291 *) Preliminary support for ENUMERATED type. This is largely copied from the
6295 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6296 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6298 *) Make sure `make rehash' target really finds the `openssl' program.
6299 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6301 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6302 like to hear about it if this slows down other processors.
6305 *) Add CygWin32 platform information to Configure script.
6306 [Alan Batie <batie@aahz.jf.intel.com>]
6308 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6309 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6311 *) New program nseq to manipulate netscape certificate sequences
6314 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6318 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6319 but the BN code had some problems that would cause failures when
6320 doing certificate verification and some other functions.
6321 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6323 *) Add ASN1 and PEM code to support netscape certificate sequences.
6326 *) Add ASN1 and PEM code to support netscape certificate sequences.
6329 *) Add several PKIX and private extended key usage OIDs.
6332 *) Modify the 'ca' program to handle the new extension code. Modify
6333 openssl.cnf for new extension format, add comments.
6336 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6337 and add a sample to openssl.cnf so req -x509 now adds appropriate
6341 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6342 error code, add initial support to X509_print() and x509 application.
6345 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6346 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6347 stuff is currently isolated and isn't even compiled yet.
6350 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6351 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6352 Removed the versions check from X509 routines when loading extensions:
6353 this allows certain broken certificates that don't set the version
6354 properly to be processed.
6357 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6358 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6359 can still be regenerated with "make depend".
6362 *) Spelling mistake in C version of CAST-128.
6363 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6365 *) Changes to the error generation code. The perl script err-code.pl
6366 now reads in the old error codes and retains the old numbers, only
6367 adding new ones if necessary. It also only changes the .err files if new
6368 codes are added. The makefiles have been modified to only insert errors
6369 when needed (to avoid needlessly modifying header files). This is done
6370 by only inserting errors if the .err file is newer than the auto generated
6371 C file. To rebuild all the error codes from scratch (the old behaviour)
6372 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6373 or delete all the .err files.
6376 *) CAST-128 was incorrectly implemented for short keys. The C version has
6377 been fixed, but is untested. The assembler versions are also fixed, but
6378 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6379 to regenerate it if needed.
6380 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6381 Hagino <itojun@kame.net>]
6383 *) File was opened incorrectly in randfile.c.
6384 [Ulf Möller <ulf@fitug.de>]
6386 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6387 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6388 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6389 al: it's just almost always a UTCTime. Note this patch adds new error
6390 codes so do a "make errors" if there are problems.
6393 *) Correct Linux 1 recognition in config.
6394 [Ulf Möller <ulf@fitug.de>]
6396 *) Remove pointless MD5 hash when using DSA keys in ca.
6397 [Anonymous <nobody@replay.com>]
6399 *) Generate an error if given an empty string as a cert directory. Also
6400 generate an error if handed NULL (previously returned 0 to indicate an
6401 error, but didn't set one).
6402 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6404 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6407 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6408 parameters. This was causing a warning which killed off the Win32 compile.
6411 *) Remove C++ style comments from crypto/bn/bn_local.h.
6412 [Neil Costigan <neil.costigan@celocom.com>]
6414 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6415 based on a text string, looking up short and long names and finally
6416 "dot" format. The "dot" format stuff didn't work. Added new function
6417 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6418 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6419 OID is not part of the table.
6422 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6423 X509_LOOKUP_by_alias().
6426 *) Sort openssl functions by name.
6429 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6430 encryption from sample DSA keys (in case anyone is interested the password
6434 *) Make _all_ *_free functions accept a NULL pointer.
6435 [Frans Heymans <fheymans@isaserver.be>]
6437 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6439 [Anonymous <nobody@replay.com>]
6441 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6442 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6444 *) Don't blow it for numeric -newkey arguments to apps/req.
6445 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6447 *) Temp key "for export" tests were wrong in s3_srvr.c.
6448 [Anonymous <nobody@replay.com>]
6450 *) Add prototype for temp key callback functions
6451 SSL_CTX_set_tmp_{rsa,dh}_callback().
6454 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6455 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6458 *) X509_name_add_entry() freed the wrong thing after an error.
6459 [Arne Ansper <arne@ats.cyber.ee>]
6461 *) rsa_eay.c would attempt to free a NULL context.
6462 [Arne Ansper <arne@ats.cyber.ee>]
6464 *) BIO_s_socket() had a broken should_retry() on Windoze.
6465 [Arne Ansper <arne@ats.cyber.ee>]
6467 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6468 [Arne Ansper <arne@ats.cyber.ee>]
6470 *) Make sure the already existing X509_STORE->depth variable is initialized
6471 in X509_STORE_new(), but document the fact that this variable is still
6472 unused in the certificate verification process.
6473 [Ralf S. Engelschall]
6475 *) Fix the various library and apps files to free up pkeys obtained from
6476 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6479 *) Fix reference counting in X509_PUBKEY_get(). This makes
6480 demos/maurice/example2.c work, amongst others, probably.
6481 [Steve Henson and Ben Laurie]
6483 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6484 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6485 are no longer created. This way we have a single and consistent command
6486 line interface `openssl <command>', similar to `cvs <command>'.
6487 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6489 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6490 BIT STRING wrapper always have zero unused bits.
6493 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6496 *) Make the top-level INSTALL documentation easier to understand.
6499 *) Makefiles updated to exit if an error occurs in a sub-directory
6500 make (including if user presses ^C) [Paul Sutton]
6502 *) Make Montgomery context stuff explicit in RSA data structure.
6505 *) Fix build order of pem and err to allow for generated pem.h.
6508 *) Fix renumbering bug in X509_NAME_delete_entry().
6511 *) Enhanced the err-ins.pl script so it makes the error library number
6512 global and can add a library name. This is needed for external ASN1 and
6513 other error libraries.
6516 *) Fixed sk_insert which never worked properly.
6519 *) Fix ASN1 macros so they can handle indefinite length construted
6520 EXPLICIT tags. Some non standard certificates use these: they can now
6524 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6525 into a single doc/ssleay.txt bundle. This way the information is still
6526 preserved but no longer messes up this directory. Now it's new room for
6527 the new set of documenation files.
6528 [Ralf S. Engelschall]
6530 *) SETs were incorrectly DER encoded. This was a major pain, because they
6531 shared code with SEQUENCEs, which aren't coded the same. This means that
6532 almost everything to do with SETs or SEQUENCEs has either changed name or
6533 number of arguments.
6534 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6536 *) Fix test data to work with the above.
6539 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6540 was already fixed by Eric for 0.9.1 it seems.
6541 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6543 *) Autodetect FreeBSD3.
6546 *) Fix various bugs in Configure. This affects the following platforms:
6550 unixware-2.0-pentium
6554 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6555 before they are needed.
6558 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6562 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6564 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6565 changed SSLeay to OpenSSL in version strings.
6566 [Ralf S. Engelschall]
6568 *) Some fixups to the top-level documents.
6571 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6572 because the symlink to include/ was missing.
6573 [Ralf S. Engelschall]
6575 *) Incorporated the popular no-RSA/DSA-only patches
6576 which allow to compile a RSA-free SSLeay.
6577 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6579 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6580 when "ssleay" is still not found.
6581 [Ralf S. Engelschall]
6583 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6584 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6586 *) Updated the README file.
6587 [Ralf S. Engelschall]
6589 *) Added various .cvsignore files in the CVS repository subdirs
6590 to make a "cvs update" really silent.
6591 [Ralf S. Engelschall]
6593 *) Recompiled the error-definition header files and added
6594 missing symbols to the Win32 linker tables.
6595 [Ralf S. Engelschall]
6597 *) Cleaned up the top-level documents;
6598 o new files: CHANGES and LICENSE
6599 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6600 o merged COPYRIGHT into LICENSE
6601 o removed obsolete TODO file
6602 o renamed MICROSOFT to INSTALL.W32
6603 [Ralf S. Engelschall]
6605 *) Removed dummy files from the 0.9.1b source tree:
6606 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6607 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6608 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6609 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6610 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6611 [Ralf S. Engelschall]
6613 *) Added various platform portability fixes.
6616 *) The Genesis of the OpenSSL rpject:
6617 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6618 Young and Tim J. Hudson created while they were working for C2Net until
6620 [The OpenSSL Project]
6623 Changes between 0.9.0b and 0.9.1b [not released]
6625 *) Updated a few CA certificates under certs/
6628 *) Changed some BIGNUM api stuff.
6631 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6632 DGUX x86, Linux Alpha, etc.
6635 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6636 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6640 *) Add -strparse option to asn1pars program which parses nested
6642 [Dr Stephen Henson <shenson@bigfoot.com>]
6644 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6647 *) DSA fix for "ca" program.
6650 *) Added "-genkey" option to "dsaparam" program.
6653 *) Added RIPE MD160 (rmd160) message digest.
6656 *) Added -a (all) option to "ssleay version" command.
6659 *) Added PLATFORM define which is the id given to Configure.
6662 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6665 *) Extended the ASN.1 parser routines.
6668 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6671 *) Added a BN_CTX to the BN library.
6674 *) Fixed the weak key values in DES library
6677 *) Changed API in EVP library for cipher aliases.
6680 *) Added support for RC2/64bit cipher.
6683 *) Converted the lhash library to the crypto/mem.c functions.
6686 *) Added more recognized ASN.1 object ids.
6689 *) Added more RSA padding checks for SSL/TLS.
6692 *) Added BIO proxy/filter functionality.
6695 *) Added extra_certs to SSL_CTX which can be used
6696 send extra CA certificates to the client in the CA cert chain sending
6697 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6700 *) Now Fortezza is denied in the authentication phase because
6701 this is key exchange mechanism is not supported by SSLeay at all.
6704 *) Additional PKCS1 checks.
6707 *) Support the string "TLSv1" for all TLS v1 ciphers.
6710 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6711 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6714 *) Fixed a few memory leaks.
6717 *) Fixed various code and comment typos.
6720 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6721 bytes sent in the client random.
6722 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob