5 Changes between 0.9.7 and 0.9.8 [xx XXX xxxx]
7 *) Added an ENGINE that implements RSA by performing private key
8 exponentiations with the GMP library. The conversions to and from
9 GMP's mpz_t format aren't optimised nor are any montgomery forms
10 cached, and on x86 it appears OpenSSL's own performance has caught up.
11 However there are likely to be other architectures where GMP could
12 provide a boost. This ENGINE is not built in by default, but it can be
13 specified at Configure time and should be accompanied by the necessary
15 ./config -DOPENSSL_USE_GMP -lgmp
18 *) "openssl engine" will not display ENGINE/DSO load failure errors when
19 testing availability of engines with "-t" - the old behaviour is
20 produced by increasing the feature's verbosity with "-tt".
23 *) ECDSA routines: under certain error conditions uninitialized BN objects
24 could be freed. Solution: make sure initialization is performed early
25 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
29 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
30 and DH_METHOD (eg. by ENGINE implementations) to override the normal
31 software implementations. For DSA and DH, parameter generation can
32 also be overriden by providing the appropriate method callbacks.
35 *) Change the "progress" mechanism used in key-generation and
36 primality testing to functions that take a new BN_GENCB pointer in
37 place of callback/argument pairs. The new API functions have "_ex"
38 postfixes and the older functions are reimplemented as wrappers for
39 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
40 declarations of the old functions to help (graceful) attempts to
41 migrate to the new functions. Also, the new key-generation API
42 functions operate on a caller-supplied key-structure and return
43 success/failure rather than returning a key or NULL - this is to
44 help make "keygen" another member function of RSA_METHOD etc.
46 Example for using the new callback interface:
48 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
52 BN_GENCB_set(&my_cb, my_callback, my_arg);
54 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
55 /* For the meaning of a, b in calls to my_callback(), see the
56 * documentation of the function that calls the callback.
57 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
58 * my_callback should return 1 if it wants BN_is_prime_ex()
59 * to continue, or 0 to stop.
64 *) Change the ZLIB compression method to be stateful, and make it
65 available to TLS with the number defined in
66 draft-ietf-tls-compression-04.txt.
69 *) Add the ASN.1 structures and functions for CertificatePair, which
70 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
72 CertificatePair ::= SEQUENCE {
73 forward [0] Certificate OPTIONAL,
74 reverse [1] Certificate OPTIONAL,
75 -- at least one of the pair shall be present -- }
77 Also implement the PEM functions to read and write certificate
78 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
80 This needed to be defined, mostly for the sake of the LDAP
81 attribute crossCertificatePair, but may prove useful elsewhere as
85 *) Make it possible to inhibit symlinking of shared libraries in
86 Makefile.shared, for Cygwin's sake.
89 *) Extend the BIGNUM API by creating new macros that behave like
92 void BN_set_sign(BIGNUM *a, int neg);
93 int BN_get_sign(const BIGNUM *a);
95 and avoid the need to access 'a->neg' directly in applications.
96 [Nils Larsch <nla@trustcenter.de>]
98 *) Implement fast modular reduction for pseudo-Mersenne primes
99 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
100 EC_GROUP_new_curve_GFp() will now automatically use this
102 [Nils Larsch <nla@trustcenter.de>]
104 *) Add new lock type (CRYPTO_LOCK_BN).
107 *) Change the ENGINE framework to automatically load engines
108 dynamically from specific directories unless they could be
109 found to already be built in or loaded. Move all the
110 current engines except for the cryptodev one to a new
112 The engines in engines/ are built as shared libraries if
113 the "shared" options was given to ./Configure or ./config.
114 Otherwise, they are inserted in libcrypto.a.
115 /usr/local/ssl/engines is the default directory for dynamic
116 engines, but that can be overriden at configure time through
117 the usual use of --prefix and/or --openssldir, and at run
118 time with the environment variable OPENSSL_ENGINES.
119 [Geoff Thorpe and Richard Levitte]
121 *) Add Makefile.shared, a helper makefile to build shared
122 libraries. Addapt Makefile.org.
125 *) Add version info to Win32 DLLs.
126 [Peter 'Luna' Runestig" <peter@runestig.com>]
128 *) Add new 'medium level' PKCS#12 API. Certificates and keys
129 can be added using this API to created arbitrary PKCS#12
130 files while avoiding the low level API.
132 New options to PKCS12_create(), key or cert can be NULL and
133 will then be omitted from the output file. The encryption
134 algorithm NIDs can be set to -1 for no encryption, the mac
135 iteration count can be set to 0 to omit the mac.
137 Enhance pkcs12 utility by making the -nokeys and -nocerts
138 options work when creating a PKCS#12 file. New option -nomac
139 to omit the mac, NONE can be set for an encryption algorithm.
140 New code is modified to use the enhanced PKCS12_create()
141 instead of the low level API.
144 *) Extend ASN1 encoder to support indefinite length constructed
145 encoding. This can output sequences tags and octet strings in
146 this form. Modify pk7_asn1.c to support indefinite length
147 encoding. This is experimental and needs additional code to
148 be useful, such as an ASN1 bio and some enhanced streaming
151 Extend template encode functionality so that tagging is passed
152 down to the template encoder.
155 *) Let 'openssl req' fail if an argument to '-newkey' is not
156 recognized instead of using RSA as a default.
159 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
160 As these are not official, they are not included in "ALL";
161 the "ECCdraft" ciphersuite group alias can be used to select them.
162 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
164 *) Add ECDH engine support.
165 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
167 *) Add ECDH in new directory crypto/ecdh/.
168 TODO: more general interface (return x coordinate, not its hash)
169 TODO: bug: pad x with leading zeros if necessary
170 [Douglas Stebila (Sun Microsystems Laboratories)]
172 *) Let BN_rand_range() abort with an error after 100 iterations
173 without success (which indicates a broken PRNG).
176 *) Change BN_mod_sqrt() so that it verifies that the input value
177 is really the square of the return value. (Previously,
178 BN_mod_sqrt would show GIGO behaviour.)
181 *) Add named elliptic curves over binary fields from X9.62, SECG,
182 and WAP/WTLS; add OIDs that were still missing.
184 [Sheueling Chang Shantz and Douglas Stebila
185 (Sun Microsystems Laboratories)]
187 *) Extend the EC library for elliptic curves over binary fields
188 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
191 EC_GF2m_simple_method
195 EC_GROUP_new_curve_GF2m
196 EC_GROUP_set_curve_GF2m
197 EC_GROUP_get_curve_GF2m
198 EC_POINT_set_affine_coordinates_GF2m
199 EC_POINT_get_affine_coordinates_GF2m
200 EC_POINT_set_compressed_coordinates_GF2m
202 Point compression for binary fields is disabled by default for
203 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
206 As binary polynomials are represented as BIGNUMs, various members
207 of the EC_GROUP and EC_POINT data structures can be shared
208 between the implementations for prime fields and binary fields;
209 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
210 are essentially identical to their ..._GFp counterparts.
211 (For simplicity, the '..._GFp' prefix has been dropped from
212 various internal method names.)
214 An internal 'field_div' method (similar to 'field_mul' and
215 'field_sqr') has been added; this is used only for binary fields.
217 [Sheueling Chang Shantz and Douglas Stebila
218 (Sun Microsystems Laboratories)]
220 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
221 through methods ('mul', 'precompute_mult').
223 The generic implementations (now internally called 'ec_wNAF_mul'
224 and 'ec_wNAF_precomputed_mult') remain the default if these
225 methods are undefined.
227 [Sheueling Chang Shantz and Douglas Stebila
228 (Sun Microsystems Laboratories)]
230 *) New function EC_GROUP_get_degree, which is defined through
231 EC_METHOD. For curves over prime fields, this returns the bit
232 length of the modulus.
234 [Sheueling Chang Shantz and Douglas Stebila
235 (Sun Microsystems Laboratories)]
237 *) New functions EC_GROUP_dup, EC_POINT_dup.
238 (These simply call ..._new and ..._copy).
240 [Sheueling Chang Shantz and Douglas Stebila
241 (Sun Microsystems Laboratories)]
243 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
244 Polynomials are represented as BIGNUMs (where the sign bit is not
245 used) in the following functions [macros]:
248 BN_GF2m_sub [= BN_GF2m_add]
249 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
250 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
251 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
253 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
254 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
255 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
256 BN_GF2m_cmp [= BN_ucmp]
258 (Note that only the 'mod' functions are actually for fields GF(2^m).
259 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
261 For some functions, an the irreducible polynomial defining a
262 field can be given as an 'unsigned int[]' with strictly
263 decreasing elements giving the indices of those bits that are set;
264 i.e., p[] represents the polynomial
265 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
267 p[0] > p[1] > ... > p[k] = 0.
268 This applies to the following functions:
273 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
274 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
277 BN_GF2m_mod_solve_quad_arr
281 Conversion can be performed by the following functions:
286 bntest.c has additional tests for binary polynomial arithmetic.
288 Two implementations for BN_GF2m_mod_div() are available.
289 The default algorithm simply uses BN_GF2m_mod_inv() and
290 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
291 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
292 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
294 [Sheueling Chang Shantz and Douglas Stebila
295 (Sun Microsystems Laboratories)]
297 *) Add new error code 'ERR_R_DISABLED' that can be used when some
298 functionality is disabled at compile-time.
299 [Douglas Stebila <douglas.stebila@sun.com>]
301 *) Change default behaviour of 'openssl asn1parse' so that more
302 information is visible when viewing, e.g., a certificate:
304 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
305 mode the content of non-printable OCTET STRINGs is output in a
306 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
307 avoid the appearance of a printable string.
308 [Nils Larsch <nla@trustcenter.de>]
310 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
312 EC_GROUP_set_asn1_flag()
313 EC_GROUP_get_asn1_flag()
314 EC_GROUP_set_point_conversion_form()
315 EC_GROUP_get_point_conversion_form()
316 These control ASN1 encoding details:
317 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
318 has been set to OPENSSL_EC_NAMED_CURVE.
319 - Points are encoded in uncompressed form by default; options for
320 asn1_for are as for point2oct, namely
321 POINT_CONVERSION_COMPRESSED
322 POINT_CONVERSION_UNCOMPRESSED
323 POINT_CONVERSION_HYBRID
325 Also add 'seed' and 'seed_len' members to EC_GROUP with access
329 EC_GROUP_get_seed_len()
330 This is used only for ASN1 purposes (so far).
331 [Nils Larsch <nla@trustcenter.de>]
333 *) Add 'field_type' member to EC_METHOD, which holds the NID
334 of the appropriate field type OID. The new function
335 EC_METHOD_get_field_type() returns this value.
336 [Nils Larsch <nla@trustcenter.de>]
343 providing useful interfaces to EC_POINT_point2oct() and
344 EC_POINT_oct2point().
345 [Nils Larsch <nla@trustcenter.de>]
347 *) Change internals of the EC library so that the functions
348 EC_GROUP_set_generator()
349 EC_GROUP_get_generator()
351 EC_GROUP_get_cofactor()
352 are implemented directly in crypto/ec/ec_lib.c and not dispatched
353 to methods, which would lead to unnecessary code duplication when
354 adding different types of curves.
355 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
357 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
358 arithmetic, and such that modified wNAFs are generated
359 (which avoid length expansion in many cases).
362 *) Add a function EC_GROUP_check_discriminant() (defined via
363 EC_METHOD) that verifies that the curve discriminant is non-zero.
365 Add a function EC_GROUP_check() that makes some sanity tests
366 on a EC_GROUP, its generator and order. This includes
367 EC_GROUP_check_discriminant().
368 [Nils Larsch <nla@trustcenter.de>]
370 *) Add ECDSA in new directory crypto/ecdsa/.
372 Add applications 'openssl ecparam' and 'openssl ecdsa'
373 (these are based on 'openssl dsaparam' and 'openssl dsa').
375 ECDSA support is also included in various other files across the
376 library. Most notably,
377 - 'openssl req' now has a '-newkey ecdsa:file' option;
378 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
379 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
380 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
381 them suitable for ECDSA where domain parameters must be
382 extracted before the specific public key;
383 - ECDSA engine support has been added.
384 [Nils Larsch <nla@trustcenter.de>]
386 *) Include some named elliptic curves, and add OIDs from X9.62,
387 SECG, and WAP/WTLS. Each curve can be obtained from the new
389 EC_GROUP_new_by_nid(),
390 and the list of available named curves can be obtained with
391 EC_get_builtin_curves().
392 Also add a 'curve_name' member to EC_GROUP objects, which can be
396 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
398 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
399 was actually never needed) and in BN_mul(). The removal in BN_mul()
400 required a small change in bn_mul_part_recursive() and the addition
401 of the functions bn_cmp_part_words(), bn_sub_part_words() and
402 bn_add_part_words(), which do the same thing as bn_cmp_words(),
403 bn_sub_words() and bn_add_words() except they take arrays with
407 Changes between 0.9.7 and 0.9.7a [XX xxx 2003]
409 *) Add the possibility to build without the ENGINE framework.
410 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
412 *) Under Win32 gmtime() can return NULL: check return value in
413 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
416 *) DSA routines: under certain error conditions uninitialized BN objects
417 could be freed. Solution: make sure initialization is performed early
418 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
419 Nils Larsch <nla@trustcenter.de> via PR#459)
422 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
423 checked on reconnect on the client side, therefore session resumption
424 could still fail with a "ssl session id is different" error. This
425 behaviour is masked when SSL_OP_ALL is used due to
426 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
427 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
431 *) IA-32 assembler support enhancements: unified ELF targets, support
432 for SCO/Caldera platforms, fix for Cygwin shared build.
435 *) Add support for FreeBSD on sparc64. As a consequence, support for
436 FreeBSD on non-x86 processors is separate from x86 processors on
437 the config script, much like the NetBSD support.
438 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
440 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
442 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
443 code (06) was taken as the first octet of the session ID and the last
444 octet was ignored consequently. As a result SSLv2 client side session
445 caching could not have worked due to the session ID mismatch between
447 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
451 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
452 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
456 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
457 seems that in spite of existing for more than a year, many application
458 author have done nothing to provide the necessary callbacks, which
459 means that this particular engine will not work properly anywhere.
460 This is a very unfortunate situation which forces us, in the name
461 of usability, to give the hw_ncipher.c a static lock, which is part
463 NOTE: This is for the 0.9.7 series ONLY. This hack will never
464 appear in 0.9.8 or later. We EXPECT application authors to have
465 dealt properly with this when 0.9.8 is released (unless we actually
466 make such changes in the libcrypto locking code that changes will
467 have to be made anyway).
470 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
471 octets have been read, EOF or an error occurs. Without this change
472 some truncated ASN1 structures will not produce an error.
475 *) Disable Heimdal support, since it hasn't been fully implemented.
476 Still give the possibility to force the use of Heimdal, but with
477 warnings and a request that patches get sent to openssl-dev.
480 *) Add the VC-CE target, introduce the WINCE sysname, and add
481 INSTALL.WCE and appropriate conditionals to make it build.
482 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
484 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
485 cygssl-x.y.z.dll, where x, y and z are the major, minor and
486 edit numbers of the version.
487 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
489 *) Introduce safe string copy and catenation functions
490 (BUF_strlcpy() and BUF_strlcat()).
491 [Ben Laurie (CHATS) and Richard Levitte]
493 *) Avoid using fixed-size buffers for one-line DNs.
496 *) Add BUF_MEM_grow_clean() to avoid information leakage when
497 resizing buffers containing secrets, and use where appropriate.
500 *) Avoid using fixed size buffers for configuration file location.
503 *) Avoid filename truncation for various CA files.
506 *) Use sizeof in preference to magic numbers.
509 *) Avoid filename truncation in cert requests.
512 *) Add assertions to check for (supposedly impossible) buffer
516 *) Don't cache truncated DNS entries in the local cache (this could
517 potentially lead to a spoofing attack).
520 *) Fix various buffers to be large enough for hex/decimal
521 representations in a platform independent manner.
524 *) Add CRYPTO_realloc_clean() to avoid information leakage when
525 resizing buffers containing secrets, and use where appropriate.
528 *) Add BIO_indent() to avoid much slightly worrying code to do
532 *) Convert sprintf()/BIO_puts() to BIO_printf().
535 *) buffer_gets() could terminate with the buffer only half
539 *) Add assertions to prevent user-supplied crypto functions from
540 overflowing internal buffers by having large block sizes, etc.
543 *) New OPENSSL_assert() macro (similar to assert(), but enabled
547 *) Eliminate unused copy of key in RC4.
550 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
553 *) Fix off-by-one error in EGD path.
556 *) If RANDFILE path is too long, ignore instead of truncating.
559 *) Eliminate unused and incorrectly sized X.509 structure
563 *) Eliminate unused and dangerous function knumber().
566 *) Eliminate unused and dangerous structure, KSSL_ERR.
569 *) Protect against overlong session ID context length in an encoded
570 session object. Since these are local, this does not appear to be
574 *) Change from security patch (see 0.9.6e below) that did not affect
575 the 0.9.6 release series:
577 Remote buffer overflow in SSL3 protocol - an attacker could
578 supply an oversized master key in Kerberos-enabled versions.
582 *) Change the SSL kerb5 codes to match RFC 2712.
585 *) Make -nameopt work fully for req and add -reqopt switch.
586 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
588 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
589 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
591 *) Make sure tests can be performed even if the corresponding algorithms
592 have been removed entirely. This was also the last step to make
593 OpenSSL compilable with DJGPP under all reasonable conditions.
594 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
596 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
597 to allow version independent disabling of normally unselected ciphers,
598 which may be activated as a side-effect of selecting a single cipher.
600 (E.g., cipher list string "RSA" enables ciphersuites that are left
601 out of "ALL" because they do not provide symmetric encryption.
602 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
603 [Lutz Jaenicke, Bodo Moeller]
605 *) Add appropriate support for separate platform-dependent build
606 directories. The recommended way to make a platform-dependent
607 build directory is the following (tested on Linux), maybe with
610 # Place yourself outside of the OpenSSL source tree. In
611 # this example, the environment variable OPENSSL_SOURCE
612 # is assumed to contain the absolute OpenSSL source directory.
613 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
614 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
615 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
616 mkdir -p `dirname $F`
617 ln -s $OPENSSL_SOURCE/$F $F
620 To be absolutely sure not to disturb the source tree, a "make clean"
621 is a good thing. If it isn't successfull, don't worry about it,
622 it probably means the source directory is very clean.
625 *) Make sure any ENGINE control commands make local copies of string
626 pointers passed to them whenever necessary. Otherwise it is possible
627 the caller may have overwritten (or deallocated) the original string
628 data when a later ENGINE operation tries to use the stored values.
629 [Götz Babin-Ebell <babinebell@trustcenter.de>]
631 *) Improve diagnostics in file reading and command-line digests.
632 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
634 *) Add AES modes CFB and OFB to the object database. Correct an
635 error in AES-CFB decryption.
638 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
639 allows existing EVP_CIPHER_CTX structures to be reused after
640 calling EVP_*Final(). This behaviour is used by encryption
641 BIOs and some applications. This has the side effect that
642 applications must explicitly clean up cipher contexts with
643 EVP_CIPHER_CTX_cleanup() or they will leak memory.
646 *) Check the values of dna and dnb in bn_mul_recursive before calling
647 bn_mul_comba (a non zero value means the a or b arrays do not contain
648 n2 elements) and fallback to bn_mul_normal if either is not zero.
651 *) Fix escaping of non-ASCII characters when using the -subj option
652 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
655 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
656 form for "surname", serialNumber has no short form.
657 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
658 therefore remove "mail" short name for "internet 7".
659 The OID for unique identifiers in X509 certificates is
660 x500UniqueIdentifier, not uniqueIdentifier.
661 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
664 *) Add an "init" command to the ENGINE config module and auto initialize
665 ENGINEs. Without any "init" command the ENGINE will be initialized
666 after all ctrl commands have been executed on it. If init=1 the
667 ENGINE is initailized at that point (ctrls before that point are run
668 on the uninitialized ENGINE and after on the initialized one). If
669 init=0 then the ENGINE will not be iniatialized at all.
672 *) Fix the 'app_verify_callback' interface so that the user-defined
673 argument is actually passed to the callback: In the
674 SSL_CTX_set_cert_verify_callback() prototype, the callback
675 declaration has been changed from
678 int (*cb)(X509_STORE_CTX *,void *);
679 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
680 i=s->ctx->app_verify_callback(&ctx)
681 has been changed into
682 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
684 To update applications using SSL_CTX_set_cert_verify_callback(),
685 a dummy argument can be added to their callback functions.
686 [D. K. Smetters <smetters@parc.xerox.com>]
688 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
689 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
691 *) Add and OPENSSL_LOAD_CONF define which will cause
692 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
693 This allows older applications to transparently support certain
694 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
695 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
696 load the config file and OPENSSL_add_all_algorithms_conf() which will
697 always load it have also been added.
700 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
701 Adjust NIDs and EVP layer.
702 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
704 *) Config modules support in openssl utility.
706 Most commands now load modules from the config file,
707 though in a few (such as version) this isn't done
708 because it couldn't be used for anything.
710 In the case of ca and req the config file used is
711 the same as the utility itself: that is the -config
712 command line option can be used to specify an
716 *) Move default behaviour from OPENSSL_config(). If appname is NULL
717 use "openssl_conf" if filename is NULL use default openssl config file.
720 *) Add an argument to OPENSSL_config() to allow the use of an alternative
721 config section name. Add a new flag to tolerate a missing config file
722 and move code to CONF_modules_load_file().
725 *) Support for crypto accelerator cards from Accelerated Encryption
726 Processing, www.aep.ie. (Use engine 'aep')
727 The support was copied from 0.9.6c [engine] and adapted/corrected
728 to work with the new engine framework.
729 [AEP Inc. and Richard Levitte]
731 *) Support for SureWare crypto accelerator cards from Baltimore
732 Technologies. (Use engine 'sureware')
733 The support was copied from 0.9.6c [engine] and adapted
734 to work with the new engine framework.
737 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
738 make the newer ENGINE framework commands for the CHIL engine work.
739 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
741 *) Make it possible to produce shared libraries on ReliantUNIX.
742 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
744 *) Add the configuration target debug-linux-ppro.
745 Make 'openssl rsa' use the general key loading routines
746 implemented in apps.c, and make those routines able to
747 handle the key format FORMAT_NETSCAPE and the variant
749 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
751 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
752 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
754 *) Add -keyform to rsautl, and document -engine.
755 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
757 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
758 BIO_R_NO_SUCH_FILE error code rather than the generic
759 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
764 ERR_peek_last_error_line
765 ERR_peek_last_error_line_data.
769 ERR_peek_error_line_data,
770 but report on the latest error recorded rather than the first one
771 still in the error queue.
772 [Ben Laurie, Bodo Moeller]
774 *) default_algorithms option in ENGINE config module. This allows things
776 default_algorithms = ALL
777 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
780 *) Prelminary ENGINE config module.
783 *) New experimental application configuration code.
786 *) Change the AES code to follow the same name structure as all other
787 symmetric ciphers, and behave the same way. Move everything to
788 the directory crypto/aes, thereby obsoleting crypto/rijndael.
789 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
791 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
792 [Ben Laurie and Theo de Raadt]
794 *) Add option to output public keys in req command.
795 [Massimiliano Pala madwolf@openca.org]
797 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
798 (up to about 10% better than before for P-192 and P-224).
801 *) New functions/macros
803 SSL_CTX_set_msg_callback(ctx, cb)
804 SSL_CTX_set_msg_callback_arg(ctx, arg)
805 SSL_set_msg_callback(ssl, cb)
806 SSL_set_msg_callback_arg(ssl, arg)
808 to request calling a callback function
810 void cb(int write_p, int version, int content_type,
811 const void *buf, size_t len, SSL *ssl, void *arg)
813 whenever a protocol message has been completely received
814 (write_p == 0) or sent (write_p == 1). Here 'version' is the
815 protocol version according to which the SSL library interprets
816 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
817 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
818 the content type as defined in the SSL 3.0/TLS 1.0 protocol
819 specification (change_cipher_spec(20), alert(21), handshake(22)).
820 'buf' and 'len' point to the actual message, 'ssl' to the
821 SSL object, and 'arg' is the application-defined value set by
822 SSL[_CTX]_set_msg_callback_arg().
824 'openssl s_client' and 'openssl s_server' have new '-msg' options
825 to enable a callback that displays all protocol messages.
828 *) Change the shared library support so shared libraries are built as
829 soon as the corresponding static library is finished, and thereby get
830 openssl and the test programs linked against the shared library.
831 This still only happens when the keyword "shard" has been given to
832 the configuration scripts.
834 NOTE: shared library support is still an experimental thing, and
835 backward binary compatibility is still not guaranteed.
836 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
838 *) Add support for Subject Information Access extension.
839 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
841 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
842 additional bytes when new memory had to be allocated, not just
843 when reusing an existing buffer.
846 *) New command line and configuration option 'utf8' for the req command.
847 This allows field values to be specified as UTF8 strings.
850 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
851 runs for the former and machine-readable output for the latter.
854 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
855 of the e-mail address in the DN (i.e., it will go into a certificate
856 extension only). The new configuration file option 'email_in_dn = no'
858 [Massimiliano Pala madwolf@openca.org]
860 *) Change all functions with names starting with des_ to be starting
861 with DES_ instead. Add wrappers that are compatible with libdes,
862 but are named _ossl_old_des_*. Finally, add macros that map the
863 des_* symbols to the corresponding _ossl_old_des_* if libdes
864 compatibility is desired. If OpenSSL 0.9.6c compatibility is
865 desired, the des_* symbols will be mapped to DES_*, with one
868 Since we provide two compatibility mappings, the user needs to
869 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
870 compatibility is desired. The default (i.e., when that macro
871 isn't defined) is OpenSSL 0.9.6c compatibility.
873 There are also macros that enable and disable the support of old
874 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
875 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
876 are defined, the default will apply: to support the old des routines.
878 In either case, one must include openssl/des.h to get the correct
879 definitions. Do not try to just include openssl/des_old.h, that
882 NOTE: This is a major break of an old API into a new one. Software
883 authors are encouraged to switch to the DES_ style functions. Some
884 time in the future, des_old.h and the libdes compatibility functions
885 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
886 default), and then completely removed.
889 *) Test for certificates which contain unsupported critical extensions.
890 If such a certificate is found during a verify operation it is
891 rejected by default: this behaviour can be overridden by either
892 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
893 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
894 X509_supported_extension() has also been added which returns 1 if a
895 particular extension is supported.
898 *) Modify the behaviour of EVP cipher functions in similar way to digests
899 to retain compatibility with existing code.
902 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
903 compatibility with existing code. In particular the 'ctx' parameter does
904 not have to be to be initialized before the call to EVP_DigestInit() and
905 it is tidied up after a call to EVP_DigestFinal(). New function
906 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
907 EVP_MD_CTX_copy() changed to not require the destination to be
908 initialized valid and new function EVP_MD_CTX_copy_ex() added which
909 requires the destination to be valid.
911 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
912 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
915 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
916 so that complete 'Handshake' protocol structures are kept in memory
917 instead of overwriting 'msg_type' and 'length' with 'body' data.
920 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
921 [Massimo Santin via Richard Levitte]
923 *) Major restructuring to the underlying ENGINE code. This includes
924 reduction of linker bloat, separation of pure "ENGINE" manipulation
925 (initialisation, etc) from functionality dealing with implementations
926 of specific crypto iterfaces. This change also introduces integrated
927 support for symmetric ciphers and digest implementations - so ENGINEs
928 can now accelerate these by providing EVP_CIPHER and EVP_MD
929 implementations of their own. This is detailed in crypto/engine/README
930 as it couldn't be adequately described here. However, there are a few
931 API changes worth noting - some RSA, DSA, DH, and RAND functions that
932 were changed in the original introduction of ENGINE code have now
933 reverted back - the hooking from this code to ENGINE is now a good
934 deal more passive and at run-time, operations deal directly with
935 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
936 dereferencing through an ENGINE pointer any more. Also, the ENGINE
937 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
938 they were not being used by the framework as there is no concept of a
939 BIGNUM_METHOD and they could not be generalised to the new
940 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
941 ENGINE_cpy() has been removed as it cannot be consistently defined in
945 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
948 *) Change mkdef.pl to sort symbols that get the same entry number,
949 and make sure the automatically generated functions ERR_load_*
950 become part of libeay.num as well.
953 *) New function SSL_renegotiate_pending(). This returns true once
954 renegotiation has been requested (either SSL_renegotiate() call
955 or HelloRequest/ClientHello receveived from the peer) and becomes
956 false once a handshake has been completed.
957 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
958 sends a HelloRequest, but does not ensure that a handshake takes
959 place. SSL_renegotiate_pending() is useful for checking if the
960 client has followed the request.)
963 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
964 By default, clients may request session resumption even during
965 renegotiation (if session ID contexts permit); with this option,
966 session resumption is possible only in the first handshake.
968 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
969 more bits available for options that should not be part of
970 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
973 *) Add some demos for certificate and certificate request creation.
976 *) Make maximum certificate chain size accepted from the peer application
977 settable (SSL*_get/set_max_cert_list()), as proposed by
978 "Douglas E. Engert" <deengert@anl.gov>.
981 *) Add support for shared libraries for Unixware-7
982 (Boyd Lynn Gerber <gerberb@zenez.com>).
985 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
986 be done prior to destruction. Use this to unload error strings from
987 ENGINEs that load their own error strings. NB: This adds two new API
988 functions to "get" and "set" this destroy handler in an ENGINE.
991 *) Alter all existing ENGINE implementations (except "openssl" and
992 "openbsd") to dynamically instantiate their own error strings. This
993 makes them more flexible to be built both as statically-linked ENGINEs
994 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
995 Also, add stub code to each that makes building them as self-contained
996 shared-libraries easier (see README.ENGINE).
999 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1000 implementations into applications that are completely implemented in
1001 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1002 commands that can be used to configure what shared-library to load and
1003 to control aspects of the way it is handled. Also, made an update to
1004 the README.ENGINE file that brings its information up-to-date and
1005 provides some information and instructions on the "dynamic" ENGINE
1006 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1009 *) Make it possible to unload ranges of ERR strings with a new
1010 "ERR_unload_strings" function.
1013 *) Add a copy() function to EVP_MD.
1016 *) Make EVP_MD routines take a context pointer instead of just the
1017 md_data void pointer.
1020 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1021 that the digest can only process a single chunk of data
1022 (typically because it is provided by a piece of
1023 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1024 is only going to provide a single chunk of data, and hence the
1025 framework needn't accumulate the data for oneshot drivers.
1028 *) As with "ERR", make it possible to replace the underlying "ex_data"
1029 functions. This change also alters the storage and management of global
1030 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1031 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1032 index counters. The API functions that use this state have been changed
1033 to take a "class_index" rather than pointers to the class's local STACK
1034 and counter, and there is now an API function to dynamically create new
1035 classes. This centralisation allows us to (a) plug a lot of the
1036 thread-safety problems that existed, and (b) makes it possible to clean
1037 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1038 such data would previously have always leaked in application code and
1039 workarounds were in place to make the memory debugging turn a blind eye
1040 to it. Application code that doesn't use this new function will still
1041 leak as before, but their memory debugging output will announce it now
1042 rather than letting it slide.
1044 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1045 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1046 has a return value to indicate success or failure.
1049 *) Make it possible to replace the underlying "ERR" functions such that the
1050 global state (2 LHASH tables and 2 locks) is only used by the "default"
1051 implementation. This change also adds two functions to "get" and "set"
1052 the implementation prior to it being automatically set the first time
1053 any other ERR function takes place. Ie. an application can call "get",
1054 pass the return value to a module it has just loaded, and that module
1055 can call its own "set" function using that value. This means the
1056 module's "ERR" operations will use (and modify) the error state in the
1057 application and not in its own statically linked copy of OpenSSL code.
1060 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1061 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1062 the operation, and provides a more encapsulated way for external code
1063 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1064 to use these functions rather than manually incrementing the counts.
1066 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1069 *) Add EVP test program.
1072 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1075 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1076 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1077 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1078 These allow a CRL to be built without having to access X509_CRL fields
1079 directly. Modify 'ca' application to use new functions.
1082 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1083 bug workarounds. Rollback attack detection is a security feature.
1084 The problem will only arise on OpenSSL servers when TLSv1 is not
1085 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1086 Software authors not wanting to support TLSv1 will have special reasons
1087 for their choice and can explicitly enable this option.
1088 [Bodo Moeller, Lutz Jaenicke]
1090 *) Rationalise EVP so it can be extended: don't include a union of
1091 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1092 (similar to those existing for EVP_CIPHER_CTX).
1097 EVP_MD_CTX_init(&md); /* new function call */
1098 EVP_DigestInit(&md, EVP_sha1());
1099 EVP_DigestUpdate(&md, in, len);
1100 EVP_DigestFinal(&md, out, NULL);
1101 EVP_MD_CTX_cleanup(&md); /* new function call */
1105 *) Make DES key schedule conform to the usual scheme, as well as
1106 correcting its structure. This means that calls to DES functions
1107 now have to pass a pointer to a des_key_schedule instead of a
1108 plain des_key_schedule (which was actually always a pointer
1111 des_key_schedule ks;
1113 des_set_key_checked(..., &ks);
1114 des_ncbc_encrypt(..., &ks, ...);
1116 (Note that a later change renames 'des_...' into 'DES_...'.)
1119 *) Initial reduction of linker bloat: the use of some functions, such as
1120 PEM causes large amounts of unused functions to be linked in due to
1121 poor organisation. For example pem_all.c contains every PEM function
1122 which has a knock on effect of linking in large amounts of (unused)
1123 ASN1 code. Grouping together similar functions and splitting unrelated
1124 functions prevents this.
1127 *) Cleanup of EVP macros.
1130 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1131 correct _ecb suffix.
1134 *) Add initial OCSP responder support to ocsp application. The
1135 revocation information is handled using the text based index
1136 use by the ca application. The responder can either handle
1137 requests generated internally, supplied in files (for example
1138 via a CGI script) or using an internal minimal server.
1141 *) Add configuration choices to get zlib compression for TLS.
1144 *) Changes to Kerberos SSL for RFC 2712 compliance:
1145 1. Implemented real KerberosWrapper, instead of just using
1146 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1147 2. Implemented optional authenticator field of KerberosWrapper.
1149 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1150 and authenticator structs; see crypto/krb5/.
1152 Generalized Kerberos calls to support multiple Kerberos libraries.
1153 [Vern Staats <staatsvr@asc.hpc.mil>,
1154 Jeffrey Altman <jaltman@columbia.edu>
1155 via Richard Levitte]
1157 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1158 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1159 values for each of the key sizes rather than having just
1160 parameters (and 'speed' generating keys each time).
1163 *) Speed up EVP routines.
1166 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1167 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1168 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1169 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1171 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1172 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1173 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1176 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1178 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1181 *) Added the OS2-EMX target.
1182 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1184 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1185 to support NCONF routines in extension code. New function CONF_set_nconf()
1186 to allow functions which take an NCONF to also handle the old LHASH
1187 structure: this means that the old CONF compatible routines can be
1188 retained (in particular wrt extensions) without having to duplicate the
1189 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1192 *) Enhance the general user interface with mechanisms for inner control
1193 and with possibilities to have yes/no kind of prompts.
1196 *) Change all calls to low level digest routines in the library and
1197 applications to use EVP. Add missing calls to HMAC_cleanup() and
1198 don't assume HMAC_CTX can be copied using memcpy().
1199 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1201 *) Add the possibility to control engines through control names but with
1202 arbitrary arguments instead of just a string.
1203 Change the key loaders to take a UI_METHOD instead of a callback
1204 function pointer. NOTE: this breaks binary compatibility with earlier
1205 versions of OpenSSL [engine].
1206 Adapt the nCipher code for these new conditions and add a card insertion
1210 *) Enhance the general user interface with mechanisms to better support
1211 dialog box interfaces, application-defined prompts, the possibility
1212 to use defaults (for example default passwords from somewhere else)
1213 and interrupts/cancellations.
1216 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1217 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1220 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1221 tidy up some unnecessarily weird code in 'sk_new()').
1222 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1224 *) Change the key loading routines for ENGINEs to use the same kind
1225 callback (pem_password_cb) as all other routines that need this
1229 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1230 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1231 than this minimum value is recommended.
1234 *) New random seeder for OpenVMS, using the system process statistics
1235 that are easily reachable.
1238 *) Windows apparently can't transparently handle global
1239 variables defined in DLLs. Initialisations such as:
1241 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1243 wont compile. This is used by the any applications that need to
1244 declare their own ASN1 modules. This was fixed by adding the option
1245 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1246 needed for static libraries under Win32.
1249 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1250 setting of purpose and trust fields. New X509_STORE trust and
1251 purpose functions and tidy up setting in other SSL functions.
1254 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1255 structure. These are inherited by X509_STORE_CTX when it is
1256 initialised. This allows various defaults to be set in the
1257 X509_STORE structure (such as flags for CRL checking and custom
1258 purpose or trust settings) for functions which only use X509_STORE_CTX
1259 internally such as S/MIME.
1261 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1262 trust settings if they are not set in X509_STORE. This allows X509_STORE
1263 purposes and trust (in S/MIME for example) to override any set by default.
1265 Add command line options for CRL checking to smime, s_client and s_server
1269 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1270 are set then the CRL is looked up in the X509_STORE structure and
1271 its validity and signature checked, then if the certificate is found
1272 in the CRL the verify fails with a revoked error.
1274 Various new CRL related callbacks added to X509_STORE_CTX structure.
1276 Command line options added to 'verify' application to support this.
1278 This needs some additional work, such as being able to handle multiple
1279 CRLs with different times, extension based lookup (rather than just
1280 by subject name) and ultimately more complete V2 CRL extension
1284 *) Add a general user interface API (crypto/ui/). This is designed
1285 to replace things like des_read_password and friends (backward
1286 compatibility functions using this new API are provided).
1287 The purpose is to remove prompting functions from the DES code
1288 section as well as provide for prompting through dialog boxes in
1289 a window system and the like.
1292 *) Add "ex_data" support to ENGINE so implementations can add state at a
1293 per-structure level rather than having to store it globally.
1296 *) Make it possible for ENGINE structures to be copied when retrieved by
1297 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1298 This causes the "original" ENGINE structure to act like a template,
1299 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1300 operational state can be localised to each ENGINE structure, despite the
1301 fact they all share the same "methods". New ENGINE structures returned in
1302 this case have no functional references and the return value is the single
1303 structural reference. This matches the single structural reference returned
1304 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1308 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1309 needs to match any other type at all we need to manually clear the
1313 *) Changes to the "openssl engine" utility to include;
1314 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1315 about an ENGINE's available control commands.
1316 - executing control commands from command line arguments using the
1317 '-pre' and '-post' switches. '-post' is only used if '-t' is
1318 specified and the ENGINE is successfully initialised. The syntax for
1319 the individual commands are colon-separated, for example;
1320 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1323 *) New dynamic control command support for ENGINEs. ENGINEs can now
1324 declare their own commands (numbers), names (strings), descriptions,
1325 and input types for run-time discovery by calling applications. A
1326 subset of these commands are implicitly classed as "executable"
1327 depending on their input type, and only these can be invoked through
1328 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1329 can be based on user input, config files, etc). The distinction is
1330 that "executable" commands cannot return anything other than a boolean
1331 result and can only support numeric or string input, whereas some
1332 discoverable commands may only be for direct use through
1333 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1334 pointers, or other custom uses. The "executable" commands are to
1335 support parameterisations of ENGINE behaviour that can be
1336 unambiguously defined by ENGINEs and used consistently across any
1337 OpenSSL-based application. Commands have been added to all the
1338 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1339 control over shared-library paths without source code alterations.
1342 *) Changed all ENGINE implementations to dynamically allocate their
1343 ENGINEs rather than declaring them statically. Apart from this being
1344 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1345 this also allows the implementations to compile without using the
1346 internal engine_int.h header.
1349 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1350 'const' value. Any code that should be able to modify a RAND_METHOD
1351 should already have non-const pointers to it (ie. they should only
1352 modify their own ones).
1355 *) Made a variety of little tweaks to the ENGINE code.
1356 - "atalla" and "ubsec" string definitions were moved from header files
1357 to C code. "nuron" string definitions were placed in variables
1358 rather than hard-coded - allowing parameterisation of these values
1359 later on via ctrl() commands.
1360 - Removed unused "#if 0"'d code.
1361 - Fixed engine list iteration code so it uses ENGINE_free() to release
1362 structural references.
1363 - Constified the RAND_METHOD element of ENGINE structures.
1364 - Constified various get/set functions as appropriate and added
1365 missing functions (including a catch-all ENGINE_cpy that duplicates
1366 all ENGINE values onto a new ENGINE except reference counts/state).
1367 - Removed NULL parameter checks in get/set functions. Setting a method
1368 or function to NULL is a way of cancelling out a previously set
1369 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1370 and doesn't justify the extra error symbols and code.
1371 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1372 flags from engine_int.h to engine.h.
1373 - Changed prototypes for ENGINE handler functions (init(), finish(),
1374 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1377 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1378 to the algorithm using long division. The binary algorithm can be
1379 used only if the modulus is odd. On 32-bit systems, it is faster
1380 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1381 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1382 up to 450 bits. In 64-bit environments, the binary algorithm
1383 appears to be advantageous for much longer moduli; here we use it
1384 for moduli up to 2048 bits.
1387 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1388 could not support the combine flag in choice fields.
1391 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1392 extensions from a certificate request to the certificate.
1395 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1396 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1397 file: this allows the display of the certificate about to be
1398 signed to be customised, to allow certain fields to be included
1399 or excluded and extension details. The old system didn't display
1400 multicharacter strings properly, omitted fields not in the policy
1401 and couldn't display additional details such as extensions.
1404 *) Function EC_POINTs_mul for multiple scalar multiplication
1405 of an arbitrary number of elliptic curve points
1406 \sum scalars[i]*points[i],
1407 optionally including the generator defined for the EC_GROUP:
1408 scalar*generator + \sum scalars[i]*points[i].
1410 EC_POINT_mul is a simple wrapper function for the typical case
1411 that the point list has just one item (besides the optional
1415 *) First EC_METHODs for curves over GF(p):
1417 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1418 operations and provides various method functions that can also
1419 operate with faster implementations of modular arithmetic.
1421 EC_GFp_mont_method() reuses most functions that are part of
1422 EC_GFp_simple_method, but uses Montgomery arithmetic.
1424 [Bodo Moeller; point addition and point doubling
1425 implementation directly derived from source code provided by
1426 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1428 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1429 crypto/ec/ec_lib.c):
1431 Curves are EC_GROUP objects (with an optional group generator)
1432 based on EC_METHODs that are built into the library.
1434 Points are EC_POINT objects based on EC_GROUP objects.
1436 Most of the framework would be able to handle curves over arbitrary
1437 finite fields, but as there are no obvious types for fields other
1438 than GF(p), some functions are limited to that for now.
1441 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1442 that the file contains a complete HTTP response.
1445 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1446 change the def and num file printf format specifier from "%-40sXXX"
1447 to "%-39s XXX". The latter will always guarantee a space after the
1448 field while the former will cause them to run together if the field
1449 is 40 of more characters long.
1452 *) Constify the cipher and digest 'method' functions and structures
1453 and modify related functions to take constant EVP_MD and EVP_CIPHER
1457 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1458 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1461 *) Modify EVP_Digest*() routines so they now return values. Although the
1462 internal software routines can never fail additional hardware versions
1466 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1468 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1469 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1472 ERR_R_NESTED_ASN1_ERROR
1474 ERR_R_MISSING_ASN1_EOS
1475 were 4 .. 9, conflicting with
1476 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1478 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1479 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1481 Add new error code 'ERR_R_INTERNAL_ERROR'.
1484 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1488 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1489 sets the subject name for a new request or supersedes the
1490 subject name in a given request. Formats that can be parsed are
1491 'CN=Some Name, OU=myOU, C=IT'
1493 'CN=Some Name/OU=myOU/C=IT'.
1495 Add options '-batch' and '-verbose' to 'openssl req'.
1496 [Massimiliano Pala <madwolf@hackmasters.net>]
1498 *) Introduce the possibility to access global variables through
1499 functions on platform were that's the best way to handle exporting
1500 global variables in shared libraries. To enable this functionality,
1501 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1502 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1503 is normally done by Configure or something similar).
1505 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1506 in the source file (foo.c) like this:
1508 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1509 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1511 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1512 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1514 OPENSSL_DECLARE_GLOBAL(int,foo);
1515 #define foo OPENSSL_GLOBAL_REF(foo)
1516 OPENSSL_DECLARE_GLOBAL(double,bar);
1517 #define bar OPENSSL_GLOBAL_REF(bar)
1519 The #defines are very important, and therefore so is including the
1520 header file everywhere where the defined globals are used.
1522 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1523 of ASN.1 items, but that structure is a bit different.
1525 The largest change is in util/mkdef.pl which has been enhanced with
1526 better and easier to understand logic to choose which symbols should
1527 go into the Windows .def files as well as a number of fixes and code
1528 cleanup (among others, algorithm keywords are now sorted
1529 lexicographically to avoid constant rewrites).
1532 *) In BN_div() keep a copy of the sign of 'num' before writing the
1533 result to 'rm' because if rm==num the value will be overwritten
1534 and produce the wrong result if 'num' is negative: this caused
1535 problems with BN_mod() and BN_nnmod().
1538 *) Function OCSP_request_verify(). This checks the signature on an
1539 OCSP request and verifies the signer certificate. The signer
1540 certificate is just checked for a generic purpose and OCSP request
1544 *) Add OCSP_check_validity() function to check the validity of OCSP
1545 responses. OCSP responses are prepared in real time and may only
1546 be a few seconds old. Simply checking that the current time lies
1547 between thisUpdate and nextUpdate max reject otherwise valid responses
1548 caused by either OCSP responder or client clock inaccuracy. Instead
1549 we allow thisUpdate and nextUpdate to fall within a certain period of
1550 the current time. The age of the response can also optionally be
1551 checked. Two new options -validity_period and -status_age added to
1555 *) If signature or public key algorithm is unrecognized print out its
1556 OID rather that just UNKNOWN.
1559 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1560 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1561 ID to be generated from the issuer certificate alone which can then be
1562 passed to OCSP_id_issuer_cmp().
1565 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1566 ASN1 modules to export functions returning ASN1_ITEM pointers
1567 instead of the ASN1_ITEM structures themselves. This adds several
1568 new macros which allow the underlying ASN1 function/structure to
1569 be accessed transparently. As a result code should not use ASN1_ITEM
1570 references directly (such as &X509_it) but instead use the relevant
1571 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1572 use of the new ASN1 code on platforms where exporting structures
1573 is problematical (for example in shared libraries) but exporting
1574 functions returning pointers to structures is not.
1577 *) Add support for overriding the generation of SSL/TLS session IDs.
1578 These callbacks can be registered either in an SSL_CTX or per SSL.
1579 The purpose of this is to allow applications to control, if they wish,
1580 the arbitrary values chosen for use as session IDs, particularly as it
1581 can be useful for session caching in multiple-server environments. A
1582 command-line switch for testing this (and any client code that wishes
1583 to use such a feature) has been added to "s_server".
1584 [Geoff Thorpe, Lutz Jaenicke]
1586 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1587 of the form '#if defined(...) || defined(...) || ...' and
1588 '#if !defined(...) && !defined(...) && ...'. This also avoids
1589 the growing number of special cases it was previously handling.
1592 *) Make all configuration macros available for application by making
1593 sure they are available in opensslconf.h, by giving them names starting
1594 with "OPENSSL_" to avoid conflicts with other packages and by making
1595 sure e_os2.h will cover all platform-specific cases together with
1597 Additionally, it is now possible to define configuration/platform-
1598 specific names (called "system identities"). In the C code, these
1599 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1600 macro with the name beginning with "OPENSSL_SYS_", which is determined
1601 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1605 *) New option -set_serial to 'req' and 'x509' this allows the serial
1606 number to use to be specified on the command line. Previously self
1607 signed certificates were hard coded with serial number 0 and the
1608 CA options of 'x509' had to use a serial number in a file which was
1612 *) New options to 'ca' utility to support V2 CRL entry extensions.
1613 Currently CRL reason, invalidity date and hold instruction are
1614 supported. Add new CRL extensions to V3 code and some new objects.
1617 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1618 disable standard block padding (aka PKCS#5 padding) in the EVP
1619 API, which was previously mandatory. This means that the data is
1620 not padded in any way and so the total length much be a multiple
1621 of the block size, otherwise an error occurs.
1624 *) Initial (incomplete) OCSP SSL support.
1627 *) New function OCSP_parse_url(). This splits up a URL into its host,
1628 port and path components: primarily to parse OCSP URLs. New -url
1629 option to ocsp utility.
1632 *) New nonce behavior. The return value of OCSP_check_nonce() now
1633 reflects the various checks performed. Applications can decide
1634 whether to tolerate certain situations such as an absent nonce
1635 in a response when one was present in a request: the ocsp application
1636 just prints out a warning. New function OCSP_add1_basic_nonce()
1637 this is to allow responders to include a nonce in a response even if
1638 the request is nonce-less.
1641 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1642 skipped when using openssl x509 multiple times on a single input file,
1643 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1646 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1647 set string type: to handle setting ASN1_TIME structures. Fix ca
1648 utility to correctly initialize revocation date of CRLs.
1651 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1652 the clients preferred ciphersuites and rather use its own preferences.
1653 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1654 Internet Explorer by ensuring unchanged hash method during stepup.
1655 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1658 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1659 to aes and add a new 'exist' option to print out symbols that don't
1663 *) Additional options to ocsp utility to allow flags to be set and
1664 additional certificates supplied.
1667 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1668 OCSP client a number of certificate to only verify the response
1672 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1673 handle the new API. Currently only ECB, CBC modes supported. Add new
1676 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1677 Encryption Standard (AES) Ciphersuites for Transport Layer
1678 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1679 not enabled by default and were not part of the "ALL" ciphersuite
1680 alias because they were not yet official; they could be
1681 explicitly requested by specifying the "AESdraft" ciphersuite
1682 group alias. In the final release of OpenSSL 0.9.7, the group
1683 alias is called "AES" and is part of "ALL".)
1684 [Ben Laurie, Steve Henson, Bodo Moeller]
1686 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1687 request to response.
1690 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1691 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1692 extract information from a certificate request. OCSP_response_create()
1693 creates a response and optionally adds a basic response structure.
1694 OCSP_basic_add1_status() adds a complete single response to a basic
1695 response and returns the OCSP_SINGLERESP structure just added (to allow
1696 extensions to be included for example). OCSP_basic_add1_cert() adds a
1697 certificate to a basic response and OCSP_basic_sign() signs a basic
1698 response with various flags. New helper functions ASN1_TIME_check()
1699 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1700 (converts ASN1_TIME to GeneralizedTime).
1703 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1704 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1705 structure from a certificate. X509_pubkey_digest() digests the public_key
1706 contents: this is used in various key identifiers.
1709 *) Make sk_sort() tolerate a NULL argument.
1710 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1712 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1713 passed by the function are trusted implicitly. If any of them signed the
1714 response then it is assumed to be valid and is not verified.
1717 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1718 to data. This was previously part of the PKCS7 ASN1 code. This
1719 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1720 [Steve Henson, reported by Kenneth R. Robinette
1721 <support@securenetterm.com>]
1723 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1724 routines: without these tracing memory leaks is very painful.
1725 Fix leaks in PKCS12 and PKCS7 routines.
1728 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1729 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1730 effectively meant GeneralizedTime would never be used. Now it
1731 is initialised to -1 but X509_time_adj() now has to check the value
1732 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1733 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1734 [Steve Henson, reported by Kenneth R. Robinette
1735 <support@securenetterm.com>]
1737 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1738 result in a zero length in the ASN1_INTEGER structure which was
1739 not consistent with the structure when d2i_ASN1_INTEGER() was used
1740 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1741 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1742 where it did not print out a minus for negative ASN1_INTEGER.
1745 *) Add summary printout to ocsp utility. The various functions which
1746 convert status values to strings have been renamed to:
1747 OCSP_response_status_str(), OCSP_cert_status_str() and
1748 OCSP_crl_reason_str() and are no longer static. New options
1749 to verify nonce values and to disable verification. OCSP response
1750 printout format cleaned up.
1753 *) Add additional OCSP certificate checks. These are those specified
1754 in RFC2560. This consists of two separate checks: the CA of the
1755 certificate being checked must either be the OCSP signer certificate
1756 or the issuer of the OCSP signer certificate. In the latter case the
1757 OCSP signer certificate must contain the OCSP signing extended key
1758 usage. This check is performed by attempting to match the OCSP
1759 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1760 in the OCSP_CERTID structures of the response.
1763 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1764 and related routines. This uses the standard OpenSSL certificate
1765 verify routines to perform initial checks (just CA validity) and
1766 to obtain the certificate chain. Then additional checks will be
1767 performed on the chain. Currently the root CA is checked to see
1768 if it is explicitly trusted for OCSP signing. This is used to set
1769 a root CA as a global signing root: that is any certificate that
1770 chains to that CA is an acceptable OCSP signing certificate.
1773 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1774 extensions from a separate configuration file.
1775 As when reading extensions from the main configuration file,
1776 the '-extensions ...' option may be used for specifying the
1778 [Massimiliano Pala <madwolf@comune.modena.it>]
1780 *) New OCSP utility. Allows OCSP requests to be generated or
1781 read. The request can be sent to a responder and the output
1782 parsed, outputed or printed in text form. Not complete yet:
1783 still needs to check the OCSP response validity.
1786 *) New subcommands for 'openssl ca':
1787 'openssl ca -status <serial>' prints the status of the cert with
1788 the given serial number (according to the index file).
1789 'openssl ca -updatedb' updates the expiry status of certificates
1791 [Massimiliano Pala <madwolf@comune.modena.it>]
1793 *) New '-newreq-nodes' command option to CA.pl. This is like
1794 '-newreq', but calls 'openssl req' with the '-nodes' option
1795 so that the resulting key is not encrypted.
1796 [Damien Miller <djm@mindrot.org>]
1798 *) New configuration for the GNU Hurd.
1799 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1801 *) Initial code to implement OCSP basic response verify. This
1802 is currently incomplete. Currently just finds the signer's
1803 certificate and verifies the signature on the response.
1806 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1807 value of OPENSSLDIR. This is available via the new '-d' option
1808 to 'openssl version', and is also included in 'openssl version -a'.
1811 *) Allowing defining memory allocation callbacks that will be given
1812 file name and line number information in additional arguments
1813 (a const char* and an int). The basic functionality remains, as
1814 well as the original possibility to just replace malloc(),
1815 realloc() and free() by functions that do not know about these
1816 additional arguments. To register and find out the current
1817 settings for extended allocation functions, the following
1818 functions are provided:
1820 CRYPTO_set_mem_ex_functions
1821 CRYPTO_set_locked_mem_ex_functions
1822 CRYPTO_get_mem_ex_functions
1823 CRYPTO_get_locked_mem_ex_functions
1825 These work the same way as CRYPTO_set_mem_functions and friends.
1826 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1827 extended allocation function is enabled.
1828 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1829 a conventional allocation function is enabled.
1830 [Richard Levitte, Bodo Moeller]
1832 *) Finish off removing the remaining LHASH function pointer casts.
1833 There should no longer be any prototype-casting required when using
1834 the LHASH abstraction, and any casts that remain are "bugs". See
1835 the callback types and macros at the head of lhash.h for details
1836 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1839 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1840 If /dev/[u]random devices are not available or do not return enough
1841 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1843 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1844 /etc/entropy will be queried once each in this sequence, quering stops
1845 when enough entropy was collected without querying more sockets.
1848 *) Change the Unix RAND_poll() variant to be able to poll several
1849 random devices, as specified by DEVRANDOM, until a sufficient amount
1850 of data has been collected. We spend at most 10 ms on each file
1851 (select timeout) and read in non-blocking mode. DEVRANDOM now
1852 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1853 (previously it was just the string "/dev/urandom"), so on typical
1854 platforms the 10 ms delay will never occur.
1855 Also separate out the Unix variant to its own file, rand_unix.c.
1856 For VMS, there's a currently-empty rand_vms.c.
1859 *) Move OCSP client related routines to ocsp_cl.c. These
1860 provide utility functions which an application needing
1861 to issue a request to an OCSP responder and analyse the
1862 response will typically need: as opposed to those which an
1863 OCSP responder itself would need which will be added later.
1865 OCSP_request_sign() signs an OCSP request with an API similar
1866 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1867 response. OCSP_response_get1_basic() extracts basic response
1868 from response. OCSP_resp_find_status(): finds and extracts status
1869 information from an OCSP_CERTID structure (which will be created
1870 when the request structure is built). These are built from lower
1871 level functions which work on OCSP_SINGLERESP structures but
1872 wont normally be used unless the application wishes to examine
1873 extensions in the OCSP response for example.
1875 Replace nonce routines with a pair of functions.
1876 OCSP_request_add1_nonce() adds a nonce value and optionally
1877 generates a random value. OCSP_check_nonce() checks the
1878 validity of the nonce in an OCSP response.
1881 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1882 This doesn't copy the supplied OCSP_CERTID and avoids the
1883 need to free up the newly created id. Change return type
1884 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1885 This can then be used to add extensions to the request.
1886 Deleted OCSP_request_new(), since most of its functionality
1887 is now in OCSP_REQUEST_new() (and the case insensitive name
1888 clash) apart from the ability to set the request name which
1889 will be added elsewhere.
1892 *) Update OCSP API. Remove obsolete extensions argument from
1893 various functions. Extensions are now handled using the new
1894 OCSP extension code. New simple OCSP HTTP function which
1895 can be used to send requests and parse the response.
1898 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1899 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1900 uses the special reorder version of SET OF to sort the attributes
1901 and reorder them to match the encoded order. This resolves a long
1902 standing problem: a verify on a PKCS7 structure just after signing
1903 it used to fail because the attribute order did not match the
1904 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1905 it uses the received order. This is necessary to tolerate some broken
1906 software that does not order SET OF. This is handled by encoding
1907 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1908 to produce the required SET OF.
1911 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1912 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1913 files to get correct declarations of the ASN.1 item variables.
1916 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1917 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1918 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1919 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1920 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1921 ASN1_ITEM and no wrapper functions.
1924 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1925 replace the old function pointer based I/O routines. Change most of
1926 the *_d2i_bio() and *_d2i_fp() functions to use these.
1929 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1930 lines, recognice more "algorithms" that can be deselected, and make
1931 it complain about algorithm deselection that isn't recognised.
1934 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1935 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1936 to use new functions. Add NO_ASN1_OLD which can be set to remove
1937 some old style ASN1 functions: this can be used to determine if old
1938 code will still work when these eventually go away.
1941 *) New extension functions for OCSP structures, these follow the
1942 same conventions as certificates and CRLs.
1945 *) New function X509V3_add1_i2d(). This automatically encodes and
1946 adds an extension. Its behaviour can be customised with various
1947 flags to append, replace or delete. Various wrappers added for
1948 certifcates and CRLs.
1951 *) Fix to avoid calling the underlying ASN1 print routine when
1952 an extension cannot be parsed. Correct a typo in the
1953 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1956 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1957 entries for variables.
1960 *) Add functionality to apps/openssl.c for detecting locking
1961 problems: As the program is single-threaded, all we have
1962 to do is register a locking callback using an array for
1963 storing which locks are currently held by the program.
1966 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1967 SSL_get_ex_data_X509_STORE_idx(), which is used in
1968 ssl_verify_cert_chain() and thus can be called at any time
1969 during TLS/SSL handshakes so that thread-safety is essential.
1970 Unfortunately, the ex_data design is not at all suited
1971 for multi-threaded use, so it probably should be abolished.
1974 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1975 [Broadcom, tweaked and integrated by Geoff Thorpe]
1977 *) Move common extension printing code to new function
1978 X509V3_print_extensions(). Reorganise OCSP print routines and
1979 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1982 *) New function X509_signature_print() to remove duplication in some
1986 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1987 set (this was treated exactly the same as SET OF previously). This
1988 is used to reorder the STACK representing the structure to match the
1989 encoding. This will be used to get round a problem where a PKCS7
1990 structure which was signed could not be verified because the STACK
1991 order did not reflect the encoded order.
1994 *) Reimplement the OCSP ASN1 module using the new code.
1997 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1998 for its ASN1 operations. The old style function pointers still exist
1999 for now but they will eventually go away.
2002 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2003 completely replaces the old ASN1 functionality with a table driven
2004 encoder and decoder which interprets an ASN1_ITEM structure describing
2005 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2006 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2007 has also been converted to the new form.
2010 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2011 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2012 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2013 for negative moduli.
2016 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2017 of not touching the result's sign bit.
2020 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2024 *) Changed the LHASH code to use prototypes for callbacks, and created
2025 macros to declare and implement thin (optionally static) functions
2026 that provide type-safety and avoid function pointer casting for the
2027 type-specific callbacks.
2030 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2032 [Veers Staats <staatsvr@asc.hpc.mil>,
2033 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2035 *) Reformat the FAQ so the different questions and answers can be divided
2036 in sections depending on the subject.
2039 *) Have the zlib compression code load ZLIB.DLL dynamically under
2043 *) New function BN_mod_sqrt for computing square roots modulo a prime
2044 (using the probabilistic Tonelli-Shanks algorithm unless
2045 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2046 be handled deterministically).
2047 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2049 *) Make BN_mod_inverse faster by explicitly handling small quotients
2050 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2051 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2054 *) New function BN_kronecker.
2057 *) Fix BN_gcd so that it works on negative inputs; the result is
2058 positive unless both parameters are zero.
2059 Previously something reasonably close to an infinite loop was
2060 possible because numbers could be growing instead of shrinking
2061 in the implementation of Euclid's algorithm.
2064 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2065 sign of the number in question.
2067 Fix BN_is_word(a,w) to work correctly for w == 0.
2069 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2070 because its test if the absolute value of 'a' equals 'w'.
2071 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2072 it exists mostly for use in the implementations of BN_is_zero(),
2073 BN_is_one(), and BN_is_word().
2076 *) New function BN_swap.
2079 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2080 the exponentiation functions are more likely to produce reasonable
2081 results on negative inputs.
2084 *) Change BN_mod_mul so that the result is always non-negative.
2085 Previously, it could be negative if one of the factors was negative;
2086 I don't think anyone really wanted that behaviour.
2089 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2090 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2091 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2092 and add new functions:
2101 BN_mod_lshift1_quick
2105 These functions always generate non-negative results.
2107 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2108 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2110 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2111 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2112 be reduced modulo m.
2113 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2116 The following entry accidentily appeared in the CHANGES file
2117 distributed with OpenSSL 0.9.7. The modifications described in
2118 it do *not* apply to OpenSSL 0.9.7.
2120 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2121 was actually never needed) and in BN_mul(). The removal in BN_mul()
2122 required a small change in bn_mul_part_recursive() and the addition
2123 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2124 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2125 bn_sub_words() and bn_add_words() except they take arrays with
2130 *) In 'openssl passwd', verify passwords read from the terminal
2131 unless the '-salt' option is used (which usually means that
2132 verification would just waste user's time since the resulting
2133 hash is going to be compared with some given password hash)
2134 or the new '-noverify' option is used.
2136 This is an incompatible change, but it does not affect
2137 non-interactive use of 'openssl passwd' (passwords on the command
2138 line, '-stdin' option, '-in ...' option) and thus should not
2142 *) Remove all references to RSAref, since there's no more need for it.
2145 *) Make DSO load along a path given through an environment variable
2146 (SHLIB_PATH) with shl_load().
2149 *) Constify the ENGINE code as a result of BIGNUM constification.
2150 Also constify the RSA code and most things related to it. In a
2151 few places, most notable in the depth of the ASN.1 code, ugly
2152 casts back to non-const were required (to be solved at a later
2156 *) Make it so the openssl application has all engines loaded by default.
2159 *) Constify the BIGNUM routines a little more.
2162 *) Add the following functions:
2164 ENGINE_load_cswift()
2166 ENGINE_load_atalla()
2168 ENGINE_load_builtin_engines()
2170 That way, an application can itself choose if external engines that
2171 are built-in in OpenSSL shall ever be used or not. The benefit is
2172 that applications won't have to be linked with libdl or other dso
2173 libraries unless it's really needed.
2175 Changed 'openssl engine' to load all engines on demand.
2176 Changed the engine header files to avoid the duplication of some
2177 declarations (they differed!).
2180 *) 'openssl engine' can now list capabilities.
2183 *) Better error reporting in 'openssl engine'.
2186 *) Never call load_dh_param(NULL) in s_server.
2189 *) Add engine application. It can currently list engines by name and
2190 identity, and test if they are actually available.
2193 *) Improve RPM specification file by forcing symbolic linking and making
2194 sure the installed documentation is also owned by root.root.
2195 [Damien Miller <djm@mindrot.org>]
2197 *) Give the OpenSSL applications more possibilities to make use of
2198 keys (public as well as private) handled by engines.
2201 *) Add OCSP code that comes from CertCo.
2204 *) Add VMS support for the Rijndael code.
2207 *) Added untested support for Nuron crypto accelerator.
2210 *) Add support for external cryptographic devices. This code was
2211 previously distributed separately as the "engine" branch.
2212 [Geoff Thorpe, Richard Levitte]
2214 *) Rework the filename-translation in the DSO code. It is now possible to
2215 have far greater control over how a "name" is turned into a filename
2216 depending on the operating environment and any oddities about the
2217 different shared library filenames on each system.
2220 *) Support threads on FreeBSD-elf in Configure.
2223 *) Fix for SHA1 assembly problem with MASM: it produces
2224 warnings about corrupt line number information when assembling
2225 with debugging information. This is caused by the overlapping
2227 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2230 NCONF_get_number() has no error checking at all. As a replacement,
2231 NCONF_get_number_e() is defined (_e for "error checking") and is
2232 promoted strongly. The old NCONF_get_number is kept around for
2233 binary backward compatibility.
2234 Make it possible for methods to load from something other than a BIO,
2235 by providing a function pointer that is given a name instead of a BIO.
2236 For example, this could be used to load configuration data from an
2240 *) Fix for non blocking accept BIOs. Added new I/O special reason
2241 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2242 with non blocking I/O was not possible because no retry code was
2243 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2247 *) Added the beginnings of Rijndael support.
2250 *) Fix for bug in DirectoryString mask setting. Add support for
2251 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2252 to allow certificate printing to more controllable, additional
2253 'certopt' option to 'x509' to allow new printing options to be
2257 *) Clean old EAY MD5 hack from e_os.h.
2260 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2262 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2263 memory from it's contents. This is done with a counter that will
2264 place alternating values in each byte. This can be used to solve
2265 two issues: 1) the removal of calls to memset() by highly optimizing
2266 compilers, and 2) cleansing with other values than 0, since those can
2267 be read through on certain media, for example a swap space on disk.
2270 *) Bugfix: client side session caching did not work with external caching,
2271 because the session->cipher setting was not restored when reloading
2272 from the external cache. This problem was masked, when
2273 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2274 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2277 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2278 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2279 [Zeev Lieber <zeev-l@yahoo.com>]
2281 *) Undo an undocumented change introduced in 0.9.6e which caused
2282 repeated calls to OpenSSL_add_all_ciphers() and
2283 OpenSSL_add_all_digests() to be ignored, even after calling
2287 *) Change the default configuration reader to deal with last line not
2288 being properly terminated.
2291 *) Change X509_NAME_cmp() so it applies the special rules on handling
2292 DN values that are of type PrintableString, as well as RDNs of type
2293 emailAddress where the value has the type ia5String.
2294 [stefank@valicert.com via Richard Levitte]
2296 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2297 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2298 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2299 the bitwise-OR of the two for use by the majority of applications
2300 wanting this behaviour, and update the docs. The documented
2301 behaviour and actual behaviour were inconsistent and had been
2302 changing anyway, so this is more a bug-fix than a behavioural
2304 [Geoff Thorpe, diagnosed by Nadav Har'El]
2306 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2307 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2310 *) Fix initialization code race conditions in
2311 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2312 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2313 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2314 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2315 ssl2_get_cipher_by_char(),
2316 ssl3_get_cipher_by_char().
2317 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2319 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2320 the cached sessions are flushed, as the remove_cb() might use ex_data
2321 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2322 (see [openssl.org #212]).
2323 [Geoff Thorpe, Lutz Jaenicke]
2325 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2326 length, instead of the encoding length to d2i_ASN1_OBJECT.
2329 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2331 *) [In 0.9.6g-engine release:]
2332 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2333 [Lynn Gazis <lgazis@rainbow.com>]
2335 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2337 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2338 and get fix the header length calculation.
2339 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2340 Alon Kantor <alonk@checkpoint.com> (and others),
2343 *) Use proper error handling instead of 'assertions' in buffer
2344 overflow checks added in 0.9.6e. This prevents DoS (the
2345 assertions could call abort()).
2346 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2348 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2350 *) Add various sanity checks to asn1_get_length() to reject
2351 the ASN1 length bytes if they exceed sizeof(long), will appear
2352 negative or the content length exceeds the length of the
2354 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2356 *) Fix cipher selection routines: ciphers without encryption had no flags
2357 for the cipher strength set and where therefore not handled correctly
2358 by the selection routines (PR #130).
2361 *) Fix EVP_dsa_sha macro.
2365 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2366 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2367 that was added in OpenSSL 0.9.6d.
2369 As the countermeasure turned out to be incompatible with some
2370 broken SSL implementations, the new option is part of SSL_OP_ALL.
2371 SSL_OP_ALL is usually employed when compatibility with weird SSL
2372 implementations is desired (e.g. '-bugs' option to 's_client' and
2373 's_server'), so the new option is automatically set in many
2377 *) Changes in security patch:
2379 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2380 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2381 Air Force Materiel Command, USAF, under agreement number
2384 *) Add various sanity checks to asn1_get_length() to reject
2385 the ASN1 length bytes if they exceed sizeof(long), will appear
2386 negative or the content length exceeds the length of the
2387 supplied buffer. (CAN-2002-0659)
2388 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2390 *) Assertions for various potential buffer overflows, not known to
2392 [Ben Laurie (CHATS)]
2394 *) Various temporary buffers to hold ASCII versions of integers were
2395 too small for 64 bit platforms. (CAN-2002-0655)
2396 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2398 *) Remote buffer overflow in SSL3 protocol - an attacker could
2399 supply an oversized session ID to a client. (CAN-2002-0656)
2400 [Ben Laurie (CHATS)]
2402 *) Remote buffer overflow in SSL2 protocol - an attacker could
2403 supply an oversized client master key. (CAN-2002-0656)
2404 [Ben Laurie (CHATS)]
2406 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2408 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2409 encoded as NULL) with id-dsa-with-sha1.
2410 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2412 *) Check various X509_...() return values in apps/req.c.
2413 [Nils Larsch <nla@trustcenter.de>]
2415 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2416 an end-of-file condition would erronously be flagged, when the CRLF
2417 was just at the end of a processed block. The bug was discovered when
2418 processing data through a buffering memory BIO handing the data to a
2419 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2420 <ptsekov@syntrex.com> and Nedelcho Stanev.
2423 *) Implement a countermeasure against a vulnerability recently found
2424 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2425 before application data chunks to avoid the use of known IVs
2426 with data potentially chosen by the attacker.
2429 *) Fix length checks in ssl3_get_client_hello().
2432 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2433 to prevent ssl3_read_internal() from incorrectly assuming that
2434 ssl3_read_bytes() found application data while handshake
2435 processing was enabled when in fact s->s3->in_read_app_data was
2436 merely automatically cleared during the initial handshake.
2437 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2439 *) Fix object definitions for Private and Enterprise: they were not
2440 recognized in their shortname (=lowercase) representation. Extend
2441 obj_dat.pl to issue an error when using undefined keywords instead
2442 of silently ignoring the problem (Svenning Sorensen
2443 <sss@sss.dnsalias.net>).
2446 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2447 generators, i.e. generators other than 2 and 5. (Previously, the
2448 code did not properly initialise the 'add' and 'rem' values to
2449 BN_generate_prime().)
2451 In the new general case, we do not insist that 'generator' is
2452 actually a primitive root: This requirement is rather pointless;
2453 a generator of the order-q subgroup is just as good, if not
2457 *) Map new X509 verification errors to alerts. Discovered and submitted by
2458 Tom Wu <tom@arcot.com>.
2461 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2462 returning non-zero before the data has been completely received
2463 when using non-blocking I/O.
2464 [Bodo Moeller; problem pointed out by John Hughes]
2466 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2467 [Ben Laurie, Lutz Jaenicke]
2469 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2470 Yoram Zahavi <YoramZ@gilian.com>).
2473 *) Add information about CygWin 1.3 and on, and preserve proper
2474 configuration for the versions before that.
2475 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2477 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2478 check whether we deal with a copy of a session and do not delete from
2479 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2480 <izhar@checkpoint.com>.
2483 *) Do not store session data into the internal session cache, if it
2484 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2485 flag is set). Proposed by Aslam <aslam@funk.com>.
2488 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2492 *) [In 0.9.6d-engine release:]
2493 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2494 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2496 *) Add the configuration target linux-s390x.
2497 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2499 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2500 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2501 variable as an indication that a ClientHello message has been
2502 received. As the flag value will be lost between multiple
2503 invocations of ssl3_accept when using non-blocking I/O, the
2504 function may not be aware that a handshake has actually taken
2505 place, thus preventing a new session from being added to the
2508 To avoid this problem, we now set s->new_session to 2 instead of
2509 using a local variable.
2510 [Lutz Jaenicke, Bodo Moeller]
2512 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2513 if the SSL_R_LENGTH_MISMATCH error is detected.
2514 [Geoff Thorpe, Bodo Moeller]
2516 *) New 'shared_ldflag' column in Configure platform table.
2519 *) Fix EVP_CIPHER_mode macro.
2520 ["Dan S. Camper" <dan@bti.net>]
2522 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2523 type, we must throw them away by setting rr->length to 0.
2524 [D P Chang <dpc@qualys.com>]
2526 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2528 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2529 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2530 worked incorrectly for those cases where range = 10..._2 and
2531 3*range is two bits longer than range.)
2534 *) Only add signing time to PKCS7 structures if it is not already
2538 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2539 OBJ_ld_ce should be OBJ_id_ce.
2540 Also some ip-pda OIDs in crypto/objects/objects.txt were
2541 incorrect (cf. RFC 3039).
2542 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2544 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2545 returns early because it has nothing to do.
2546 [Andy Schneider <andy.schneider@bjss.co.uk>]
2548 *) [In 0.9.6c-engine release:]
2549 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2550 [Andy Schneider <andy.schneider@bjss.co.uk>]
2552 *) [In 0.9.6c-engine release:]
2553 Add support for Cryptographic Appliance's keyserver technology.
2554 (Use engine 'keyclient')
2555 [Cryptographic Appliances and Geoff Thorpe]
2557 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2558 is called via tools/c89.sh because arguments have to be
2559 rearranged (all '-L' options must appear before the first object
2561 [Richard Shapiro <rshapiro@abinitio.com>]
2563 *) [In 0.9.6c-engine release:]
2564 Add support for Broadcom crypto accelerator cards, backported
2566 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2568 *) [In 0.9.6c-engine release:]
2569 Add support for SureWare crypto accelerator cards from
2570 Baltimore Technologies. (Use engine 'sureware')
2571 [Baltimore Technologies and Mark Cox]
2573 *) [In 0.9.6c-engine release:]
2574 Add support for crypto accelerator cards from Accelerated
2575 Encryption Processing, www.aep.ie. (Use engine 'aep')
2576 [AEP Inc. and Mark Cox]
2578 *) Add a configuration entry for gcc on UnixWare.
2579 [Gary Benson <gbenson@redhat.com>]
2581 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2582 messages are stored in a single piece (fixed-length part and
2583 variable-length part combined) and fix various bugs found on the way.
2586 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2587 instead. BIO_gethostbyname() does not know what timeouts are
2588 appropriate, so entries would stay in cache even when they have
2590 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2592 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2593 faced with a pathologically small ClientHello fragment that does
2594 not contain client_version: Instead of aborting with an error,
2595 simply choose the highest available protocol version (i.e.,
2596 TLS 1.0 unless it is disabled). In practice, ClientHello
2597 messages are never sent like this, but this change gives us
2598 strictly correct behaviour at least for TLS.
2601 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2602 never resets s->method to s->ctx->method when called from within
2603 one of the SSL handshake functions.
2604 [Bodo Moeller; problem pointed out by Niko Baric]
2606 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2607 (sent using the client's version number) if client_version is
2608 smaller than the protocol version in use. Also change
2609 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2610 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2611 the client will at least see that alert.
2614 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2618 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2619 client receives HelloRequest while in a handshake.
2620 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2622 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2623 should end in 'break', not 'goto end' which circuments various
2624 cleanups done in state SSL_ST_OK. But session related stuff
2625 must be disabled for SSL_ST_OK in the case that we just sent a
2628 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2629 before just sending a HelloRequest.
2630 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2632 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2633 reveal whether illegal block cipher padding was found or a MAC
2634 verification error occured. (Neither SSLerr() codes nor alerts
2635 are directly visible to potential attackers, but the information
2636 may leak via logfiles.)
2638 Similar changes are not required for the SSL 2.0 implementation
2639 because the number of padding bytes is sent in clear for SSL 2.0,
2640 and the extra bytes are just ignored. However ssl/s2_pkt.c
2641 failed to verify that the purported number of padding bytes is in
2645 *) Add OpenUNIX-8 support including shared libraries
2646 (Boyd Lynn Gerber <gerberb@zenez.com>).
2649 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2650 'wristwatch attack' using huge encoding parameters (cf.
2651 James H. Manger's CRYPTO 2001 paper). Note that the
2652 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2653 encoding parameters and hence was not vulnerable.
2656 *) BN_sqr() bug fix.
2657 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2659 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2660 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2661 followed by modular reduction.
2662 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2664 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2665 equivalent based on BN_pseudo_rand() instead of BN_rand().
2668 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2669 This function was broken, as the check for a new client hello message
2670 to handle SGC did not allow these large messages.
2671 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2674 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2677 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2678 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2681 *) Rework the configuration and shared library support for Tru64 Unix.
2682 The configuration part makes use of modern compiler features and
2683 still retains old compiler behavior for those that run older versions
2684 of the OS. The shared library support part includes a variant that
2685 uses the RPATH feature, and is available through the special
2686 configuration target "alpha-cc-rpath", which will never be selected
2688 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2690 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2691 with the same message size as in ssl3_get_certificate_request().
2692 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2693 messages might inadvertently be reject as too long.
2694 [Petr Lampa <lampa@fee.vutbr.cz>]
2696 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2699 *) Modified SSL library such that the verify_callback that has been set
2700 specificly for an SSL object with SSL_set_verify() is actually being
2701 used. Before the change, a verify_callback set with this function was
2702 ignored and the verify_callback() set in the SSL_CTX at the time of
2703 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2704 to allow the necessary settings.
2707 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2708 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2709 done automatically (in contradiction to the requirements of the C
2710 standard). This made problems when used from OpenSSH.
2713 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2714 dh->length and always used
2716 BN_rand_range(priv_key, dh->p).
2718 BN_rand_range() is not necessary for Diffie-Hellman, and this
2719 specific range makes Diffie-Hellman unnecessarily inefficient if
2720 dh->length (recommended exponent length) is much smaller than the
2721 length of dh->p. We could use BN_rand_range() if the order of
2722 the subgroup was stored in the DH structure, but we only have
2727 BN_rand(priv_key, l, ...)
2729 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2735 RSA_eay_public_encrypt
2736 RSA_eay_private_decrypt
2737 RSA_eay_private_encrypt (signing)
2738 RSA_eay_public_decrypt (signature verification)
2740 (default implementations for RSA_public_encrypt,
2741 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2742 always reject numbers >= n.
2745 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2746 to synchronize access to 'locking_thread'. This is necessary on
2747 systems where access to 'locking_thread' (an 'unsigned long'
2748 variable) is not atomic.
2751 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2752 *before* setting the 'crypto_lock_rand' flag. The previous code had
2753 a race condition if 0 is a valid thread ID.
2754 [Travis Vitek <vitek@roguewave.com>]
2756 *) Add support for shared libraries under Irix.
2757 [Albert Chin-A-Young <china@thewrittenword.com>]
2759 *) Add configuration option to build on Linux on both big-endian and
2761 [Ralf Baechle <ralf@uni-koblenz.de>]
2763 *) Add the possibility to create shared libraries on HP-UX.
2766 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2768 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2769 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2770 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2771 PRNG state recovery was possible based on the output of
2772 one PRNG request appropriately sized to gain knowledge on
2773 'md' followed by enough consecutive 1-byte PRNG requests
2774 to traverse all of 'state'.
2776 1. When updating 'md_local' (the current thread's copy of 'md')
2777 during PRNG output generation, hash all of the previous
2778 'md_local' value, not just the half used for PRNG output.
2780 2. Make the number of bytes from 'state' included into the hash
2781 independent from the number of PRNG bytes requested.
2783 The first measure alone would be sufficient to avoid
2784 Markku-Juhani's attack. (Actually it had never occurred
2785 to me that the half of 'md_local' used for chaining was the
2786 half from which PRNG output bytes were taken -- I had always
2787 assumed that the secret half would be used.) The second
2788 measure makes sure that additional data from 'state' is never
2789 mixed into 'md_local' in small portions; this heuristically
2790 further strengthens the PRNG.
2793 *) Fix crypto/bn/asm/mips3.s.
2796 *) When only the key is given to "enc", the IV is undefined. Print out
2797 an error message in this case.
2800 *) Handle special case when X509_NAME is empty in X509 printing routines.
2803 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2804 positive and less than q.
2807 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2808 used: it isn't thread safe and the add_lock_callback should handle
2810 [Paul Rose <Paul.Rose@bridge.com>]
2812 *) Verify that incoming data obeys the block size in
2813 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2817 [Ulf Möller, Bodo Möller]
2819 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2820 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2821 when fixing the server behaviour for backwards-compatible 'client
2822 hello' messages. (Note that the attack is impractical against
2823 SSL 3.0 and TLS 1.0 anyway because length and version checking
2824 means that the probability of guessing a valid ciphertext is
2825 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2828 Before 0.9.5, the countermeasure (hide the error by generating a
2829 random 'decryption result') did not work properly because
2830 ERR_clear_error() was missing, meaning that SSL_get_error() would
2831 detect the supposedly ignored error.
2833 Both problems are now fixed.
2836 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2837 (previously it was 1024).
2840 *) Fix for compatibility mode trust settings: ignore trust settings
2841 unless some valid trust or reject settings are present.
2844 *) Fix for blowfish EVP: its a variable length cipher.
2847 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2848 parameters in DSA public key structures and return an error in the
2849 DSA routines if parameters are absent.
2852 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2853 in the current directory if neither $RANDFILE nor $HOME was set.
2854 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2855 caused some confusion to Windows users who haven't defined $HOME.
2856 Thus RAND_file_name() is changed again: e_os.h can define a
2857 DEFAULT_HOME, which will be used if $HOME is not set.
2858 For Windows, we use "C:"; on other platforms, we still require
2859 environment variables.
2861 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2862 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2863 having multiple threads call RAND_poll() concurrently.
2866 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2867 combination of a flag and a thread ID variable.
2868 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2869 flag), *other* threads can enter ssleay_add_bytes without obeying
2870 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2871 that they do not hold after the first thread unsets add_do_not_lock).
2874 *) Change bctest again: '-x' expressions are not available in all
2878 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2880 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2881 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2883 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2884 the default extension for executables, if any. Also, make the perl
2885 scripts that use symlink() to test if it really exists and use "cp"
2886 if it doesn't. All this made OpenSSL compilable and installable in
2890 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2891 If SEQUENCE is length is indefinite just set c->slen to the total
2892 amount of data available.
2893 [Steve Henson, reported by shige@FreeBSD.org]
2894 [This change does not apply to 0.9.7.]
2896 *) Change bctest to avoid here-documents inside command substitution
2897 (workaround for FreeBSD /bin/sh bug).
2898 For compatibility with Ultrix, avoid shell functions (introduced
2899 in the bctest version that searches along $PATH).
2902 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2903 with des_encrypt() defined on some operating systems, like Solaris
2907 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2908 On the Importance of Eliminating Errors in Cryptographic
2909 Computations, J. Cryptology 14 (2001) 2, 101-119,
2910 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2913 *) MIPS assembler BIGNUM division bug fix.
2916 *) Disabled incorrect Alpha assembler code.
2919 *) Fix PKCS#7 decode routines so they correctly update the length
2920 after reading an EOC for the EXPLICIT tag.
2922 [This change does not apply to 0.9.7.]
2924 *) Fix bug in PKCS#12 key generation routines. This was triggered
2925 if a 3DES key was generated with a 0 initial byte. Include
2926 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2927 (but broken) behaviour.
2930 *) Enhance bctest to search for a working bc along $PATH and print
2932 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2934 *) Fix memory leaks in err.c: free err_data string if necessary;
2935 don't write to the wrong index in ERR_set_error_data.
2938 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2942 *) Replace rdtsc with _emit statements for VC++ version 5.
2943 [Jeremy Cooper <jeremy@baymoo.org>]
2945 *) Make it possible to reuse SSLv2 sessions.
2948 *) In copy_email() check for >= 0 as a return value for
2949 X509_NAME_get_index_by_NID() since 0 is a valid index.
2950 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2952 *) Avoid coredump with unsupported or invalid public keys by checking if
2953 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2954 PKCS7_verify() fails with non detached data.
2957 *) Don't use getenv in library functions when run as setuid/setgid.
2958 New function OPENSSL_issetugid().
2961 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2962 due to incorrect handling of multi-threading:
2964 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2966 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2968 3. Count how many times MemCheck_off() has been called so that
2969 nested use can be treated correctly. This also avoids
2970 inband-signalling in the previous code (which relied on the
2971 assumption that thread ID 0 is impossible).
2974 *) Add "-rand" option also to s_client and s_server.
2977 *) Fix CPU detection on Irix 6.x.
2978 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2979 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2981 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2984 [This change does not apply to 0.9.7.]
2986 *) Use the cached encoding of an X509_NAME structure rather than
2987 copying it. This is apparently the reason for the libsafe "errors"
2988 but the code is actually correct.
2991 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2992 Bleichenbacher's DSA attack.
2993 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2994 to be set and top=0 forces the highest bit to be set; top=-1 is new
2995 and leaves the highest bit random.
2996 [Ulf Moeller, Bodo Moeller]
2998 *) In the NCONF_...-based implementations for CONF_... queries
2999 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3000 a temporary CONF structure with the data component set to NULL
3001 (which gives segmentation faults in lh_retrieve).
3002 Instead, use NULL for the CONF pointer in CONF_get_string and
3003 CONF_get_number (which may use environment variables) and directly
3004 return NULL from CONF_get_section.
3007 *) Fix potential buffer overrun for EBCDIC.
3010 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3011 keyUsage if basicConstraints absent for a CA.
3014 *) Make SMIME_write_PKCS7() write mail header values with a format that
3015 is more generally accepted (no spaces before the semicolon), since
3016 some programs can't parse those values properly otherwise. Also make
3017 sure BIO's that break lines after each write do not create invalid
3021 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3022 macros previously used would not encode an empty SEQUENCE OF
3023 and break the signature.
3025 [This change does not apply to 0.9.7.]
3027 *) Zero the premaster secret after deriving the master secret in
3031 *) Add some EVP_add_digest_alias registrations (as found in
3032 OpenSSL_add_all_digests()) to SSL_library_init()
3033 aka OpenSSL_add_ssl_algorithms(). This provides improved
3034 compatibility with peers using X.509 certificates
3035 with unconventional AlgorithmIdentifier OIDs.
3038 *) Fix for Irix with NO_ASM.
3039 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3041 *) ./config script fixes.
3042 [Ulf Moeller, Richard Levitte]
3044 *) Fix 'openssl passwd -1'.
3047 *) Change PKCS12_key_gen_asc() so it can cope with non null
3048 terminated strings whose length is passed in the passlen
3049 parameter, for example from PEM callbacks. This was done
3050 by adding an extra length parameter to asc2uni().
3051 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3053 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3054 call failed, free the DSA structure.
3057 *) Fix to uni2asc() to cope with zero length Unicode strings.
3058 These are present in some PKCS#12 files.
3061 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3062 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3063 when writing a 32767 byte record.
3064 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3066 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3067 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3069 (RSA objects have a reference count access to which is protected
3070 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3071 so they are meant to be shared between threads.)
3072 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3073 "Reddie, Steven" <Steven.Reddie@ca.com>]
3075 *) Fix a deadlock in CRYPTO_mem_leaks().
3078 *) Use better test patterns in bntest.
3081 *) rand_win.c fix for Borland C.
3084 *) BN_rshift bugfix for n == 0.
3087 *) Add a 'bctest' script that checks for some known 'bc' bugs
3088 so that 'make test' does not abort just because 'bc' is broken.
3091 *) Store verify_result within SSL_SESSION also for client side to
3092 avoid potential security hole. (Re-used sessions on the client side
3093 always resulted in verify_result==X509_V_OK, not using the original
3094 result of the server certificate verification.)
3097 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3098 SSL3_RT_APPLICATION_DATA, return 0.
3099 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3103 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3104 releases, have been re-implemented by renaming the previous
3105 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3106 and ssl3_read_internal, respectively, and adding 'peek' parameters
3107 to them. The new ssl[23]_{read,peek} functions are calls to
3108 ssl[23]_read_internal with the 'peek' flag set appropriately.
3109 A 'peek' parameter has also been added to ssl3_read_bytes, which
3110 does the actual work for ssl3_read_internal.
3113 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3114 the method-specific "init()" handler. Also clean up ex_data after
3115 calling the method-specific "finish()" handler. Previously, this was
3116 happening the other way round.
3119 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3120 The previous value, 12, was not always sufficient for BN_mod_exp().
3123 *) Make sure that shared libraries get the internal name engine with
3124 the full version number and not just 0. This should mark the
3125 shared libraries as not backward compatible. Of course, this should
3126 be changed again when we can guarantee backward binary compatibility.
3129 *) Fix typo in get_cert_by_subject() in by_dir.c
3130 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3132 *) Rework the system to generate shared libraries:
3134 - Make note of the expected extension for the shared libraries and
3135 if there is a need for symbolic links from for example libcrypto.so.0
3136 to libcrypto.so.0.9.7. There is extended info in Configure for
3139 - Make as few rebuilds of the shared libraries as possible.
3141 - Still avoid linking the OpenSSL programs with the shared libraries.
3143 - When installing, install the shared libraries separately from the
3147 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3149 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3150 and not in SSL_clear because the latter is also used by the
3151 accept/connect functions; previously, the settings made by
3152 SSL_set_read_ahead would be lost during the handshake.
3153 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3155 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3156 Previously, it would create entries for disableed algorithms no
3160 *) Added several new manual pages for SSL_* function.
3163 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3165 *) In ssl23_get_client_hello, generate an error message when faced
3166 with an initial SSL 3.0/TLS record that is too small to contain the
3167 first two bytes of the ClientHello message, i.e. client_version.
3168 (Note that this is a pathologic case that probably has never happened
3169 in real life.) The previous approach was to use the version number
3170 from the record header as a substitute; but our protocol choice
3171 should not depend on that one because it is not authenticated
3172 by the Finished messages.
3175 *) More robust randomness gathering functions for Windows.
3176 [Jeffrey Altman <jaltman@columbia.edu>]
3178 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3179 not set then we don't setup the error code for issuer check errors
3180 to avoid possibly overwriting other errors which the callback does
3181 handle. If an application does set the flag then we assume it knows
3182 what it is doing and can handle the new informational codes
3186 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3187 a general "ANY" type, as such it should be able to decode anything
3188 including tagged types. However it didn't check the class so it would
3189 wrongly interpret tagged types in the same way as their universal
3190 counterpart and unknown types were just rejected. Changed so that the
3191 tagged and unknown types are handled in the same way as a SEQUENCE:
3192 that is the encoding is stored intact. There is also a new type
3193 "V_ASN1_OTHER" which is used when the class is not universal, in this
3194 case we have no idea what the actual type is so we just lump them all
3198 *) On VMS, stdout may very well lead to a file that is written to
3199 in a record-oriented fashion. That means that every write() will
3200 write a separate record, which will be read separately by the
3201 programs trying to read from it. This can be very confusing.
3203 The solution is to put a BIO filter in the way that will buffer
3204 text until a linefeed is reached, and then write everything a
3205 line at a time, so every record written will be an actual line,
3206 not chunks of lines and not (usually doesn't happen, but I've
3207 seen it once) several lines in one record. BIO_f_linebuffer() is
3210 Currently, it's a VMS-only method, because that's where it has
3211 been tested well enough.
3214 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3215 it can return incorrect results.
3216 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3217 but it was in 0.9.6-beta[12].)
3220 *) Disable the check for content being present when verifying detached
3221 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3222 include zero length content when signing messages.
3225 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3226 BIO_ctrl (for BIO pairs).
3229 *) Add DSO method for VMS.
3232 *) Bug fix: Montgomery multiplication could produce results with the
3236 *) Add RPM specification openssl.spec and modify it to build three
3237 packages. The default package contains applications, application
3238 documentation and run-time libraries. The devel package contains
3239 include files, static libraries and function documentation. The
3240 doc package contains the contents of the doc directory. The original
3241 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3244 *) Add a large number of documentation files for many SSL routines.
3245 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3247 *) Add a configuration entry for Sony News 4.
3248 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3250 *) Don't set the two most significant bits to one when generating a
3251 random number < q in the DSA library.
3254 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3255 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3256 the underlying transport is blocking) if a handshake took place.
3257 (The default behaviour is needed by applications such as s_client
3258 and s_server that use select() to determine when to use SSL_read;
3259 but for applications that know in advance when to expect data, it
3260 just makes things more complicated.)
3263 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3267 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3268 work better on such systems.
3269 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3271 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3272 Update PKCS12_parse() so it copies the friendlyName and the
3273 keyid to the certificates aux info.
3276 *) Fix bug in PKCS7_verify() which caused an infinite loop
3277 if there was more than one signature.
3278 [Sven Uszpelkat <su@celocom.de>]
3280 *) Major change in util/mkdef.pl to include extra information
3281 about each symbol, as well as presentig variables as well
3282 as functions. This change means that there's n more need
3283 to rebuild the .num files when some algorithms are excluded.
3286 *) Allow the verify time to be set by an application,
3287 rather than always using the current time.
3290 *) Phase 2 verify code reorganisation. The certificate
3291 verify code now looks up an issuer certificate by a
3292 number of criteria: subject name, authority key id
3293 and key usage. It also verifies self signed certificates
3294 by the same criteria. The main comparison function is
3295 X509_check_issued() which performs these checks.
3297 Lot of changes were necessary in order to support this
3298 without completely rewriting the lookup code.
3300 Authority and subject key identifier are now cached.
3302 The LHASH 'certs' is X509_STORE has now been replaced
3303 by a STACK_OF(X509_OBJECT). This is mainly because an
3304 LHASH can't store or retrieve multiple objects with
3305 the same hash value.
3307 As a result various functions (which were all internal
3308 use only) have changed to handle the new X509_STORE
3309 structure. This will break anything that messed round
3310 with X509_STORE internally.
3312 The functions X509_STORE_add_cert() now checks for an
3313 exact match, rather than just subject name.
3315 The X509_STORE API doesn't directly support the retrieval
3316 of multiple certificates matching a given criteria, however
3317 this can be worked round by performing a lookup first
3318 (which will fill the cache with candidate certificates)
3319 and then examining the cache for matches. This is probably
3320 the best we can do without throwing out X509_LOOKUP
3321 entirely (maybe later...).
3323 The X509_VERIFY_CTX structure has been enhanced considerably.
3325 All certificate lookup operations now go via a get_issuer()
3326 callback. Although this currently uses an X509_STORE it
3327 can be replaced by custom lookups. This is a simple way
3328 to bypass the X509_STORE hackery necessary to make this
3329 work and makes it possible to use more efficient techniques
3330 in future. A very simple version which uses a simple
3331 STACK for its trusted certificate store is also provided
3332 using X509_STORE_CTX_trusted_stack().
3334 The verify_cb() and verify() callbacks now have equivalents
3335 in the X509_STORE_CTX structure.
3337 X509_STORE_CTX also has a 'flags' field which can be used
3338 to customise the verify behaviour.
3341 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3342 excludes S/MIME capabilities.
3345 *) When a certificate request is read in keep a copy of the
3346 original encoding of the signed data and use it when outputing
3347 again. Signatures then use the original encoding rather than
3348 a decoded, encoded version which may cause problems if the
3349 request is improperly encoded.
3352 *) For consistency with other BIO_puts implementations, call
3353 buffer_write(b, ...) directly in buffer_puts instead of calling
3356 In BIO_puts, increment b->num_write as in BIO_write.
3357 [Peter.Sylvester@EdelWeb.fr]
3359 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3360 BN_zero, we may not return a BIGNUM with an array consisting of
3364 *) Avoid calling abort() from within the library when problems are
3365 detected, except if preprocessor symbols have been defined
3366 (such as REF_CHECK, BN_DEBUG etc.).
3369 *) New openssl application 'rsautl'. This utility can be
3370 used for low level RSA operations. DER public key
3371 BIO/fp routines also added.
3374 *) New Configure entry and patches for compiling on QNX 4.
3375 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3377 *) A demo state-machine implementation was sponsored by
3378 Nuron (http://www.nuron.com/) and is now available in
3379 demos/state_machine.
3382 *) New options added to the 'dgst' utility for signature
3383 generation and verification.
3386 *) Unrecognized PKCS#7 content types are now handled via a
3387 catch all ASN1_TYPE structure. This allows unsupported
3388 types to be stored as a "blob" and an application can
3389 encode and decode it manually.
3392 *) Fix various signed/unsigned issues to make a_strex.c
3394 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3396 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3397 length if passed a buffer. ASN1_INTEGER_to_BN failed
3398 if passed a NULL BN and its argument was negative.
3399 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3401 *) Modification to PKCS#7 encoding routines to output definite
3402 length encoding. Since currently the whole structures are in
3403 memory there's not real point in using indefinite length
3404 constructed encoding. However if OpenSSL is compiled with
3405 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3408 *) Added BIO_vprintf() and BIO_vsnprintf().
3411 *) Added more prefixes to parse for in the the strings written
3412 through a logging bio, to cover all the levels that are available
3413 through syslog. The prefixes are now:
3415 PANIC, EMERG, EMR => LOG_EMERG
3416 ALERT, ALR => LOG_ALERT
3417 CRIT, CRI => LOG_CRIT
3418 ERROR, ERR => LOG_ERR
3419 WARNING, WARN, WAR => LOG_WARNING
3420 NOTICE, NOTE, NOT => LOG_NOTICE
3421 INFO, INF => LOG_INFO
3422 DEBUG, DBG => LOG_DEBUG
3424 and as before, if none of those prefixes are present at the
3425 beginning of the string, LOG_ERR is chosen.
3427 On Win32, the LOG_* levels are mapped according to this:
3429 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3430 LOG_WARNING => EVENTLOG_WARNING_TYPE
3431 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3435 *) Made it possible to reconfigure with just the configuration
3436 argument "reconf" or "reconfigure". The command line arguments
3437 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3438 and are retrieved from there when reconfiguring.
3442 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3444 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3447 *) The obj_dat.pl script was messing up the sorting of object
3448 names. The reason was that it compared the quoted version
3449 of strings as a result "OCSP" > "OCSP Signing" because
3450 " > SPACE. Changed script to store unquoted versions of
3451 names and add quotes on output. It was also omitting some
3452 names from the lookup table if they were given a default
3453 value (that is if SN is missing it is given the same
3454 value as LN and vice versa), these are now added on the
3455 grounds that if an object has a name we should be able to
3456 look it up. Finally added warning output when duplicate
3457 short or long names are found.
3460 *) Changes needed for Tandem NSK.
3461 [Scott Uroff <scott@xypro.com>]
3463 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3464 RSA_padding_check_SSLv23(), special padding was never detected
3465 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3466 version rollback attacks was not effective.
3468 In s23_clnt.c, don't use special rollback-attack detection padding
3469 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3470 client; similarly, in s23_srvr.c, don't do the rollback check if
3471 SSL 2.0 is the only protocol enabled in the server.
3474 *) Make it possible to get hexdumps of unprintable data with 'openssl
3475 asn1parse'. By implication, the functions ASN1_parse_dump() and
3476 BIO_dump_indent() are added.
3479 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3480 these print out strings and name structures based on various
3481 flags including RFC2253 support and proper handling of
3482 multibyte characters. Added options to the 'x509' utility
3483 to allow the various flags to be set.
3486 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3487 Also change the functions X509_cmp_current_time() and
3488 X509_gmtime_adj() work with an ASN1_TIME structure,
3489 this will enable certificates using GeneralizedTime in validity
3490 dates to be checked.
3493 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3494 negative public key encodings) on by default,
3495 NO_NEG_PUBKEY_BUG can be set to disable it.
3498 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3499 content octets. An i2c_ASN1_OBJECT is unnecessary because
3500 the encoding can be trivially obtained from the structure.
3503 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3504 not read locks (CRYPTO_r_[un]lock).
3507 *) A first attempt at creating official support for shared
3508 libraries through configuration. I've kept it so the
3509 default is static libraries only, and the OpenSSL programs
3510 are always statically linked for now, but there are
3511 preparations for dynamic linking in place.
3512 This has been tested on Linux and Tru64.
3515 *) Randomness polling function for Win9x, as described in:
3516 Peter Gutmann, Software Generation of Practically Strong
3520 *) Fix so PRNG is seeded in req if using an already existing
3524 *) New options to smime application. -inform and -outform
3525 allow alternative formats for the S/MIME message including
3526 PEM and DER. The -content option allows the content to be
3527 specified separately. This should allow things like Netscape
3528 form signing output easier to verify.
3531 *) Fix the ASN1 encoding of tags using the 'long form'.
3534 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3535 STRING types. These convert content octets to and from the
3536 underlying type. The actual tag and length octets are
3537 already assumed to have been read in and checked. These
3538 are needed because all other string types have virtually
3539 identical handling apart from the tag. By having versions
3540 of the ASN1 functions that just operate on content octets
3541 IMPLICIT tagging can be handled properly. It also allows
3542 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3543 and ASN1_INTEGER are identical apart from the tag.
3546 *) Change the handling of OID objects as follows:
3548 - New object identifiers are inserted in objects.txt, following
3549 the syntax given in objects.README.
3550 - objects.pl is used to process obj_mac.num and create a new
3552 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3555 This is currently kind of a hack, and the perl code in objects.pl
3556 isn't very elegant, but it works as I intended. The simplest way
3557 to check that it worked correctly is to look in obj_dat.h and
3558 check the array nid_objs and make sure the objects haven't moved
3559 around (this is important!). Additions are OK, as well as
3560 consistent name changes.
3563 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3566 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3567 The given file adds to whatever has already been seeded into the
3568 random pool through the RANDFILE configuration file option or
3569 environment variable, or the default random state file.
3572 *) mkstack.pl now sorts each macro group into lexical order.
3573 Previously the output order depended on the order the files
3574 appeared in the directory, resulting in needless rewriting
3578 *) Patches to make OpenSSL compile under Win32 again. Mostly
3579 work arounds for the VC++ problem that it treats func() as
3580 func(void). Also stripped out the parts of mkdef.pl that
3581 added extra typesafe functions: these no longer exist.
3584 *) Reorganisation of the stack code. The macros are now all
3585 collected in safestack.h . Each macro is defined in terms of
3586 a "stack macro" of the form SKM_<name>(type, a, b). The
3587 DEBUG_SAFESTACK is now handled in terms of function casts,
3588 this has the advantage of retaining type safety without the
3589 use of additional functions. If DEBUG_SAFESTACK is not defined
3590 then the non typesafe macros are used instead. Also modified the
3591 mkstack.pl script to handle the new form. Needs testing to see
3592 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3593 the default if no major problems. Similar behaviour for ASN1_SET_OF
3594 and PKCS12_STACK_OF.
3597 *) When some versions of IIS use the 'NET' form of private key the
3598 key derivation algorithm is different. Normally MD5(password) is
3599 used as a 128 bit RC4 key. In the modified case
3600 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3601 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3602 as the old Netscape_RSA functions except they have an additional
3603 'sgckey' parameter which uses the modified algorithm. Also added
3604 an -sgckey command line option to the rsa utility. Thanks to
3605 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3606 algorithm to openssl-dev.
3609 *) The evp_local.h macros were using 'c.##kname' which resulted in
3610 invalid expansion on some systems (SCO 5.0.5 for example).
3611 Corrected to 'c.kname'.
3612 [Phillip Porch <root@theporch.com>]
3614 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3615 a STACK of email addresses from a certificate or request, these look
3616 in the subject name and the subject alternative name extensions and
3617 omit any duplicate addresses.
3620 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3621 This makes DSA verification about 2 % faster.
3624 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3625 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3626 plus overhead for 1024 bit moduli).
3627 This makes exponentiations about 0.5 % faster for 1024 bit
3628 exponents (as measured by "openssl speed rsa2048").
3631 *) Rename memory handling macros to avoid conflicts with other
3633 Malloc => OPENSSL_malloc
3634 Malloc_locked => OPENSSL_malloc_locked
3635 Realloc => OPENSSL_realloc
3636 Free => OPENSSL_free
3639 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3640 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3643 *) CygWin32 support.
3644 [John Jarvie <jjarvie@newsguy.com>]
3646 *) The type-safe stack code has been rejigged. It is now only compiled
3647 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3648 by default all type-specific stack functions are "#define"d back to
3649 standard stack functions. This results in more streamlined output
3650 but retains the type-safety checking possibilities of the original
3654 *) The STACK code has been cleaned up, and certain type declarations
3655 that didn't make a lot of sense have been brought in line. This has
3656 also involved a cleanup of sorts in safestack.h to more correctly
3657 map type-safe stack functions onto their plain stack counterparts.
3658 This work has also resulted in a variety of "const"ifications of
3659 lots of the code, especially "_cmp" operations which should normally
3660 be prototyped with "const" parameters anyway.
3663 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3664 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3665 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3666 where all of 'md' is used each time the PRNG is used, but 'state'
3667 is used only indexed by a cyclic counter. As entropy may not be
3668 well distributed from the beginning, 'md' is important as a
3669 chaining variable. However, the output function chains only half
3670 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3671 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3672 in all of 'state' being rewritten, with the new values depending
3673 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3676 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3677 the handshake is continued after ssl_verify_cert_chain();
3678 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3679 can lead to 'unexplainable' connection aborts later.
3680 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3682 *) Major EVP API cipher revision.
3683 Add hooks for extra EVP features. This allows various cipher
3684 parameters to be set in the EVP interface. Support added for variable
3685 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3686 setting of RC2 and RC5 parameters.
3688 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3691 Remove lots of duplicated code from the EVP library. For example *every*
3692 cipher init() function handles the 'iv' in the same way according to the
3693 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3694 for CFB and OFB modes they zero ctx->num.
3696 New functionality allows removal of S/MIME code RC2 hack.
3698 Most of the routines have the same form and so can be declared in terms
3701 By shifting this to the top level EVP_CipherInit() it can be removed from
3702 all individual ciphers. If the cipher wants to handle IVs or keys
3703 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3706 Change lots of functions like EVP_EncryptUpdate() to now return a
3707 value: although software versions of the algorithms cannot fail
3708 any installed hardware versions can.
3711 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3712 this option is set, tolerate broken clients that send the negotiated
3713 protocol version number instead of the requested protocol version
3717 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3718 i.e. non-zero for export ciphersuites, zero otherwise.
3719 Previous versions had this flag inverted, inconsistent with
3720 rsa_tmp_cb (..._TMP_RSA_CB).
3721 [Bodo Moeller; problem reported by Amit Chopra]
3723 *) Add missing DSA library text string. Work around for some IIS
3724 key files with invalid SEQUENCE encoding.
3727 *) Add a document (doc/standards.txt) that list all kinds of standards
3728 and so on that are implemented in OpenSSL.
3731 *) Enhance c_rehash script. Old version would mishandle certificates
3732 with the same subject name hash and wouldn't handle CRLs at all.
3733 Added -fingerprint option to crl utility, to support new c_rehash
3737 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3740 *) Fix for SSL server purpose checking. Server checking was
3741 rejecting certificates which had extended key usage present
3742 but no ssl client purpose.
3743 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3745 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3746 is a little unclear about how a blank password is handled.
3747 Since the password in encoded as a BMPString with terminating
3748 double NULL a zero length password would end up as just the
3749 double NULL. However no password at all is different and is
3750 handled differently in the PKCS#12 key generation code. NS
3751 treats a blank password as zero length. MSIE treats it as no
3752 password on export: but it will try both on import. We now do
3753 the same: PKCS12_parse() tries zero length and no password if
3754 the password is set to "" or NULL (NULL is now a valid password:
3755 it wasn't before) as does the pkcs12 application.
3758 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3759 perror when PEM_read_bio_X509_REQ fails, the error message must
3760 be obtained from the error queue.
3763 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3764 it in ERR_remove_state if appropriate, and change ERR_get_state
3765 accordingly to avoid race conditions (this is necessary because
3766 thread_hash is no longer constant once set).
3769 *) Bugfix for linux-elf makefile.one.
3772 *) RSA_get_default_method() will now cause a default
3773 RSA_METHOD to be chosen if one doesn't exist already.
3774 Previously this was only set during a call to RSA_new()
3775 or RSA_new_method(NULL) meaning it was possible for
3776 RSA_get_default_method() to return NULL.
3779 *) Added native name translation to the existing DSO code
3780 that will convert (if the flag to do so is set) filenames
3781 that are sufficiently small and have no path information
3782 into a canonical native form. Eg. "blah" converted to
3783 "libblah.so" or "blah.dll" etc.
3786 *) New function ERR_error_string_n(e, buf, len) which is like
3787 ERR_error_string(e, buf), but writes at most 'len' bytes
3788 including the 0 terminator. For ERR_error_string_n, 'buf'
3790 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3792 *) CONF library reworked to become more general. A new CONF
3793 configuration file reader "class" is implemented as well as a
3794 new functions (NCONF_*, for "New CONF") to handle it. The now
3795 old CONF_* functions are still there, but are reimplemented to
3796 work in terms of the new functions. Also, a set of functions
3797 to handle the internal storage of the configuration data is
3798 provided to make it easier to write new configuration file
3799 reader "classes" (I can definitely see something reading a
3800 configuration file in XML format, for example), called _CONF_*,
3801 or "the configuration storage API"...
3803 The new configuration file reading functions are:
3805 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3806 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3808 NCONF_default, NCONF_WIN32
3810 NCONF_dump_fp, NCONF_dump_bio
3812 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3813 NCONF_new creates a new CONF object. This works in the same way
3814 as other interfaces in OpenSSL, like the BIO interface.
3815 NCONF_dump_* dump the internal storage of the configuration file,
3816 which is useful for debugging. All other functions take the same
3817 arguments as the old CONF_* functions wth the exception of the
3818 first that must be a `CONF *' instead of a `LHASH *'.
3820 To make it easer to use the new classes with the old CONF_* functions,
3821 the function CONF_set_default_method is provided.
3824 *) Add '-tls1' option to 'openssl ciphers', which was already
3825 mentioned in the documentation but had not been implemented.
3826 (This option is not yet really useful because even the additional
3827 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3830 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3831 OpenSSL-based applications) load shared libraries and bind to
3832 them in a portable way.
3833 [Geoff Thorpe, with contributions from Richard Levitte]
3835 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3837 *) Make sure _lrotl and _lrotr are only used with MSVC.
3839 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3840 (the default implementation of RAND_status).
3842 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3843 to '-clrext' (= clear extensions), as intended and documented.
3844 [Bodo Moeller; inconsistency pointed out by Michael Attili
3847 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3848 was larger than the MD block size.
3849 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3851 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3852 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3853 using the passed key: if the passed key was a private key the result
3854 of X509_print(), for example, would be to print out all the private key
3858 *) des_quad_cksum() byte order bug fix.
3859 [Ulf Möller, using the problem description in krb4-0.9.7, where
3860 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3862 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3864 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3866 *) For easily testing in shell scripts whether some command
3867 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3868 returns with exit code 0 iff no command of the given name is available.
3869 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3870 the output goes to stdout and nothing is printed to stderr.
3871 Additional arguments are always ignored.
3873 Since for each cipher there is a command of the same name,
3874 the 'no-cipher' compilation switches can be tested this way.
3876 ('openssl no-XXX' is not able to detect pseudo-commands such
3877 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3880 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3883 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3884 is set; it will be thrown away anyway because each handshake creates
3886 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3887 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3888 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3889 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3892 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3893 'Q' and 'R' lose their special meanings (quit/renegotiate).
3894 This is part of what -quiet does; unlike -quiet, -ign_eof
3895 does not suppress any output.
3898 *) Add compatibility options to the purpose and trust code. The
3899 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3900 accepts a certificate or CA, this was the previous behaviour,
3901 with all the associated security issues.
3903 X509_TRUST_COMPAT is the old trust behaviour: only and
3904 automatically trust self signed roots in certificate store. A
3905 new trust setting X509_TRUST_DEFAULT is used to specify that
3906 a purpose has no associated trust setting and it should instead
3907 use the value in the default purpose.
3910 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3911 and fix a memory leak.
3914 *) In util/mkerr.pl (which implements 'make errors'), preserve
3915 reason strings from the previous version of the .c file, as
3916 the default to have only downcase letters (and digits) in
3917 automatically generated reasons codes is not always appropriate.
3920 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3921 using strerror. Previously, ERR_reason_error_string() returned
3922 library names as reason strings for SYSerr; but SYSerr is a special
3923 case where small numbers are errno values, not library numbers.
3926 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3927 converts DSA parameters into DH parameters. (When creating parameters,
3928 DSA_generate_parameters is used.)
3931 *) Include 'length' (recommended exponent length) in C code generated
3932 by 'openssl dhparam -C'.
3935 *) The second argument to set_label in perlasm was already being used
3936 so couldn't be used as a "file scope" flag. Moved to third argument
3940 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3941 instead of RAND_bytes for encryption IVs and salts.
3944 *) Include RAND_status() into RAND_METHOD instead of implementing
3945 it only for md_rand.c Otherwise replacing the PRNG by calling
3946 RAND_set_rand_method would be impossible.
3949 *) Don't let DSA_generate_key() enter an infinite loop if the random
3950 number generation fails.
3953 *) New 'rand' application for creating pseudo-random output.
3956 *) Added configuration support for Linux/IA64
3957 [Rolf Haberrecker <rolf@suse.de>]
3959 *) Assembler module support for Mingw32.
3962 *) Shared library support for HPUX (in shlib/).
3963 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3965 *) Shared library support for Solaris gcc.
3966 [Lutz Behnke <behnke@trustcenter.de>]
3968 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3970 *) PKCS7_encrypt() was adding text MIME headers twice because they
3971 were added manually and by SMIME_crlf_copy().
3974 *) In bntest.c don't call BN_rand with zero bits argument.
3975 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3977 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3978 case was implemented. This caused BN_div_recp() to fail occasionally.
3981 *) Add an optional second argument to the set_label() in the perl
3982 assembly language builder. If this argument exists and is set
3983 to 1 it signals that the assembler should use a symbol whose
3984 scope is the entire file, not just the current function. This
3985 is needed with MASM which uses the format label:: for this scope.
3986 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3988 *) Change the ASN1 types so they are typedefs by default. Before
3989 almost all types were #define'd to ASN1_STRING which was causing
3990 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3994 *) Change names of new functions to the new get1/get0 naming
3995 convention: After 'get1', the caller owns a reference count
3996 and has to call ..._free; 'get0' returns a pointer to some
3997 data structure without incrementing reference counters.
3998 (Some of the existing 'get' functions increment a reference
3999 counter, some don't.)
4000 Similarly, 'set1' and 'add1' functions increase reference
4001 counters or duplicate objects.
4004 *) Allow for the possibility of temp RSA key generation failure:
4005 the code used to assume it always worked and crashed on failure.
4008 *) Fix potential buffer overrun problem in BIO_printf().
4009 [Ulf Möller, using public domain code by Patrick Powell; problem
4010 pointed out by David Sacerdote <das33@cornell.edu>]
4012 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4013 RAND_egd() and RAND_status(). In the command line application,
4014 the EGD socket can be specified like a seed file using RANDFILE
4018 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4019 Some CAs (e.g. Verisign) distribute certificates in this form.
4022 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4023 list to exclude them. This means that no special compilation option
4024 is needed to use anonymous DH: it just needs to be included in the
4028 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4029 EVP_MD_type. The old functionality is available in a new macro called
4030 EVP_MD_md(). Change code that uses it and update docs.
4033 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4034 where the 'void *' argument is replaced by a function pointer argument.
4035 Previously 'void *' was abused to point to functions, which works on
4036 many platforms, but is not correct. As these functions are usually
4037 called by macros defined in OpenSSL header files, most source code
4038 should work without changes.
4041 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4042 sections with information on -D... compiler switches used for
4043 compiling the library so that applications can see them. To enable
4044 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4045 must be defined. E.g.,
4046 #define OPENSSL_ALGORITHM_DEFINES
4047 #include <openssl/opensslconf.h>
4048 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4049 [Richard Levitte, Ulf and Bodo Möller]
4051 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4055 *) Change the 'other' type in certificate aux info to a STACK_OF
4056 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4057 the required ASN1 format: arbitrary types determined by an OID.
4060 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4061 argument to 'req'. This is not because the function is newer or
4062 better than others it just uses the work 'NEW' in the certificate
4063 request header lines. Some software needs this.
4066 *) Reorganise password command line arguments: now passwords can be
4067 obtained from various sources. Delete the PEM_cb function and make
4068 it the default behaviour: i.e. if the callback is NULL and the
4069 usrdata argument is not NULL interpret it as a null terminated pass
4070 phrase. If usrdata and the callback are NULL then the pass phrase
4071 is prompted for as usual.
4074 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4075 the support is automatically enabled. The resulting binaries will
4076 autodetect the card and use it if present.
4077 [Ben Laurie and Compaq Inc.]
4079 *) Work around for Netscape hang bug. This sends certificate request
4080 and server done in one record. Since this is perfectly legal in the
4081 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4082 the bugs/SSLv3 entry for more info.
4085 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4088 *) Add -rand argument to smime and pkcs12 applications and read/write
4092 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4095 *) Add command line password options to the remaining applications.
4098 *) Bug fix for BN_div_recp() for numerators with an even number of
4102 *) More tests in bntest.c, and changed test_bn output.
4105 *) ./config recognizes MacOS X now.
4108 *) Bug fix for BN_div() when the first words of num and divsor are
4109 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4112 *) Add support for various broken PKCS#8 formats, and command line
4113 options to produce them.
4116 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4117 get temporary BIGNUMs from a BN_CTX.
4120 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4124 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4125 include a #define from the old name to the new. The original intent
4126 was that statically linked binaries could for example just call
4127 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4128 link with digests. This never worked becayse SSLeay_add_all_digests()
4129 and SSLeay_add_all_ciphers() were in the same source file so calling
4130 one would link with the other. They are now in separate source files.
4133 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4136 *) Use a less unusual form of the Miller-Rabin primality test (it used
4137 a binary algorithm for exponentiation integrated into the Miller-Rabin
4138 loop, our standard modexp algorithms are faster).
4141 *) Support for the EBCDIC character set completed.
4142 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4144 *) Source code cleanups: use const where appropriate, eliminate casts,
4145 use void * instead of char * in lhash.
4148 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4149 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4150 this the server could overwrite ephemeral keys that the client
4154 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4155 using 50 iterations of the Rabin-Miller test.
4157 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4158 iterations of the Rabin-Miller test as required by the appendix
4159 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4160 As BN_is_prime_fasttest includes trial division, DSA parameter
4161 generation becomes much faster.
4163 This implies a change for the callback functions in DSA_is_prime
4164 and DSA_generate_parameters: The callback function is called once
4165 for each positive witness in the Rabin-Miller test, not just
4166 occasionally in the inner loop; and the parameters to the
4167 callback function now provide an iteration count for the outer
4168 loop rather than for the current invocation of the inner loop.
4169 DSA_generate_parameters additionally can call the callback
4170 function with an 'iteration count' of -1, meaning that a
4171 candidate has passed the trial division test (when q is generated
4172 from an application-provided seed, trial division is skipped).
4175 *) New function BN_is_prime_fasttest that optionally does trial
4176 division before starting the Rabin-Miller test and has
4177 an additional BN_CTX * argument (whereas BN_is_prime always
4178 has to allocate at least one BN_CTX).
4179 'callback(1, -1, cb_arg)' is called when a number has passed the
4180 trial division stage.
4183 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4187 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4190 *) New function BN_pseudo_rand().
4193 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4194 bignum version of BN_from_montgomery() with the working code from
4195 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4199 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4200 made it impossible to use the same SSL_SESSION data structure in
4201 SSL2 clients in multiple threads.
4204 *) The return value of RAND_load_file() no longer counts bytes obtained
4205 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4206 to seed the PRNG (previously an explicit byte count was required).
4207 [Ulf Möller, Bodo Möller]
4209 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4210 used (char *) instead of (void *) and had casts all over the place.
4213 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4216 *) Retain source code compatibility for BN_prime_checks macro:
4217 BN_is_prime(..., BN_prime_checks, ...) now uses
4218 BN_prime_checks_for_size to determine the appropriate number of
4219 Rabin-Miller iterations.
4222 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4223 DH_CHECK_P_NOT_SAFE_PRIME.
4224 (Check if this is true? OpenPGP calls them "strong".)
4227 *) Merge the functionality of "dh" and "gendh" programs into a new program
4228 "dhparam". The old programs are retained for now but will handle DH keys
4229 (instead of parameters) in future.
4232 *) Make the ciphers, s_server and s_client programs check the return values
4233 when a new cipher list is set.
4236 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4237 ciphers. Before when the 56bit ciphers were enabled the sorting was
4240 The syntax for the cipher sorting has been extended to support sorting by
4241 cipher-strength (using the strength_bits hard coded in the tables).
4242 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4244 Fix a bug in the cipher-command parser: when supplying a cipher command
4245 string with an "undefined" symbol (neither command nor alphanumeric
4246 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4247 an error is flagged.
4249 Due to the strength-sorting extension, the code of the
4250 ssl_create_cipher_list() function was completely rearranged. I hope that
4251 the readability was also increased :-)
4252 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4254 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4255 for the first serial number and places 2 in the serial number file. This
4256 avoids problems when the root CA is created with serial number zero and
4257 the first user certificate has the same issuer name and serial number
4261 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4262 the new code. Add documentation for this stuff.
4265 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4266 X509_*() to X509at_*() on the grounds that they don't handle X509
4267 structures and behave in an analagous way to the X509v3 functions:
4268 they shouldn't be called directly but wrapper functions should be used
4271 So we also now have some wrapper functions that call the X509at functions
4272 when passed certificate requests. (TO DO: similar things can be done with
4273 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4274 things. Some of these need some d2i or i2d and print functionality
4275 because they handle more complex structures.)
4278 *) Add missing #ifndefs that caused missing symbols when building libssl
4279 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4280 NO_RSA in ssl/s2*.c.
4281 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4283 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4284 has a return value which indicates the quality of the random data
4285 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4286 error queue. New function RAND_pseudo_bytes() generates output that is
4287 guaranteed to be unique but not unpredictable. RAND_add is like
4288 RAND_seed, but takes an extra argument for an entropy estimate
4289 (RAND_seed always assumes full entropy).
4292 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4293 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4294 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4295 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4296 false-positive rate of at most 2^-80 for random input.
4299 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4302 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4303 in the 0.9.5 release), this returns the chain
4304 from an X509_CTX structure with a dup of the stack and all
4305 the X509 reference counts upped: so the stack will exist
4306 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4309 Also make SSL_SESSION_print() print out the verify return
4313 *) Add manpage for the pkcs12 command. Also change the default
4314 behaviour so MAC iteration counts are used unless the new
4315 -nomaciter option is used. This improves file security and
4316 only older versions of MSIE (4.0 for example) need it.
4319 *) Honor the no-xxx Configure options when creating .DEF files.
4322 *) Add PKCS#10 attributes to field table: challengePassword,
4323 unstructuredName and unstructuredAddress. These are taken from
4324 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4325 international characters are used.
4327 More changes to X509_ATTRIBUTE code: allow the setting of types
4328 based on strings. Remove the 'loc' parameter when adding
4329 attributes because these will be a SET OF encoding which is sorted
4333 *) Initial changes to the 'req' utility to allow request generation
4334 automation. This will allow an application to just generate a template
4335 file containing all the field values and have req construct the
4338 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4339 used all over the place including certificate requests and PKCS#7
4340 structures. They are currently handled manually where necessary with
4341 some primitive wrappers for PKCS#7. The new functions behave in a
4342 manner analogous to the X509 extension functions: they allow
4343 attributes to be looked up by NID and added.
4345 Later something similar to the X509V3 code would be desirable to
4346 automatically handle the encoding, decoding and printing of the
4347 more complex types. The string types like challengePassword can
4348 be handled by the string table functions.
4350 Also modified the multi byte string table handling. Now there is
4351 a 'global mask' which masks out certain types. The table itself
4352 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4353 is useful when for example there is only one permissible type
4354 (as in countryName) and using the mask might result in no valid
4358 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4359 SSL_get_peer_finished to allow applications to obtain the latest
4360 Finished messages sent to the peer or expected from the peer,
4361 respectively. (SSL_get_peer_finished is usually the Finished message
4362 actually received from the peer, otherwise the protocol will be aborted.)
4364 As the Finished message are message digests of the complete handshake
4365 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4366 be used for external authentication procedures when the authentication
4367 provided by SSL/TLS is not desired or is not enough.
4370 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4371 the host supports BWX extension and if Compaq C is present on the
4372 $PATH. Just exploiting of the BWX extension results in 20-30%
4373 performance kick for some algorithms, e.g. DES and RC4 to mention
4374 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4378 *) Add support for MS "fast SGC". This is arguably a violation of the
4379 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4380 weak crypto and after checking the certificate is SGC a second one
4381 with strong crypto. MS SGC stops the first handshake after receiving
4382 the server certificate message and sends a second client hello. Since
4383 a server will typically do all the time consuming operations before
4384 expecting any further messages from the client (server key exchange
4385 is the most expensive) there is little difference between the two.
4387 To get OpenSSL to support MS SGC we have to permit a second client
4388 hello message after we have sent server done. In addition we have to
4389 reset the MAC if we do get this second client hello.
4392 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4393 if a DER encoded private key is RSA or DSA traditional format. Changed
4394 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4395 format DER encoded private key. Newer code should use PKCS#8 format which
4396 has the key type encoded in the ASN1 structure. Added DER private key
4397 support to pkcs8 application.
4400 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4401 ciphersuites has been selected (as required by the SSL 3/TLS 1
4402 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4403 is set, we interpret this as a request to violate the specification
4404 (the worst that can happen is a handshake failure, and 'correct'
4405 behaviour would result in a handshake failure anyway).
4408 *) In SSL_CTX_add_session, take into account that there might be multiple
4409 SSL_SESSION structures with the same session ID (e.g. when two threads
4410 concurrently obtain them from an external cache).
4411 The internal cache can handle only one SSL_SESSION with a given ID,
4412 so if there's a conflict, we now throw out the old one to achieve
4416 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4417 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4418 some routines that use cipher OIDs: some ciphers do not have OIDs
4419 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4423 *) Simplify the trust setting structure and code. Now we just have
4424 two sequences of OIDs for trusted and rejected settings. These will
4425 typically have values the same as the extended key usage extension
4426 and any application specific purposes.
4428 The trust checking code now has a default behaviour: it will just
4429 check for an object with the same NID as the passed id. Functions can
4430 be provided to override either the default behaviour or the behaviour
4431 for a given id. SSL client, server and email already have functions
4432 in place for compatibility: they check the NID and also return "trusted"
4433 if the certificate is self signed.
4436 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4437 traditional format into an EVP_PKEY structure.
4440 *) Add a password callback function PEM_cb() which either prompts for
4441 a password if usr_data is NULL or otherwise assumes it is a null
4442 terminated password. Allow passwords to be passed on command line
4443 environment or config files in a few more utilities.
4446 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4447 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4448 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4449 Update documentation.
4452 *) Support for ASN1 "NULL" type. This could be handled before by using
4453 ASN1_TYPE but there wasn't any function that would try to read a NULL
4454 and produce an error if it couldn't. For compatibility we also have
4455 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4456 don't allocate anything because they don't need to.
4459 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4461 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4463 *) Rebuild of the memory allocation routines used by OpenSSL code and
4464 possibly others as well. The purpose is to make an interface that
4465 provide hooks so anyone can build a separate set of allocation and
4466 deallocation routines to be used by OpenSSL, for example memory
4467 pool implementations, or something else, which was previously hard
4468 since Malloc(), Realloc() and Free() were defined as macros having
4469 the values malloc, realloc and free, respectively (except for Win32
4470 compilations). The same is provided for memory debugging code.
4471 OpenSSL already comes with functionality to find memory leaks, but
4472 this gives people a chance to debug other memory problems.
4474 With these changes, a new set of functions and macros have appeared:
4476 CRYPTO_set_mem_debug_functions() [F]
4477 CRYPTO_get_mem_debug_functions() [F]
4478 CRYPTO_dbg_set_options() [F]
4479 CRYPTO_dbg_get_options() [F]
4480 CRYPTO_malloc_debug_init() [M]
4482 The memory debug functions are NULL by default, unless the library
4483 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4484 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4485 gives the standard debugging functions that come with OpenSSL) or
4486 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4487 provided by the library user) must be used. When the standard
4488 debugging functions are used, CRYPTO_dbg_set_options can be used to
4489 request additional information:
4490 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4491 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4493 Also, things like CRYPTO_set_mem_functions will always give the
4494 expected result (the new set of functions is used for allocation
4495 and deallocation) at all times, regardless of platform and compiler
4498 To finish it up, some functions that were never use in any other
4499 way than through macros have a new API and new semantic:
4502 CRYPTO_dbg_realloc()
4505 All macros of value have retained their old syntax.
4506 [Richard Levitte and Bodo Moeller]
4508 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4509 ordering of SMIMECapabilities wasn't in "strength order" and there
4510 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4514 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4515 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4516 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4518 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4519 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4520 functionality to handle multipart/signed properly) and a utility
4521 called 'smime' to call all this stuff. This is based on code I
4522 originally wrote for Celo who have kindly allowed it to be
4523 included in OpenSSL.
4526 *) Add variants des_set_key_checked and des_set_key_unchecked of
4527 des_set_key (aka des_key_sched). Global variable des_check_key
4528 decides which of these is called by des_set_key; this way
4529 des_check_key behaves as it always did, but applications and
4530 the library itself, which was buggy for des_check_key == 1,
4531 have a cleaner way to pick the version they need.
4534 *) New function PKCS12_newpass() which changes the password of a
4538 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4539 dynamic mix. In both cases the ids can be used as an index into the
4540 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4541 functions so they accept a list of the field values and the
4542 application doesn't need to directly manipulate the X509_TRUST
4546 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4550 *) Modify the way the V3 extension code looks up extensions. This now
4551 works in a similar way to the object code: we have some "standard"
4552 extensions in a static table which is searched with OBJ_bsearch()
4553 and the application can add dynamic ones if needed. The file
4554 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4555 updated whenever a new extension is added to the core code and kept
4556 in ext_nid order. There is a simple program 'tabtest.c' which checks
4557 this. New extensions are not added too often so this file can readily
4558 be maintained manually.
4560 There are two big advantages in doing things this way. The extensions
4561 can be looked up immediately and no longer need to be "added" using
4562 X509V3_add_standard_extensions(): this function now does nothing.
4563 [Side note: I get *lots* of email saying the extension code doesn't
4564 work because people forget to call this function]
4565 Also no dynamic allocation is done unless new extensions are added:
4566 so if we don't add custom extensions there is no need to call
4567 X509V3_EXT_cleanup().
4570 *) Modify enc utility's salting as follows: make salting the default. Add a
4571 magic header, so unsalted files fail gracefully instead of just decrypting
4572 to garbage. This is because not salting is a big security hole, so people
4573 should be discouraged from doing it.
4576 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4577 digest to be passed on the command line but it only used this
4578 parameter when signing a certificate. Modified so all relevant
4579 operations are affected by the digest parameter including the
4580 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4581 DSA key was used because it didn't fix the digest.
4584 *) Initial certificate chain verify code. Currently tests the untrusted
4585 certificates for consistency with the verify purpose (which is set
4586 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4588 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4589 this is because it will reject chains with invalid extensions whereas
4590 every previous version of OpenSSL and SSLeay made no checks at all.
4592 Trust code: checks the root CA for the relevant trust settings. Trust
4593 settings have an initial value consistent with the verify purpose: e.g.
4594 if the verify purpose is for SSL client use it expects the CA to be
4595 trusted for SSL client use. However the default value can be changed to
4596 permit custom trust settings: one example of this would be to only trust
4597 certificates from a specific "secure" set of CAs.
4599 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4600 which should be used for version portability: especially since the
4601 verify structure is likely to change more often now.
4603 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4604 to set them. If not set then assume SSL clients will verify SSL servers
4607 Two new options to the verify program: -untrusted allows a set of
4608 untrusted certificates to be passed in and -purpose which sets the
4609 intended purpose of the certificate. If a purpose is set then the
4610 new chain verify code is used to check extension consistency.
4613 *) Support for the authority information access extension.
4616 *) Modify RSA and DSA PEM read routines to transparently handle
4617 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4618 public keys in a format compatible with certificate
4619 SubjectPublicKeyInfo structures. Unfortunately there were already
4620 functions called *_PublicKey_* which used various odd formats so
4621 these are retained for compatibility: however the DSA variants were
4622 never in a public release so they have been deleted. Changed dsa/rsa
4623 utilities to handle the new format: note no releases ever handled public
4624 keys so we should be OK.
4626 The primary motivation for this change is to avoid the same fiasco
4627 that dogs private keys: there are several incompatible private key
4628 formats some of which are standard and some OpenSSL specific and
4629 require various evil hacks to allow partial transparent handling and
4630 even then it doesn't work with DER formats. Given the option anything
4631 other than PKCS#8 should be dumped: but the other formats have to
4632 stay in the name of compatibility.
4634 With public keys and the benefit of hindsight one standard format
4635 is used which works with EVP_PKEY, RSA or DSA structures: though
4636 it clearly returns an error if you try to read the wrong kind of key.
4638 Added a -pubkey option to the 'x509' utility to output the public key.
4639 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4640 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4641 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4642 that do the same as the EVP_PKEY_assign_*() except they up the
4643 reference count of the added key (they don't "swallow" the
4647 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4648 CRLs would fail if the file contained no certificates or no CRLs:
4649 added a new function to read in both types and return the number
4650 read: this means that if none are read it will be an error. The
4651 DER versions of the certificate and CRL reader would always fail
4652 because it isn't possible to mix certificates and CRLs in DER format
4653 without choking one or the other routine. Changed this to just read
4654 a certificate: this is the best we can do. Also modified the code
4655 in apps/verify.c to take notice of return codes: it was previously
4656 attempting to read in certificates from NULL pointers and ignoring
4657 any errors: this is one reason why the cert and CRL reader seemed
4658 to work. It doesn't check return codes from the default certificate
4659 routines: these may well fail if the certificates aren't installed.
4662 *) Code to support otherName option in GeneralName.
4665 *) First update to verify code. Change the verify utility
4666 so it warns if it is passed a self signed certificate:
4667 for consistency with the normal behaviour. X509_verify
4668 has been modified to it will now verify a self signed
4669 certificate if *exactly* the same certificate appears
4670 in the store: it was previously impossible to trust a
4671 single self signed certificate. This means that:
4672 openssl verify ss.pem
4673 now gives a warning about a self signed certificate but
4674 openssl verify -CAfile ss.pem ss.pem
4678 *) For servers, store verify_result in SSL_SESSION data structure
4679 (and add it to external session representation).
4680 This is needed when client certificate verifications fails,
4681 but an application-provided verification callback (set by
4682 SSL_CTX_set_cert_verify_callback) allows accepting the session
4683 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4684 but returns 1): When the session is reused, we have to set
4685 ssl->verify_result to the appropriate error code to avoid
4687 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4689 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4690 case in PKCS7_dataInit() where the signed PKCS7 structure
4691 didn't contain any existing data because it was being created.
4692 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4694 *) Add a salt to the key derivation routines in enc.c. This
4695 forms the first 8 bytes of the encrypted file. Also add a
4696 -S option to allow a salt to be input on the command line.
4699 *) New function X509_cmp(). Oddly enough there wasn't a function
4700 to compare two certificates. We do this by working out the SHA1
4701 hash and comparing that. X509_cmp() will be needed by the trust
4705 *) SSL_get1_session() is like SSL_get_session(), but increments
4706 the reference count in the SSL_SESSION returned.
4707 [Geoff Thorpe <geoff@eu.c2.net>]
4709 *) Fix for 'req': it was adding a null to request attributes.
4710 Also change the X509_LOOKUP and X509_INFO code to handle
4711 certificate auxiliary information.
4714 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4718 *) Add the possibility to add extra information to the memory leak
4719 detecting output, to form tracebacks, showing from where each
4720 allocation was originated: CRYPTO_push_info("constant string") adds
4721 the string plus current file name and line number to a per-thread
4722 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4723 is like calling CYRPTO_pop_info() until the stack is empty.
4724 Also updated memory leak detection code to be multi-thread-safe.
4727 *) Add options -text and -noout to pkcs7 utility and delete the
4728 encryption options which never did anything. Update docs.
4731 *) Add options to some of the utilities to allow the pass phrase
4732 to be included on either the command line (not recommended on
4733 OSes like Unix) or read from the environment. Update the
4734 manpages and fix a few bugs.
4737 *) Add a few manpages for some of the openssl commands.
4740 *) Fix the -revoke option in ca. It was freeing up memory twice,
4741 leaking and not finding already revoked certificates.
4744 *) Extensive changes to support certificate auxiliary information.
4745 This involves the use of X509_CERT_AUX structure and X509_AUX
4746 functions. An X509_AUX function such as PEM_read_X509_AUX()
4747 can still read in a certificate file in the usual way but it
4748 will also read in any additional "auxiliary information". By
4749 doing things this way a fair degree of compatibility can be
4750 retained: existing certificates can have this information added
4751 using the new 'x509' options.
4753 Current auxiliary information includes an "alias" and some trust
4754 settings. The trust settings will ultimately be used in enhanced
4755 certificate chain verification routines: currently a certificate
4756 can only be trusted if it is self signed and then it is trusted
4760 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4761 The problem was that one of the replacement routines had not been working
4762 since SSLeay releases. For now the offending routine has been replaced
4763 with non-optimised assembler. Even so, this now gives around 95%
4764 performance improvement for 1024 bit RSA signs.
4767 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4768 handling. Most clients have the effective key size in bits equal to
4769 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4770 A few however don't do this and instead use the size of the decrypted key
4771 to determine the RC2 key length and the AlgorithmIdentifier to determine
4772 the effective key length. In this case the effective key length can still
4773 be 40 bits but the key length can be 168 bits for example. This is fixed
4774 by manually forcing an RC2 key into the EVP_PKEY structure because the
4775 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4776 the key length and effective key length are equal.
4779 *) Add a bunch of functions that should simplify the creation of
4780 X509_NAME structures. Now you should be able to do:
4781 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4782 and have it automatically work out the correct field type and fill in
4783 the structures. The more adventurous can try:
4784 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4785 and it will (hopefully) work out the correct multibyte encoding.
4788 *) Change the 'req' utility to use the new field handling and multibyte
4789 copy routines. Before the DN field creation was handled in an ad hoc
4790 way in req, ca, and x509 which was rather broken and didn't support
4791 BMPStrings or UTF8Strings. Since some software doesn't implement
4792 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4793 using the dirstring_type option. See the new comment in the default
4794 openssl.cnf for more info.
4797 *) Make crypto/rand/md_rand.c more robust:
4798 - Assure unique random numbers after fork().
4799 - Make sure that concurrent threads access the global counter and
4800 md serializably so that we never lose entropy in them
4801 or use exactly the same state in multiple threads.
4802 Access to the large state is not always serializable because
4803 the additional locking could be a performance killer, and
4804 md should be large enough anyway.
4807 *) New file apps/app_rand.c with commonly needed functionality
4808 for handling the random seed file.
4810 Use the random seed file in some applications that previously did not:
4812 dsaparam -genkey (which also ignored its '-rand' option),
4815 x509 (when signing).
4816 Except on systems with /dev/urandom, it is crucial to have a random
4817 seed file at least for key creation, DSA signing, and for DH exchanges;
4818 for RSA signatures we could do without one.
4820 gendh and gendsa (unlike genrsa) used to read only the first byte
4821 of each file listed in the '-rand' option. The function as previously
4822 found in genrsa is now in app_rand.c and is used by all programs
4823 that support '-rand'.
4826 *) In RAND_write_file, use mode 0600 for creating files;
4827 don't just chmod when it may be too late.
4830 *) Report an error from X509_STORE_load_locations
4831 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4834 *) New function ASN1_mbstring_copy() this copies a string in either
4835 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4836 into an ASN1_STRING type. A mask of permissible types is passed
4837 and it chooses the "minimal" type to use or an error if not type
4841 *) Add function equivalents to the various macros in asn1.h. The old
4842 macros are retained with an M_ prefix. Code inside the library can
4843 use the M_ macros. External code (including the openssl utility)
4844 should *NOT* in order to be "shared library friendly".
4847 *) Add various functions that can check a certificate's extensions
4848 to see if it usable for various purposes such as SSL client,
4849 server or S/MIME and CAs of these types. This is currently
4850 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4851 verification. Also added a -purpose flag to x509 utility to
4852 print out all the purposes.
4855 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4859 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4860 for, obtain and decode and extension and obtain its critical flag.
4861 This allows all the necessary extension code to be handled in a
4862 single function call.
4865 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4866 platforms. See crypto/rc4/rc4_enc.c for further details.
4869 *) New -noout option to asn1parse. This causes no output to be produced
4870 its main use is when combined with -strparse and -out to extract data
4871 from a file (which may not be in ASN.1 format).
4874 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4875 when producing the local key id.
4876 [Richard Levitte <levitte@stacken.kth.se>]
4878 *) New option -dhparam in s_server. This allows a DH parameter file to be
4879 stated explicitly. If it is not stated then it tries the first server
4880 certificate file. The previous behaviour hard coded the filename
4884 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4885 a public key to be input or output. For example:
4886 openssl rsa -in key.pem -pubout -out pubkey.pem
4887 Also added necessary DSA public key functions to handle this.
4890 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4891 in the message. This was handled by allowing
4892 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4893 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4895 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4896 to the end of the strings whereas this didn't. This would cause problems
4897 if strings read with d2i_ASN1_bytes() were later modified.
4898 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4900 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4901 data and it contains EOF it will end up returning an error. This is
4902 caused by input 46 bytes long. The cause is due to the way base64
4903 BIOs find the start of base64 encoded data. They do this by trying a
4904 trial decode on each line until they find one that works. When they
4905 do a flag is set and it starts again knowing it can pass all the
4906 data directly through the decoder. Unfortunately it doesn't reset
4907 the context it uses. This means that if EOF is reached an attempt
4908 is made to pass two EOFs through the context and this causes the
4909 resulting error. This can also cause other problems as well. As is
4910 usual with these problems it takes *ages* to find and the fix is
4911 trivial: move one line.
4912 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4914 *) Ugly workaround to get s_client and s_server working under Windows. The
4915 old code wouldn't work because it needed to select() on sockets and the
4916 tty (for keypresses and to see if data could be written). Win32 only
4917 supports select() on sockets so we select() with a 1s timeout on the
4918 sockets and then see if any characters are waiting to be read, if none
4919 are present then we retry, we also assume we can always write data to
4920 the tty. This isn't nice because the code then blocks until we've
4921 received a complete line of data and it is effectively polling the
4922 keyboard at 1s intervals: however it's quite a bit better than not
4923 working at all :-) A dedicated Windows application might handle this
4924 with an event loop for example.
4927 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4928 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4929 will be called when RSA_sign() and RSA_verify() are used. This is useful
4930 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4931 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4932 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4933 This necessitated the support of an extra signature type NID_md5_sha1
4934 for SSL signatures and modifications to the SSL library to use it instead
4935 of calling RSA_public_decrypt() and RSA_private_encrypt().
4938 *) Add new -verify -CAfile and -CApath options to the crl program, these
4939 will lookup a CRL issuers certificate and verify the signature in a
4940 similar way to the verify program. Tidy up the crl program so it
4941 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4942 less strict. It will now permit CRL extensions even if it is not
4943 a V2 CRL: this will allow it to tolerate some broken CRLs.
4946 *) Initialize all non-automatic variables each time one of the openssl
4947 sub-programs is started (this is necessary as they may be started
4948 multiple times from the "OpenSSL>" prompt).
4949 [Lennart Bang, Bodo Moeller]
4951 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4952 removing all other RSA functionality (this is what NO_RSA does). This
4953 is so (for example) those in the US can disable those operations covered
4954 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4958 *) Non-copying interface to BIO pairs.
4959 (still largely untested)
4962 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4963 ASCII string. This was handled independently in various places before.
4966 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4967 UTF8 strings a character at a time.
4970 *) Use client_version from client hello to select the protocol
4971 (s23_srvr.c) and for RSA client key exchange verification
4972 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4975 *) Add various utility functions to handle SPKACs, these were previously
4976 handled by poking round in the structure internals. Added new function
4977 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4978 print, verify and generate SPKACs. Based on an original idea from
4979 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4982 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4985 *) Allow the config file extension section to be overwritten on the
4986 command line. Based on an original idea from Massimiliano Pala
4987 <madwolf@comune.modena.it>. The new option is called -extensions
4988 and can be applied to ca, req and x509. Also -reqexts to override
4989 the request extensions in req and -crlexts to override the crl extensions
4993 *) Add new feature to the SPKAC handling in ca. Now you can include
4994 the same field multiple times by preceding it by "XXXX." for example:
4997 this is the same syntax as used in the req config file.
5000 *) Allow certificate extensions to be added to certificate requests. These
5001 are specified in a 'req_extensions' option of the req section of the
5002 config file. They can be printed out with the -text option to req but
5003 are otherwise ignored at present.
5006 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5007 data read consists of only the final block it would not decrypted because
5008 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5009 A misplaced 'break' also meant the decrypted final block might not be
5010 copied until the next read.
5013 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5014 a few extra parameters to the DH structure: these will be useful if
5015 for example we want the value of 'q' or implement X9.42 DH.
5018 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5019 provides hooks that allow the default DSA functions or functions on a
5020 "per key" basis to be replaced. This allows hardware acceleration and
5021 hardware key storage to be handled without major modification to the
5022 library. Also added low level modexp hooks and CRYPTO_EX structure and
5023 associated functions.
5026 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5027 as "read only": it can't be written to and the buffer it points to will
5028 not be freed. Reading from a read only BIO is much more efficient than
5029 a normal memory BIO. This was added because there are several times when
5030 an area of memory needs to be read from a BIO. The previous method was
5031 to create a memory BIO and write the data to it, this results in two
5032 copies of the data and an O(n^2) reading algorithm. There is a new
5033 function BIO_new_mem_buf() which creates a read only memory BIO from
5034 an area of memory. Also modified the PKCS#7 routines to use read only
5038 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5039 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5040 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5041 but a retry condition occured while trying to read the rest.
5044 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5045 NID_pkcs7_encrypted by default: this was wrong since this should almost
5046 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5047 the encrypted data type: this is a more sensible place to put it and it
5048 allows the PKCS#12 code to be tidied up that duplicated this
5052 *) Changed obj_dat.pl script so it takes its input and output files on
5053 the command line. This should avoid shell escape redirection problems
5057 *) Initial support for certificate extension requests, these are included
5058 in things like Xenroll certificate requests. Included functions to allow
5059 extensions to be obtained and added.
5062 *) -crlf option to s_client and s_server for sending newlines as
5063 CRLF (as required by many protocols).
5066 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5068 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5069 [Ralf S. Engelschall]
5071 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5072 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5074 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5078 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5079 DH parameters/keys (q is lost during that conversion, but the resulting
5080 DH parameters contain its length).
5082 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5083 much faster than DH_generate_parameters (which creates parameters
5084 where p = 2*q + 1), and also the smaller q makes DH computations
5085 much more efficient (160-bit exponentiation instead of 1024-bit
5086 exponentiation); so this provides a convenient way to support DHE
5087 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5088 utter importance to use
5089 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5091 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5092 when such DH parameters are used, because otherwise small subgroup
5093 attacks may become possible!
5096 *) Avoid memory leak in i2d_DHparams.
5099 *) Allow the -k option to be used more than once in the enc program:
5100 this allows the same encrypted message to be read by multiple recipients.
5103 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5104 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5105 it will always use the numerical form of the OID, even if it has a short
5109 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5110 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5111 otherwise bn_mod_exp was called. In the case of hardware keys for example
5112 no private key components need be present and it might store extra data
5113 in the RSA structure, which cannot be accessed from bn_mod_exp.
5114 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5115 private key operations.
5118 *) Added support for SPARC Linux.
5121 *) pem_password_cb function type incompatibly changed from
5122 typedef int pem_password_cb(char *buf, int size, int rwflag);
5124 ....(char *buf, int size, int rwflag, void *userdata);
5125 so that applications can pass data to their callbacks:
5126 The PEM[_ASN1]_{read,write}... functions and macros now take an
5127 additional void * argument, which is just handed through whenever
5128 the password callback is called.
5129 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5131 New function SSL_CTX_set_default_passwd_cb_userdata.
5133 Compatibility note: As many C implementations push function arguments
5134 onto the stack in reverse order, the new library version is likely to
5135 interoperate with programs that have been compiled with the old
5136 pem_password_cb definition (PEM_whatever takes some data that
5137 happens to be on the stack as its last argument, and the callback
5138 just ignores this garbage); but there is no guarantee whatsoever that
5141 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5142 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5143 problems not only on Windows, but also on some Unix platforms.
5144 To avoid problematic command lines, these definitions are now in an
5145 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5146 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5149 *) MIPS III/IV assembler module is reimplemented.
5152 *) More DES library cleanups: remove references to srand/rand and
5153 delete an unused file.
5156 *) Add support for the the free Netwide assembler (NASM) under Win32,
5157 since not many people have MASM (ml) and it can be hard to obtain.
5158 This is currently experimental but it seems to work OK and pass all
5159 the tests. Check out INSTALL.W32 for info.
5162 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5163 without temporary keys kept an extra copy of the server key,
5164 and connections with temporary keys did not free everything in case
5168 *) New function RSA_check_key and new openssl rsa option -check
5169 for verifying the consistency of RSA keys.
5170 [Ulf Moeller, Bodo Moeller]
5172 *) Various changes to make Win32 compile work:
5173 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5174 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5175 comparison" warnings.
5176 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5179 *) Add a debugging option to PKCS#5 v2 key generation function: when
5180 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5181 derived keys are printed to stderr.
5184 *) Copy the flags in ASN1_STRING_dup().
5185 [Roman E. Pavlov <pre@mo.msk.ru>]
5187 *) The x509 application mishandled signing requests containing DSA
5188 keys when the signing key was also DSA and the parameters didn't match.
5190 It was supposed to omit the parameters when they matched the signing key:
5191 the verifying software was then supposed to automatically use the CA's
5192 parameters if they were absent from the end user certificate.
5194 Omitting parameters is no longer recommended. The test was also
5195 the wrong way round! This was probably due to unusual behaviour in
5196 EVP_cmp_parameters() which returns 1 if the parameters match.
5197 This meant that parameters were omitted when they *didn't* match and
5198 the certificate was useless. Certificates signed with 'ca' didn't have
5200 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5202 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5203 The interface is as follows:
5204 Applications can use
5205 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5206 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5207 "off" is now the default.
5208 The library internally uses
5209 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5210 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5211 to disable memory-checking temporarily.
5213 Some inconsistent states that previously were possible (and were
5214 even the default) are now avoided.
5216 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5217 with each memory chunk allocated; this is occasionally more helpful
5218 than just having a counter.
5220 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5222 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5226 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5227 which largely parallels "options", but is for changing API behaviour,
5228 whereas "options" are about protocol behaviour.
5229 Initial "mode" flags are:
5231 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5232 a single record has been written.
5233 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5234 retries use the same buffer location.
5235 (But all of the contents must be
5239 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5242 *) Fix problems with no-hmac etc.
5243 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5245 *) New functions RSA_get_default_method(), RSA_set_method() and
5246 RSA_get_method(). These allows replacement of RSA_METHODs without having
5247 to mess around with the internals of an RSA structure.
5250 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5251 Also really enable memory leak checks in openssl.c and in some
5253 [Chad C. Mulligan, Bodo Moeller]
5255 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5256 up the length of negative integers. This has now been simplified to just
5257 store the length when it is first determined and use it later, rather
5258 than trying to keep track of where data is copied and updating it to
5260 [Steve Henson, reported by Brien Wheeler
5261 <bwheeler@authentica-security.com>]
5263 *) Add a new function PKCS7_signatureVerify. This allows the verification
5264 of a PKCS#7 signature but with the signing certificate passed to the
5265 function itself. This contrasts with PKCS7_dataVerify which assumes the
5266 certificate is present in the PKCS#7 structure. This isn't always the
5267 case: certificates can be omitted from a PKCS#7 structure and be
5268 distributed by "out of band" means (such as a certificate database).
5271 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5272 function prototypes in pem.h, also change util/mkdef.pl to add the
5273 necessary function names.
5276 *) mk1mf.pl (used by Windows builds) did not properly read the
5277 options set by Configure in the top level Makefile, and Configure
5278 was not even able to write more than one option correctly.
5279 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5282 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5283 file to be loaded from a BIO or FILE pointer. The BIO version will
5284 for example allow memory BIOs to contain config info.
5287 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5288 Whoever hopes to achieve shared-library compatibility across versions
5289 must use this, not the compile-time macro.
5290 (Exercise 0.9.4: Which is the minimum library version required by
5292 Note: All this applies only to multi-threaded programs, others don't
5296 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5297 through a BIO pair triggered the default case, i.e.
5298 SSLerr(...,SSL_R_UNKNOWN_STATE).
5301 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5302 can use the SSL library even if none of the specific BIOs is
5306 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5307 for the encoded length.
5308 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5310 *) Add initial documentation of the X509V3 functions.
5313 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5314 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5315 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5316 secure PKCS#8 private key format with a high iteration count.
5319 *) Fix determination of Perl interpreter: A perl or perl5
5320 _directory_ in $PATH was also accepted as the interpreter.
5321 [Ralf S. Engelschall]
5323 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5324 wrong with it but it was very old and did things like calling
5325 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5329 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5330 to use the new extension code.
5333 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5334 with macros. This should make it easier to change their form, add extra
5335 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5339 *) Add to configuration table a new entry that can specify an alternative
5340 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5341 according to Mark Crispin <MRC@Panda.COM>.
5345 *) DES CBC did not update the IV. Weird.
5348 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5349 Changing the behaviour of the former might break existing programs --
5350 where IV updating is needed, des_ncbc_encrypt can be used.
5353 *) When bntest is run from "make test" it drives bc to check its
5354 calculations, as well as internally checking them. If an internal check
5355 fails, it needs to cause bc to give a non-zero result or make test carries
5356 on without noticing the failure. Fixed.
5359 *) DES library cleanups.
5362 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5363 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5364 ciphers. NOTE: although the key derivation function has been verified
5365 against some published test vectors it has not been extensively tested
5366 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5370 *) Instead of "mkdir -p", which is not fully portable, use new
5371 Perl script "util/mkdir-p.pl".
5374 *) Rewrite the way password based encryption (PBE) is handled. It used to
5375 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5376 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5377 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5378 the 'parameter' field of the AlgorithmIdentifier is passed to the
5379 underlying key generation function so it must do its own ASN1 parsing.
5380 This has also changed the EVP_PBE_CipherInit() function which now has a
5381 'parameter' argument instead of literal salt and iteration count values
5382 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5385 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5386 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5387 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5388 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5389 value was just used as a "magic string" and not used directly its
5390 value doesn't matter.
5393 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5397 *) "linux-sparc64" configuration (ultrapenguin).
5398 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5399 "linux-sparc" configuration.
5400 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5402 *) config now generates no-xxx options for missing ciphers.
5405 *) Support the EBCDIC character set (work in progress).
5406 File ebcdic.c not yet included because it has a different license.
5407 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5409 *) Support BS2000/OSD-POSIX.
5410 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5412 *) Make callbacks for key generation use void * instead of char *.
5415 *) Make S/MIME samples compile (not yet tested).
5418 *) Additional typesafe stacks.
5421 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5425 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5427 *) New configuration variant "sco5-gcc".
5429 *) Updated some demos.
5430 [Sean O Riordain, Wade Scholine]
5432 *) Add missing BIO_free at exit of pkcs12 application.
5435 *) Fix memory leak in conf.c.
5438 *) Updates for Win32 to assembler version of MD5.
5441 *) Set #! path to perl in apps/der_chop to where we found it
5442 instead of using a fixed path.
5445 *) SHA library changes for irix64-mips4-cc.
5448 *) Improvements for VMS support.
5452 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5454 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5455 This also avoids the problems with SC4.2 and unpatched SC5.
5456 [Andy Polyakov <appro@fy.chalmers.se>]
5458 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5459 These are required because of the typesafe stack would otherwise break
5460 existing code. If old code used a structure member which used to be STACK
5461 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5462 sk_num or sk_value it would produce an error because the num, data members
5463 are not present in STACK_OF. Now it just produces a warning. sk_set
5464 replaces the old method of assigning a value to sk_value
5465 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5466 that does this will no longer work (and should use sk_set instead) but
5467 this could be regarded as a "questionable" behaviour anyway.
5470 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5471 correctly handle encrypted S/MIME data.
5474 *) Change type of various DES function arguments from des_cblock
5475 (which means, in function argument declarations, pointer to char)
5476 to des_cblock * (meaning pointer to array with 8 char elements),
5477 which allows the compiler to do more typechecking; it was like
5478 that back in SSLeay, but with lots of ugly casts.
5480 Introduce new type const_des_cblock.
5483 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5484 problems: find RecipientInfo structure that matches recipient certificate
5485 and initialise the ASN1 structures properly based on passed cipher.
5488 *) Belatedly make the BN tests actually check the results.
5491 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5492 to and from BNs: it was completely broken. New compilation option
5493 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5494 key elements as negative integers.
5497 *) Reorganize and speed up MD5.
5498 [Andy Polyakov <appro@fy.chalmers.se>]
5501 [Richard Levitte <richard@levitte.org>]
5503 *) New option -out to asn1parse to allow the parsed structure to be
5504 output to a file. This is most useful when combined with the -strparse
5505 option to examine the output of things like OCTET STRINGS.
5508 *) Make SSL library a little more fool-proof by not requiring any longer
5509 that SSL_set_{accept,connect}_state be called before
5510 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5511 in many applications because usually everything *appeared* to work as
5512 intended anyway -- now it really works as intended).
5515 *) Move openssl.cnf out of lib/.
5518 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5519 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5520 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5521 [Ralf S. Engelschall]
5523 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5524 handle PKCS#7 enveloped data properly.
5525 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5527 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5528 copying pointers. The cert_st handling is changed by this in
5529 various ways (and thus what used to be known as ctx->default_cert
5530 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5531 any longer when s->cert does not give us what we need).
5532 ssl_cert_instantiate becomes obsolete by this change.
5533 As soon as we've got the new code right (possibly it already is?),
5534 we have solved a couple of bugs of the earlier code where s->cert
5535 was used as if it could not have been shared with other SSL structures.
5537 Note that using the SSL API in certain dirty ways now will result
5538 in different behaviour than observed with earlier library versions:
5539 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5540 does not influence s as it used to.
5542 In order to clean up things more thoroughly, inside SSL_SESSION
5543 we don't use CERT any longer, but a new structure SESS_CERT
5544 that holds per-session data (if available); currently, this is
5545 the peer's certificate chain and, for clients, the server's certificate
5546 and temporary key. CERT holds only those values that can have
5547 meaningful defaults in an SSL_CTX.
5550 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5551 from the internal representation. Various PKCS#7 fixes: remove some
5552 evil casts and set the enc_dig_alg field properly based on the signing
5556 *) Allow PKCS#12 password to be set from the command line or the
5557 environment. Let 'ca' get its config file name from the environment
5558 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5562 *) Allow certificate policies extension to use an IA5STRING for the
5563 organization field. This is contrary to the PKIX definition but
5564 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5568 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5569 without disallowing inline assembler and the like for non-pedantic builds.
5572 *) Support Borland C++ builder.
5573 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5578 *) SHA-1 cleanups and performance enhancements.
5579 [Andy Polyakov <appro@fy.chalmers.se>]
5581 *) Sparc v8plus assembler for the bignum library.
5582 [Andy Polyakov <appro@fy.chalmers.se>]
5584 *) Accept any -xxx and +xxx compiler options in Configure.
5587 *) Update HPUX configuration.
5590 *) Add missing sk_<type>_unshift() function to safestack.h
5591 [Ralf S. Engelschall]
5593 *) New function SSL_CTX_use_certificate_chain_file that sets the
5594 "extra_cert"s in addition to the certificate. (This makes sense
5595 only for "PEM" format files, as chains as a whole are not
5599 *) Support verify_depth from the SSL API.
5600 x509_vfy.c had what can be considered an off-by-one-error:
5601 Its depth (which was not part of the external interface)
5602 was actually counting the number of certificates in a chain;
5603 now it really counts the depth.
5606 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5607 instead of X509err, which often resulted in confusing error
5608 messages since the error codes are not globally unique
5609 (e.g. an alleged error in ssl3_accept when a certificate
5610 didn't match the private key).
5612 *) New function SSL_CTX_set_session_id_context that allows to set a default
5613 value (so that you don't need SSL_set_session_id_context for each
5614 connection using the SSL_CTX).
5617 *) OAEP decoding bug fix.
5620 *) Support INSTALL_PREFIX for package builders, as proposed by
5624 *) New Configure options "threads" and "no-threads". For systems
5625 where the proper compiler options are known (currently Solaris
5626 and Linux), "threads" is the default.
5629 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5632 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5633 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5634 such as /usr/local/bin.
5637 *) "make linux-shared" to build shared libraries.
5638 [Niels Poppe <niels@netbox.org>]
5640 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5643 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5644 extension adding in x509 utility.
5647 *) Remove NOPROTO sections and error code comments.
5650 *) Partial rewrite of the DEF file generator to now parse the ANSI
5654 *) New Configure options --prefix=DIR and --openssldir=DIR.
5657 *) Complete rewrite of the error code script(s). It is all now handled
5658 by one script at the top level which handles error code gathering,
5659 header rewriting and C source file generation. It should be much better
5660 than the old method: it now uses a modified version of Ulf's parser to
5661 read the ANSI prototypes in all header files (thus the old K&R definitions
5662 aren't needed for error creation any more) and do a better job of
5663 translating function codes into names. The old 'ASN1 error code imbedded
5664 in a comment' is no longer necessary and it doesn't use .err files which
5665 have now been deleted. Also the error code call doesn't have to appear all
5666 on one line (which resulted in some large lines...).
5669 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5672 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5673 0 (which usually indicates a closed connection), but continue reading.
5676 *) Fix some race conditions.
5679 *) Add support for CRL distribution points extension. Add Certificate
5680 Policies and CRL distribution points documentation.
5683 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5686 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5687 8 of keying material. Merlin has also confirmed interop with this fix
5688 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5689 [Merlin Hughes <merlin@baltimore.ie>]
5691 *) Fix lots of warnings.
5692 [Richard Levitte <levitte@stacken.kth.se>]
5694 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5695 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5696 [Richard Levitte <levitte@stacken.kth.se>]
5698 *) Fix problems with sizeof(long) == 8.
5699 [Andy Polyakov <appro@fy.chalmers.se>]
5701 *) Change functions to ANSI C.
5704 *) Fix typos in error codes.
5705 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5707 *) Remove defunct assembler files from Configure.
5710 *) SPARC v8 assembler BIGNUM implementation.
5711 [Andy Polyakov <appro@fy.chalmers.se>]
5713 *) Support for Certificate Policies extension: both print and set.
5714 Various additions to support the r2i method this uses.
5717 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5718 return a const string when you are expecting an allocated buffer.
5721 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5722 types DirectoryString and DisplayText.
5725 *) Add code to allow r2i extensions to access the configuration database,
5726 add an LHASH database driver and add several ctx helper functions.
5729 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5730 fail when they extended the size of a BIGNUM.
5733 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5734 support typesafe stack.
5737 *) Fix typo in SSL_[gs]et_options().
5738 [Nils Frostberg <nils@medcom.se>]
5740 *) Delete various functions and files that belonged to the (now obsolete)
5741 old X509V3 handling code.
5744 *) New Configure option "rsaref".
5747 *) Don't auto-generate pem.h.
5750 *) Introduce type-safe ASN.1 SETs.
5753 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5754 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5756 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5757 that links with OpenSSL (well at least cause lots of warnings), but fear
5758 not: the conversion is trivial, and it eliminates loads of evil casts. A
5759 few STACKed things have been converted already. Feel free to convert more.
5760 In the fullness of time, I'll do away with the STACK type altogether.
5763 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5764 specified in <certfile> by updating the entry in the index.txt file.
5765 This way one no longer has to edit the index.txt file manually for
5766 revoking a certificate. The -revoke option does the gory details now.
5767 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5769 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5770 `-text' option at all and this way the `-noout -text' combination was
5771 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5772 [Ralf S. Engelschall]
5774 *) Make sure a corresponding plain text error message exists for the
5775 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5776 verify callback function determined that a certificate was revoked.
5777 [Ralf S. Engelschall]
5779 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5780 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5781 all available cipers including rc5, which was forgotten until now.
5782 In order to let the testing shell script know which algorithms
5783 are available, a new (up to now undocumented) command
5784 "openssl list-cipher-commands" is used.
5787 *) Bugfix: s_client occasionally would sleep in select() when
5788 it should have checked SSL_pending() first.
5791 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5792 the raw DSA values prior to ASN.1 encoding.
5795 *) Tweaks to Configure
5796 [Niels Poppe <niels@netbox.org>]
5798 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5802 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5805 *) New config option to avoid instructions that are illegal on the 80386.
5806 The default code is faster, but requires at least a 486.
5809 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5810 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5811 same as SSL2_VERSION anyway.
5814 *) New "-showcerts" option for s_client.
5817 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5818 application. Various cleanups and fixes.
5821 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5822 modify error routines to work internally. Add error codes and PBE init
5823 to library startup routines.
5826 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5827 packing functions to asn1 and evp. Changed function names and error
5828 codes along the way.
5831 *) PKCS12 integration: and so it begins... First of several patches to
5832 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5833 objects to objects.h
5836 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5837 and display support for Thawte strong extranet extension.
5840 *) Add LinuxPPC support.
5841 [Jeff Dubrule <igor@pobox.org>]
5843 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5844 bn_div_words in alpha.s.
5845 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5847 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5848 OAEP isn't supported when OpenSSL is built with RSAref.
5849 [Ulf Moeller <ulf@fitug.de>]
5851 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5852 so they no longer are missing under -DNOPROTO.
5853 [Soren S. Jorvang <soren@t.dk>]
5856 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5858 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5859 doesn't work when the session is reused. Coming soon!
5862 *) Fix a security hole, that allows sessions to be reused in the wrong
5863 context thus bypassing client cert protection! All software that uses
5864 client certs and session caches in multiple contexts NEEDS PATCHING to
5865 allow session reuse! A fuller solution is in the works.
5866 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5868 *) Some more source tree cleanups (removed obsolete files
5869 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5870 permission on "config" script to be executable) and a fix for the INSTALL
5872 [Ulf Moeller <ulf@fitug.de>]
5874 *) Remove some legacy and erroneous uses of malloc, free instead of
5876 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5878 *) Make rsa_oaep_test return non-zero on error.
5879 [Ulf Moeller <ulf@fitug.de>]
5881 *) Add support for native Solaris shared libraries. Configure
5882 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5883 if someone would make that last step automatic.
5884 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5886 *) ctx_size was not built with the right compiler during "make links". Fixed.
5889 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5890 except NULL ciphers". This means the default cipher list will no longer
5891 enable NULL ciphers. They need to be specifically enabled e.g. with
5892 the string "DEFAULT:eNULL".
5895 *) Fix to RSA private encryption routines: if p < q then it would
5896 occasionally produce an invalid result. This will only happen with
5897 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5900 *) Be less restrictive and allow also `perl util/perlpath.pl
5901 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5902 because this way one can also use an interpreter named `perl5' (which is
5903 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5904 installed as `perl').
5905 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5907 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5908 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5910 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5911 advapi32.lib to Win32 build and change the pem test comparision
5912 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5913 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5914 and crypto/des/ede_cbcm_enc.c.
5917 *) DES quad checksum was broken on big-endian architectures. Fixed.
5920 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5921 Win32 test batch file so it (might) work again. The Win32 test batch file
5922 is horrible: I feel ill....
5925 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5926 in e_os.h. Audit of header files to check ANSI and non ANSI
5927 sections: 10 functions were absent from non ANSI section and not exported
5928 from Windows DLLs. Fixed up libeay.num for new functions.
5931 *) Make `openssl version' output lines consistent.
5932 [Ralf S. Engelschall]
5934 *) Fix Win32 symbol export lists for BIO functions: Added
5935 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5936 to ms/libeay{16,32}.def.
5937 [Ralf S. Engelschall]
5939 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5940 fine under Unix and passes some trivial tests I've now added. But the
5941 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5942 added to make sure no one expects that this stuff really works in the
5943 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5944 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5946 [Ralf S. Engelschall]
5948 *) Fix the generation of two part addresses in perl.
5949 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5951 *) Add config entry for Linux on MIPS.
5952 [John Tobey <jtobey@channel1.com>]
5954 *) Make links whenever Configure is run, unless we are on Windoze.
5957 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5958 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5962 *) Add a useful kludge to allow package maintainers to specify compiler and
5963 other platforms details on the command line without having to patch the
5964 Configure script everytime: One now can use ``perl Configure
5965 <id>:<details>'', i.e. platform ids are allowed to have details appended
5966 to them (seperated by colons). This is treated as there would be a static
5967 pre-configured entry in Configure's %table under key <id> with value
5968 <details> and ``perl Configure <id>'' is called. So, when you want to
5969 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5970 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5971 now, which overrides the FreeBSD-elf entry on-the-fly.
5972 [Ralf S. Engelschall]
5974 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5977 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5978 on the `perl Configure ...' command line. This way one can compile
5979 OpenSSL libraries with Position Independent Code (PIC) which is needed
5980 for linking it into DSOs.
5981 [Ralf S. Engelschall]
5983 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5987 *) Cleaned up the LICENSE document: The official contact for any license
5988 questions now is the OpenSSL core team under openssl-core@openssl.org.
5989 And add a paragraph about the dual-license situation to make sure people
5990 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5991 to the OpenSSL toolkit.
5992 [Ralf S. Engelschall]
5994 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5995 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5996 Additonally cleaned up the `make links' target: Remove unnecessary
5997 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5998 to speed processing and no longer clutter the display with confusing
5999 stuff. Instead only the actually done links are displayed.
6000 [Ralf S. Engelschall]
6002 *) Permit null encryption ciphersuites, used for authentication only. It used
6003 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6004 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6008 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6009 signed attributes when verifying signatures (this would break them),
6010 the detached data encoding was wrong and public keys obtained using
6011 X509_get_pubkey() weren't freed.
6014 *) Add text documentation for the BUFFER functions. Also added a work around
6015 to a Win95 console bug. This was triggered by the password read stuff: the
6016 last character typed gets carried over to the next fread(). If you were
6017 generating a new cert request using 'req' for example then the last
6018 character of the passphrase would be CR which would then enter the first
6022 *) Added the new `Includes OpenSSL Cryptography Software' button as
6023 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6024 button and can be used by applications based on OpenSSL to show the
6025 relationship to the OpenSSL project.
6026 [Ralf S. Engelschall]
6028 *) Remove confusing variables in function signatures in files
6029 ssl/ssl_lib.c and ssl/ssl.h.
6030 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6032 *) Don't install bss_file.c under PREFIX/include/
6033 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6035 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6036 functions that return function pointers and has support for NT specific
6037 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6038 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6039 unsigned to signed types: this was killing the Win32 compile.
6042 *) Add new certificate file to stack functions,
6043 SSL_add_dir_cert_subjects_to_stack() and
6044 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6045 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6046 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6047 This means that Apache-SSL and similar packages don't have to mess around
6048 to add as many CAs as they want to the preferred list.
6051 *) Experiment with doxygen documentation. Currently only partially applied to
6053 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6054 openssl.doxy as the configuration file.
6057 *) Get rid of remaining C++-style comments which strict C compilers hate.
6058 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6060 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6061 compiled in by default: it has problems with large keys.
6064 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6065 DH private keys and/or callback functions which directly correspond to
6066 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6067 is needed for applications which have to configure certificates on a
6068 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6070 For the RSA certificate situation is makes no difference, but
6071 for the DSA certificate situation this fixes the "no shared cipher"
6072 problem where the OpenSSL cipher selection procedure failed because the
6073 temporary keys were not overtaken from the context and the API provided
6074 no way to reconfigure them.
6075 The new functions now let applications reconfigure the stuff and they
6076 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6077 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6078 non-public-API function ssl_cert_instantiate() is used as a helper
6079 function and also to reduce code redundancy inside ssl_rsa.c.
6080 [Ralf S. Engelschall]
6082 *) Move s_server -dcert and -dkey options out of the undocumented feature
6083 area because they are useful for the DSA situation and should be
6084 recognized by the users.
6085 [Ralf S. Engelschall]
6087 *) Fix the cipher decision scheme for export ciphers: the export bits are
6088 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6089 SSL_EXP_MASK. So, the original variable has to be used instead of the
6090 already masked variable.
6091 [Richard Levitte <levitte@stacken.kth.se>]
6093 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6094 [Richard Levitte <levitte@stacken.kth.se>]
6096 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6097 from `int' to `unsigned int' because it's a length and initialized by
6098 EVP_DigestFinal() which expects an `unsigned int *'.
6099 [Richard Levitte <levitte@stacken.kth.se>]
6101 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6102 script. Instead use the usual Shell->Perl transition trick.
6103 [Ralf S. Engelschall]
6105 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6106 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6107 -noout -modulus' as it's already the case for `openssl rsa -noout
6108 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6109 currently the public key is printed (a decision which was already done by
6110 `openssl dsa -modulus' in the past) which serves a similar purpose.
6111 Additionally the NO_RSA no longer completely removes the whole -modulus
6112 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6114 [Ralf S. Engelschall]
6116 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6117 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6118 [Arne Ansper <arne@ats.cyber.ee>]
6120 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6121 to be added. Now both 'req' and 'ca' can use new objects defined in the
6125 *) Add cool BIO that does syslog (or event log on NT).
6126 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6128 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6129 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6130 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6131 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6134 *) Add preliminary config info for new extension code.
6137 *) Make RSA_NO_PADDING really use no padding.
6138 [Ulf Moeller <ulf@fitug.de>]
6140 *) Generate errors when private/public key check is done.
6143 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6144 for some CRL extensions and new objects added.
6147 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6148 key usage extension and fuller support for authority key id.
6151 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6152 padding method for RSA, which is recommended for new applications in PKCS
6153 #1 v2.0 (RFC 2437, October 1998).
6154 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6155 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6156 against Bleichbacher's attack on RSA.
6157 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6160 *) Updates to the new SSL compression code
6161 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6163 *) Fix so that the version number in the master secret, when passed
6164 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6165 (because the server will not accept higher), that the version number
6166 is 0x03,0x01, not 0x03,0x00
6167 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6169 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6170 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6171 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6174 *) Support for RAW extensions where an arbitrary extension can be
6175 created by including its DER encoding. See apps/openssl.cnf for
6179 *) Make sure latest Perl versions don't interpret some generated C array
6180 code as Perl array code in the crypto/err/err_genc.pl script.
6181 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6183 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6184 not many people have the assembler. Various Win32 compilation fixes and
6185 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6189 *) Modify configure script 'Configure' to automatically create crypto/date.h
6190 file under Win32 and also build pem.h from pem.org. New script
6191 util/mkfiles.pl to create the MINFO file on environments that can't do a
6192 'make files': perl util/mkfiles.pl >MINFO should work.
6195 *) Major rework of DES function declarations, in the pursuit of correctness
6196 and purity. As a result, many evil casts evaporated, and some weirdness,
6197 too. You may find this causes warnings in your code. Zapping your evil
6198 casts will probably fix them. Mostly.
6201 *) Fix for a typo in asn1.h. Bug fix to object creation script
6202 obj_dat.pl. It considered a zero in an object definition to mean
6203 "end of object": none of the objects in objects.h have any zeros
6204 so it wasn't spotted.
6205 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6207 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6208 Masking (CBCM). In the absence of test vectors, the best I have been able
6209 to do is check that the decrypt undoes the encrypt, so far. Send me test
6210 vectors if you have them.
6213 *) Correct calculation of key length for export ciphers (too much space was
6214 allocated for null ciphers). This has not been tested!
6217 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6218 message is now correct (it understands "crypto" and "ssl" on its
6219 command line). There is also now an "update" option. This will update
6220 the util/ssleay.num and util/libeay.num files with any new functions.
6222 perl util/mkdef.pl crypto ssl update
6223 it will update them.
6226 *) Overhauled the Perl interface (perl/*):
6227 - ported BN stuff to OpenSSL's different BN library
6228 - made the perl/ source tree CVS-aware
6229 - renamed the package from SSLeay to OpenSSL (the files still contain
6230 their history because I've copied them in the repository)
6231 - removed obsolete files (the test scripts will be replaced
6232 by better Test::Harness variants in the future)
6233 [Ralf S. Engelschall]
6235 *) First cut for a very conservative source tree cleanup:
6236 1. merge various obsolete readme texts into doc/ssleay.txt
6237 where we collect the old documents and readme texts.
6238 2. remove the first part of files where I'm already sure that we no
6239 longer need them because of three reasons: either they are just temporary
6240 files which were left by Eric or they are preserved original files where
6241 I've verified that the diff is also available in the CVS via "cvs diff
6242 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6243 the crypto/md/ stuff).
6244 [Ralf S. Engelschall]
6246 *) More extension code. Incomplete support for subject and issuer alt
6247 name, issuer and authority key id. Change the i2v function parameters
6248 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6249 what that's for :-) Fix to ASN1 macro which messed up
6250 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6253 *) Preliminary support for ENUMERATED type. This is largely copied from the
6257 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6258 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6260 *) Make sure `make rehash' target really finds the `openssl' program.
6261 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6263 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6264 like to hear about it if this slows down other processors.
6267 *) Add CygWin32 platform information to Configure script.
6268 [Alan Batie <batie@aahz.jf.intel.com>]
6270 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6271 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6273 *) New program nseq to manipulate netscape certificate sequences
6276 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6280 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6281 but the BN code had some problems that would cause failures when
6282 doing certificate verification and some other functions.
6283 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6285 *) Add ASN1 and PEM code to support netscape certificate sequences.
6288 *) Add ASN1 and PEM code to support netscape certificate sequences.
6291 *) Add several PKIX and private extended key usage OIDs.
6294 *) Modify the 'ca' program to handle the new extension code. Modify
6295 openssl.cnf for new extension format, add comments.
6298 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6299 and add a sample to openssl.cnf so req -x509 now adds appropriate
6303 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6304 error code, add initial support to X509_print() and x509 application.
6307 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6308 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6309 stuff is currently isolated and isn't even compiled yet.
6312 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6313 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6314 Removed the versions check from X509 routines when loading extensions:
6315 this allows certain broken certificates that don't set the version
6316 properly to be processed.
6319 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6320 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6321 can still be regenerated with "make depend".
6324 *) Spelling mistake in C version of CAST-128.
6325 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6327 *) Changes to the error generation code. The perl script err-code.pl
6328 now reads in the old error codes and retains the old numbers, only
6329 adding new ones if necessary. It also only changes the .err files if new
6330 codes are added. The makefiles have been modified to only insert errors
6331 when needed (to avoid needlessly modifying header files). This is done
6332 by only inserting errors if the .err file is newer than the auto generated
6333 C file. To rebuild all the error codes from scratch (the old behaviour)
6334 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6335 or delete all the .err files.
6338 *) CAST-128 was incorrectly implemented for short keys. The C version has
6339 been fixed, but is untested. The assembler versions are also fixed, but
6340 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6341 to regenerate it if needed.
6342 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6343 Hagino <itojun@kame.net>]
6345 *) File was opened incorrectly in randfile.c.
6346 [Ulf Möller <ulf@fitug.de>]
6348 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6349 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6350 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6351 al: it's just almost always a UTCTime. Note this patch adds new error
6352 codes so do a "make errors" if there are problems.
6355 *) Correct Linux 1 recognition in config.
6356 [Ulf Möller <ulf@fitug.de>]
6358 *) Remove pointless MD5 hash when using DSA keys in ca.
6359 [Anonymous <nobody@replay.com>]
6361 *) Generate an error if given an empty string as a cert directory. Also
6362 generate an error if handed NULL (previously returned 0 to indicate an
6363 error, but didn't set one).
6364 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6366 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6369 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6370 parameters. This was causing a warning which killed off the Win32 compile.
6373 *) Remove C++ style comments from crypto/bn/bn_local.h.
6374 [Neil Costigan <neil.costigan@celocom.com>]
6376 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6377 based on a text string, looking up short and long names and finally
6378 "dot" format. The "dot" format stuff didn't work. Added new function
6379 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6380 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6381 OID is not part of the table.
6384 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6385 X509_LOOKUP_by_alias().
6388 *) Sort openssl functions by name.
6391 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6392 encryption from sample DSA keys (in case anyone is interested the password
6396 *) Make _all_ *_free functions accept a NULL pointer.
6397 [Frans Heymans <fheymans@isaserver.be>]
6399 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6401 [Anonymous <nobody@replay.com>]
6403 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6404 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6406 *) Don't blow it for numeric -newkey arguments to apps/req.
6407 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6409 *) Temp key "for export" tests were wrong in s3_srvr.c.
6410 [Anonymous <nobody@replay.com>]
6412 *) Add prototype for temp key callback functions
6413 SSL_CTX_set_tmp_{rsa,dh}_callback().
6416 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6417 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6420 *) X509_name_add_entry() freed the wrong thing after an error.
6421 [Arne Ansper <arne@ats.cyber.ee>]
6423 *) rsa_eay.c would attempt to free a NULL context.
6424 [Arne Ansper <arne@ats.cyber.ee>]
6426 *) BIO_s_socket() had a broken should_retry() on Windoze.
6427 [Arne Ansper <arne@ats.cyber.ee>]
6429 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6430 [Arne Ansper <arne@ats.cyber.ee>]
6432 *) Make sure the already existing X509_STORE->depth variable is initialized
6433 in X509_STORE_new(), but document the fact that this variable is still
6434 unused in the certificate verification process.
6435 [Ralf S. Engelschall]
6437 *) Fix the various library and apps files to free up pkeys obtained from
6438 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6441 *) Fix reference counting in X509_PUBKEY_get(). This makes
6442 demos/maurice/example2.c work, amongst others, probably.
6443 [Steve Henson and Ben Laurie]
6445 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6446 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6447 are no longer created. This way we have a single and consistent command
6448 line interface `openssl <command>', similar to `cvs <command>'.
6449 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6451 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6452 BIT STRING wrapper always have zero unused bits.
6455 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6458 *) Make the top-level INSTALL documentation easier to understand.
6461 *) Makefiles updated to exit if an error occurs in a sub-directory
6462 make (including if user presses ^C) [Paul Sutton]
6464 *) Make Montgomery context stuff explicit in RSA data structure.
6467 *) Fix build order of pem and err to allow for generated pem.h.
6470 *) Fix renumbering bug in X509_NAME_delete_entry().
6473 *) Enhanced the err-ins.pl script so it makes the error library number
6474 global and can add a library name. This is needed for external ASN1 and
6475 other error libraries.
6478 *) Fixed sk_insert which never worked properly.
6481 *) Fix ASN1 macros so they can handle indefinite length construted
6482 EXPLICIT tags. Some non standard certificates use these: they can now
6486 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6487 into a single doc/ssleay.txt bundle. This way the information is still
6488 preserved but no longer messes up this directory. Now it's new room for
6489 the new set of documenation files.
6490 [Ralf S. Engelschall]
6492 *) SETs were incorrectly DER encoded. This was a major pain, because they
6493 shared code with SEQUENCEs, which aren't coded the same. This means that
6494 almost everything to do with SETs or SEQUENCEs has either changed name or
6495 number of arguments.
6496 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6498 *) Fix test data to work with the above.
6501 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6502 was already fixed by Eric for 0.9.1 it seems.
6503 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6505 *) Autodetect FreeBSD3.
6508 *) Fix various bugs in Configure. This affects the following platforms:
6512 unixware-2.0-pentium
6516 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6517 before they are needed.
6520 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6524 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6526 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6527 changed SSLeay to OpenSSL in version strings.
6528 [Ralf S. Engelschall]
6530 *) Some fixups to the top-level documents.
6533 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6534 because the symlink to include/ was missing.
6535 [Ralf S. Engelschall]
6537 *) Incorporated the popular no-RSA/DSA-only patches
6538 which allow to compile a RSA-free SSLeay.
6539 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6541 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6542 when "ssleay" is still not found.
6543 [Ralf S. Engelschall]
6545 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6546 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6548 *) Updated the README file.
6549 [Ralf S. Engelschall]
6551 *) Added various .cvsignore files in the CVS repository subdirs
6552 to make a "cvs update" really silent.
6553 [Ralf S. Engelschall]
6555 *) Recompiled the error-definition header files and added
6556 missing symbols to the Win32 linker tables.
6557 [Ralf S. Engelschall]
6559 *) Cleaned up the top-level documents;
6560 o new files: CHANGES and LICENSE
6561 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6562 o merged COPYRIGHT into LICENSE
6563 o removed obsolete TODO file
6564 o renamed MICROSOFT to INSTALL.W32
6565 [Ralf S. Engelschall]
6567 *) Removed dummy files from the 0.9.1b source tree:
6568 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6569 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6570 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6571 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6572 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6573 [Ralf S. Engelschall]
6575 *) Added various platform portability fixes.
6578 *) The Genesis of the OpenSSL rpject:
6579 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6580 Young and Tim J. Hudson created while they were working for C2Net until
6582 [The OpenSSL Project]
6585 Changes between 0.9.0b and 0.9.1b [not released]
6587 *) Updated a few CA certificates under certs/
6590 *) Changed some BIGNUM api stuff.
6593 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6594 DGUX x86, Linux Alpha, etc.
6597 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6598 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6602 *) Add -strparse option to asn1pars program which parses nested
6604 [Dr Stephen Henson <shenson@bigfoot.com>]
6606 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6609 *) DSA fix for "ca" program.
6612 *) Added "-genkey" option to "dsaparam" program.
6615 *) Added RIPE MD160 (rmd160) message digest.
6618 *) Added -a (all) option to "ssleay version" command.
6621 *) Added PLATFORM define which is the id given to Configure.
6624 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6627 *) Extended the ASN.1 parser routines.
6630 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6633 *) Added a BN_CTX to the BN library.
6636 *) Fixed the weak key values in DES library
6639 *) Changed API in EVP library for cipher aliases.
6642 *) Added support for RC2/64bit cipher.
6645 *) Converted the lhash library to the crypto/mem.c functions.
6648 *) Added more recognized ASN.1 object ids.
6651 *) Added more RSA padding checks for SSL/TLS.
6654 *) Added BIO proxy/filter functionality.
6657 *) Added extra_certs to SSL_CTX which can be used
6658 send extra CA certificates to the client in the CA cert chain sending
6659 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6662 *) Now Fortezza is denied in the authentication phase because
6663 this is key exchange mechanism is not supported by SSLeay at all.
6666 *) Additional PKCS1 checks.
6669 *) Support the string "TLSv1" for all TLS v1 ciphers.
6672 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6673 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6676 *) Fixed a few memory leaks.
6679 *) Fixed various code and comment typos.
6682 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6683 bytes sent in the client random.
6684 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob