5 Changes between 0.9.7 and 0.9.8 [xx XXX xxxx]
7 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
8 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9 software implementations. For DSA and DH, parameter generation can
10 also be overriden by providing the appropriate method callbacks.
13 *) Change the "progress" mechanism used in key-generation and
14 primality testing to functions that take a new BN_GENCB pointer in
15 place of callback/argument pairs. The new API functions have "_ex"
16 postfixes and the older functions are reimplemented as wrappers for
17 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
18 declarations of the old functions to help (graceful) attempts to
19 migrate to the new functions. Also, the new key-generation API
20 functions operate on a caller-supplied key-structure and return
21 success/failure rather than returning a key or NULL - this is to
22 help make "keygen" another member function of RSA_METHOD etc.
24 Example for using the new callback interface:
26 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
30 BN_GENCB_set(&my_cb, my_callback, my_arg);
32 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
33 /* For the meaning of a, b in calls to my_callback(), see the
34 * documentation of the function that calls the callback.
35 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
36 * my_callback should return 1 if it wants BN_is_prime_ex()
37 * to continue, or 0 to stop.
42 *) Change the ZLIB compression method to be stateful, and make it
43 available to TLS with the number defined in
44 draft-ietf-tls-compression-04.txt.
47 *) Add the ASN.1 structures and functions for CertificatePair, which
48 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
50 CertificatePair ::= SEQUENCE {
51 forward [0] Certificate OPTIONAL,
52 reverse [1] Certificate OPTIONAL,
53 -- at least one of the pair shall be present -- }
55 Also implement the PEM functions to read and write certificate
56 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
58 This needed to be defined, mostly for the sake of the LDAP
59 attribute crossCertificatePair, but may prove useful elsewhere as
63 *) Make it possible to inhibit symlinking of shared libraries in
64 Makefile.shared, for Cygwin's sake.
67 *) Extend the BIGNUM API by creating new macros that behave like
70 void BN_set_sign(BIGNUM *a, int neg);
71 int BN_get_sign(const BIGNUM *a);
73 and avoid the need to access 'a->neg' directly in applications.
74 [Nils Larsch <nla@trustcenter.de>]
76 *) Implement fast modular reduction for pseudo-Mersenne primes
77 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
78 EC_GROUP_new_curve_GFp() will now automatically use this
80 [Nils Larsch <nla@trustcenter.de>]
82 *) Add new lock type (CRYPTO_LOCK_BN).
85 *) Change the ENGINE framework to automatically load engines
86 dynamically from specific directories unless they could be
87 found to already be built in or loaded. Move all the
88 current engines except for the cryptodev one to a new
90 The engines in engines/ are built as shared libraries if
91 the "shared" options was given to ./Configure or ./config.
92 Otherwise, they are inserted in libcrypto.a.
93 /usr/local/ssl/engines is the default directory for dynamic
94 engines, but that can be overriden at configure time through
95 the usual use of --prefix and/or --openssldir, and at run
96 time with the environment variable OPENSSL_ENGINES.
97 [Geoff Thorpe and Richard Levitte]
99 *) Add Makefile.shared, a helper makefile to build shared
100 libraries. Addapt Makefile.org.
103 *) Add version info to Win32 DLLs.
104 [Peter 'Luna' Runestig" <peter@runestig.com>]
106 *) Add new 'medium level' PKCS#12 API. Certificates and keys
107 can be added using this API to created arbitrary PKCS#12
108 files while avoiding the low level API.
110 New options to PKCS12_create(), key or cert can be NULL and
111 will then be omitted from the output file. The encryption
112 algorithm NIDs can be set to -1 for no encryption, the mac
113 iteration count can be set to 0 to omit the mac.
115 Enhance pkcs12 utility by making the -nokeys and -nocerts
116 options work when creating a PKCS#12 file. New option -nomac
117 to omit the mac, NONE can be set for an encryption algorithm.
118 New code is modified to use the enhanced PKCS12_create()
119 instead of the low level API.
122 *) Extend ASN1 encoder to support indefinite length constructed
123 encoding. This can output sequences tags and octet strings in
124 this form. Modify pk7_asn1.c to support indefinite length
125 encoding. This is experimental and needs additional code to
126 be useful, such as an ASN1 bio and some enhanced streaming
129 Extend template encode functionality so that tagging is passed
130 down to the template encoder.
133 *) Let 'openssl req' fail if an argument to '-newkey' is not
134 recognized instead of using RSA as a default.
137 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
138 As these are not official, they are not included in "ALL";
139 the "ECCdraft" ciphersuite group alias can be used to select them.
140 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
142 *) Add ECDH engine support.
143 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
145 *) Add ECDH in new directory crypto/ecdh/.
146 TODO: more general interface (return x coordinate, not its hash)
147 TODO: bug: pad x with leading zeros if necessary
148 [Douglas Stebila (Sun Microsystems Laboratories)]
150 *) Let BN_rand_range() abort with an error after 100 iterations
151 without success (which indicates a broken PRNG).
154 *) Change BN_mod_sqrt() so that it verifies that the input value
155 is really the square of the return value. (Previously,
156 BN_mod_sqrt would show GIGO behaviour.)
159 *) Add named elliptic curves over binary fields from X9.62, SECG,
160 and WAP/WTLS; add OIDs that were still missing.
162 [Sheueling Chang Shantz and Douglas Stebila
163 (Sun Microsystems Laboratories)]
165 *) Extend the EC library for elliptic curves over binary fields
166 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
169 EC_GF2m_simple_method
173 EC_GROUP_new_curve_GF2m
174 EC_GROUP_set_curve_GF2m
175 EC_GROUP_get_curve_GF2m
176 EC_POINT_set_affine_coordinates_GF2m
177 EC_POINT_get_affine_coordinates_GF2m
178 EC_POINT_set_compressed_coordinates_GF2m
180 Point compression for binary fields is disabled by default for
181 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
184 As binary polynomials are represented as BIGNUMs, various members
185 of the EC_GROUP and EC_POINT data structures can be shared
186 between the implementations for prime fields and binary fields;
187 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
188 are essentially identical to their ..._GFp counterparts.
189 (For simplicity, the '..._GFp' prefix has been dropped from
190 various internal method names.)
192 An internal 'field_div' method (similar to 'field_mul' and
193 'field_sqr') has been added; this is used only for binary fields.
195 [Sheueling Chang Shantz and Douglas Stebila
196 (Sun Microsystems Laboratories)]
198 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
199 through methods ('mul', 'precompute_mult').
201 The generic implementations (now internally called 'ec_wNAF_mul'
202 and 'ec_wNAF_precomputed_mult') remain the default if these
203 methods are undefined.
205 [Sheueling Chang Shantz and Douglas Stebila
206 (Sun Microsystems Laboratories)]
208 *) New function EC_GROUP_get_degree, which is defined through
209 EC_METHOD. For curves over prime fields, this returns the bit
210 length of the modulus.
212 [Sheueling Chang Shantz and Douglas Stebila
213 (Sun Microsystems Laboratories)]
215 *) New functions EC_GROUP_dup, EC_POINT_dup.
216 (These simply call ..._new and ..._copy).
218 [Sheueling Chang Shantz and Douglas Stebila
219 (Sun Microsystems Laboratories)]
221 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
222 Polynomials are represented as BIGNUMs (where the sign bit is not
223 used) in the following functions [macros]:
226 BN_GF2m_sub [= BN_GF2m_add]
227 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
228 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
229 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
231 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
232 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
233 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
234 BN_GF2m_cmp [= BN_ucmp]
236 (Note that only the 'mod' functions are actually for fields GF(2^m).
237 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
239 For some functions, an the irreducible polynomial defining a
240 field can be given as an 'unsigned int[]' with strictly
241 decreasing elements giving the indices of those bits that are set;
242 i.e., p[] represents the polynomial
243 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
245 p[0] > p[1] > ... > p[k] = 0.
246 This applies to the following functions:
251 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
252 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
255 BN_GF2m_mod_solve_quad_arr
259 Conversion can be performed by the following functions:
264 bntest.c has additional tests for binary polynomial arithmetic.
266 Two implementations for BN_GF2m_mod_div() are available.
267 The default algorithm simply uses BN_GF2m_mod_inv() and
268 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
269 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
270 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
272 [Sheueling Chang Shantz and Douglas Stebila
273 (Sun Microsystems Laboratories)]
275 *) Add new error code 'ERR_R_DISABLED' that can be used when some
276 functionality is disabled at compile-time.
277 [Douglas Stebila <douglas.stebila@sun.com>]
279 *) Change default behaviour of 'openssl asn1parse' so that more
280 information is visible when viewing, e.g., a certificate:
282 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
283 mode the content of non-printable OCTET STRINGs is output in a
284 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
285 avoid the appearance of a printable string.
286 [Nils Larsch <nla@trustcenter.de>]
288 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
290 EC_GROUP_set_asn1_flag()
291 EC_GROUP_get_asn1_flag()
292 EC_GROUP_set_point_conversion_form()
293 EC_GROUP_get_point_conversion_form()
294 These control ASN1 encoding details:
295 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
296 has been set to OPENSSL_EC_NAMED_CURVE.
297 - Points are encoded in uncompressed form by default; options for
298 asn1_for are as for point2oct, namely
299 POINT_CONVERSION_COMPRESSED
300 POINT_CONVERSION_UNCOMPRESSED
301 POINT_CONVERSION_HYBRID
303 Also add 'seed' and 'seed_len' members to EC_GROUP with access
307 EC_GROUP_get_seed_len()
308 This is used only for ASN1 purposes (so far).
309 [Nils Larsch <nla@trustcenter.de>]
311 *) Add 'field_type' member to EC_METHOD, which holds the NID
312 of the appropriate field type OID. The new function
313 EC_METHOD_get_field_type() returns this value.
314 [Nils Larsch <nla@trustcenter.de>]
321 providing useful interfaces to EC_POINT_point2oct() and
322 EC_POINT_oct2point().
323 [Nils Larsch <nla@trustcenter.de>]
325 *) Change internals of the EC library so that the functions
326 EC_GROUP_set_generator()
327 EC_GROUP_get_generator()
329 EC_GROUP_get_cofactor()
330 are implemented directly in crypto/ec/ec_lib.c and not dispatched
331 to methods, which would lead to unnecessary code duplication when
332 adding different types of curves.
333 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
335 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
336 arithmetic, and such that modified wNAFs are generated
337 (which avoid length expansion in many cases).
340 *) Add a function EC_GROUP_check_discriminant() (defined via
341 EC_METHOD) that verifies that the curve discriminant is non-zero.
343 Add a function EC_GROUP_check() that makes some sanity tests
344 on a EC_GROUP, its generator and order. This includes
345 EC_GROUP_check_discriminant().
346 [Nils Larsch <nla@trustcenter.de>]
348 *) Add ECDSA in new directory crypto/ecdsa/.
350 Add applications 'openssl ecparam' and 'openssl ecdsa'
351 (these are based on 'openssl dsaparam' and 'openssl dsa').
353 ECDSA support is also included in various other files across the
354 library. Most notably,
355 - 'openssl req' now has a '-newkey ecdsa:file' option;
356 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
357 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
358 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
359 them suitable for ECDSA where domain parameters must be
360 extracted before the specific public key;
361 - ECDSA engine support has been added.
362 [Nils Larsch <nla@trustcenter.de>]
364 *) Include some named elliptic curves, and add OIDs from X9.62,
365 SECG, and WAP/WTLS. Each curve can be obtained from the new
367 EC_GROUP_new_by_nid(),
368 and the list of available named curves can be obtained with
369 EC_get_builtin_curves().
370 Also add a 'curve_name' member to EC_GROUP objects, which can be
374 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
376 Changes between 0.9.7 and 0.9.7a [XX xxx 2003]
378 *) IA-32 assembler support enhancements: unified ELF targets, support
379 for SCO/Caldera platforms, fix for Cygwin shared build.
382 *) Add support for FreeBSD on sparc64. As a consequence, support for
383 FreeBSD on non-x86 processors is separate from x86 processors on
384 the config script, much like the NetBSD support.
385 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
387 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
389 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
390 code (06) was taken as the first octet of the session ID and the last
391 octet was ignored consequently. As a result SSLv2 client side session
392 caching could not have worked due to the session ID mismatch between
394 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
398 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
399 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
403 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
404 seems that in spite of existing for more than a year, many application
405 author have done nothing to provide the necessary callbacks, which
406 means that this particular engine will not work properly anywhere.
407 This is a very unfortunate situation which forces us, in the name
408 of usability, to give the hw_ncipher.c a static lock, which is part
410 NOTE: This is for the 0.9.7 series ONLY. This hack will never
411 appear in 0.9.8 or later. We EXPECT application authors to have
412 dealt properly with this when 0.9.8 is released (unless we actually
413 make such changes in the libcrypto locking code that changes will
414 have to be made anyway).
417 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
418 octets have been read, EOF or an error occurs. Without this change
419 some truncated ASN1 structures will not produce an error.
422 *) Disable Heimdal support, since it hasn't been fully implemented.
423 Still give the possibility to force the use of Heimdal, but with
424 warnings and a request that patches get sent to openssl-dev.
427 *) Add the VC-CE target, introduce the WINCE sysname, and add
428 INSTALL.WCE and appropriate conditionals to make it build.
429 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
431 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
432 cygssl-x.y.z.dll, where x, y and z are the major, minor and
433 edit numbers of the version.
434 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
436 *) Introduce safe string copy and catenation functions
437 (BUF_strlcpy() and BUF_strlcat()).
438 [Ben Laurie (CHATS) and Richard Levitte]
440 *) Avoid using fixed-size buffers for one-line DNs.
443 *) Add BUF_MEM_grow_clean() to avoid information leakage when
444 resizing buffers containing secrets, and use where appropriate.
447 *) Avoid using fixed size buffers for configuration file location.
450 *) Avoid filename truncation for various CA files.
453 *) Use sizeof in preference to magic numbers.
456 *) Avoid filename truncation in cert requests.
459 *) Add assertions to check for (supposedly impossible) buffer
463 *) Don't cache truncated DNS entries in the local cache (this could
464 potentially lead to a spoofing attack).
467 *) Fix various buffers to be large enough for hex/decimal
468 representations in a platform independent manner.
471 *) Add CRYPTO_realloc_clean() to avoid information leakage when
472 resizing buffers containing secrets, and use where appropriate.
475 *) Add BIO_indent() to avoid much slightly worrying code to do
479 *) Convert sprintf()/BIO_puts() to BIO_printf().
482 *) buffer_gets() could terminate with the buffer only half
486 *) Add assertions to prevent user-supplied crypto functions from
487 overflowing internal buffers by having large block sizes, etc.
490 *) New OPENSSL_assert() macro (similar to assert(), but enabled
494 *) Eliminate unused copy of key in RC4.
497 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
500 *) Fix off-by-one error in EGD path.
503 *) If RANDFILE path is too long, ignore instead of truncating.
506 *) Eliminate unused and incorrectly sized X.509 structure
510 *) Eliminate unused and dangerous function knumber().
513 *) Eliminate unused and dangerous structure, KSSL_ERR.
516 *) Protect against overlong session ID context length in an encoded
517 session object. Since these are local, this does not appear to be
521 *) Change from security patch (see 0.9.6e below) that did not affect
522 the 0.9.6 release series:
524 Remote buffer overflow in SSL3 protocol - an attacker could
525 supply an oversized master key in Kerberos-enabled versions.
529 *) Change the SSL kerb5 codes to match RFC 2712.
532 *) Make -nameopt work fully for req and add -reqopt switch.
533 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
535 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
536 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
538 *) Make sure tests can be performed even if the corresponding algorithms
539 have been removed entirely. This was also the last step to make
540 OpenSSL compilable with DJGPP under all reasonable conditions.
541 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
543 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
544 to allow version independent disabling of normally unselected ciphers,
545 which may be activated as a side-effect of selecting a single cipher.
547 (E.g., cipher list string "RSA" enables ciphersuites that are left
548 out of "ALL" because they do not provide symmetric encryption.
549 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
550 [Lutz Jaenicke, Bodo Moeller]
552 *) Add appropriate support for separate platform-dependent build
553 directories. The recommended way to make a platform-dependent
554 build directory is the following (tested on Linux), maybe with
557 # Place yourself outside of the OpenSSL source tree. In
558 # this example, the environment variable OPENSSL_SOURCE
559 # is assumed to contain the absolute OpenSSL source directory.
560 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
561 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
562 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
563 mkdir -p `dirname $F`
564 ln -s $OPENSSL_SOURCE/$F $F
567 To be absolutely sure not to disturb the source tree, a "make clean"
568 is a good thing. If it isn't successfull, don't worry about it,
569 it probably means the source directory is very clean.
572 *) Make sure any ENGINE control commands make local copies of string
573 pointers passed to them whenever necessary. Otherwise it is possible
574 the caller may have overwritten (or deallocated) the original string
575 data when a later ENGINE operation tries to use the stored values.
576 [Götz Babin-Ebell <babinebell@trustcenter.de>]
578 *) Improve diagnostics in file reading and command-line digests.
579 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
581 *) Add AES modes CFB and OFB to the object database. Correct an
582 error in AES-CFB decryption.
585 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
586 allows existing EVP_CIPHER_CTX structures to be reused after
587 calling EVP_*Final(). This behaviour is used by encryption
588 BIOs and some applications. This has the side effect that
589 applications must explicitly clean up cipher contexts with
590 EVP_CIPHER_CTX_cleanup() or they will leak memory.
593 *) Check the values of dna and dnb in bn_mul_recursive before calling
594 bn_mul_comba (a non zero value means the a or b arrays do not contain
595 n2 elements) and fallback to bn_mul_normal if either is not zero.
598 *) Fix escaping of non-ASCII characters when using the -subj option
599 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
602 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
603 form for "surname", serialNumber has no short form.
604 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
605 therefore remove "mail" short name for "internet 7".
606 The OID for unique identifiers in X509 certificates is
607 x500UniqueIdentifier, not uniqueIdentifier.
608 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
611 *) Add an "init" command to the ENGINE config module and auto initialize
612 ENGINEs. Without any "init" command the ENGINE will be initialized
613 after all ctrl commands have been executed on it. If init=1 the
614 ENGINE is initailized at that point (ctrls before that point are run
615 on the uninitialized ENGINE and after on the initialized one). If
616 init=0 then the ENGINE will not be iniatialized at all.
619 *) Fix the 'app_verify_callback' interface so that the user-defined
620 argument is actually passed to the callback: In the
621 SSL_CTX_set_cert_verify_callback() prototype, the callback
622 declaration has been changed from
625 int (*cb)(X509_STORE_CTX *,void *);
626 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
627 i=s->ctx->app_verify_callback(&ctx)
628 has been changed into
629 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
631 To update applications using SSL_CTX_set_cert_verify_callback(),
632 a dummy argument can be added to their callback functions.
633 [D. K. Smetters <smetters@parc.xerox.com>]
635 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
636 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
638 *) Add and OPENSSL_LOAD_CONF define which will cause
639 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
640 This allows older applications to transparently support certain
641 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
642 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
643 load the config file and OPENSSL_add_all_algorithms_conf() which will
644 always load it have also been added.
647 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
648 Adjust NIDs and EVP layer.
649 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
651 *) Config modules support in openssl utility.
653 Most commands now load modules from the config file,
654 though in a few (such as version) this isn't done
655 because it couldn't be used for anything.
657 In the case of ca and req the config file used is
658 the same as the utility itself: that is the -config
659 command line option can be used to specify an
663 *) Move default behaviour from OPENSSL_config(). If appname is NULL
664 use "openssl_conf" if filename is NULL use default openssl config file.
667 *) Add an argument to OPENSSL_config() to allow the use of an alternative
668 config section name. Add a new flag to tolerate a missing config file
669 and move code to CONF_modules_load_file().
672 *) Support for crypto accelerator cards from Accelerated Encryption
673 Processing, www.aep.ie. (Use engine 'aep')
674 The support was copied from 0.9.6c [engine] and adapted/corrected
675 to work with the new engine framework.
676 [AEP Inc. and Richard Levitte]
678 *) Support for SureWare crypto accelerator cards from Baltimore
679 Technologies. (Use engine 'sureware')
680 The support was copied from 0.9.6c [engine] and adapted
681 to work with the new engine framework.
684 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
685 make the newer ENGINE framework commands for the CHIL engine work.
686 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
688 *) Make it possible to produce shared libraries on ReliantUNIX.
689 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
691 *) Add the configuration target debug-linux-ppro.
692 Make 'openssl rsa' use the general key loading routines
693 implemented in apps.c, and make those routines able to
694 handle the key format FORMAT_NETSCAPE and the variant
696 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
698 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
699 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
701 *) Add -keyform to rsautl, and document -engine.
702 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
704 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
705 BIO_R_NO_SUCH_FILE error code rather than the generic
706 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
711 ERR_peek_last_error_line
712 ERR_peek_last_error_line_data.
716 ERR_peek_error_line_data,
717 but report on the latest error recorded rather than the first one
718 still in the error queue.
719 [Ben Laurie, Bodo Moeller]
721 *) default_algorithms option in ENGINE config module. This allows things
723 default_algorithms = ALL
724 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
727 *) Prelminary ENGINE config module.
730 *) New experimental application configuration code.
733 *) Change the AES code to follow the same name structure as all other
734 symmetric ciphers, and behave the same way. Move everything to
735 the directory crypto/aes, thereby obsoleting crypto/rijndael.
736 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
738 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
739 [Ben Laurie and Theo de Raadt]
741 *) Add option to output public keys in req command.
742 [Massimiliano Pala madwolf@openca.org]
744 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
745 (up to about 10% better than before for P-192 and P-224).
748 *) New functions/macros
750 SSL_CTX_set_msg_callback(ctx, cb)
751 SSL_CTX_set_msg_callback_arg(ctx, arg)
752 SSL_set_msg_callback(ssl, cb)
753 SSL_set_msg_callback_arg(ssl, arg)
755 to request calling a callback function
757 void cb(int write_p, int version, int content_type,
758 const void *buf, size_t len, SSL *ssl, void *arg)
760 whenever a protocol message has been completely received
761 (write_p == 0) or sent (write_p == 1). Here 'version' is the
762 protocol version according to which the SSL library interprets
763 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
764 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
765 the content type as defined in the SSL 3.0/TLS 1.0 protocol
766 specification (change_cipher_spec(20), alert(21), handshake(22)).
767 'buf' and 'len' point to the actual message, 'ssl' to the
768 SSL object, and 'arg' is the application-defined value set by
769 SSL[_CTX]_set_msg_callback_arg().
771 'openssl s_client' and 'openssl s_server' have new '-msg' options
772 to enable a callback that displays all protocol messages.
775 *) Change the shared library support so shared libraries are built as
776 soon as the corresponding static library is finished, and thereby get
777 openssl and the test programs linked against the shared library.
778 This still only happens when the keyword "shard" has been given to
779 the configuration scripts.
781 NOTE: shared library support is still an experimental thing, and
782 backward binary compatibility is still not guaranteed.
783 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
785 *) Add support for Subject Information Access extension.
786 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
788 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
789 additional bytes when new memory had to be allocated, not just
790 when reusing an existing buffer.
793 *) New command line and configuration option 'utf8' for the req command.
794 This allows field values to be specified as UTF8 strings.
797 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
798 runs for the former and machine-readable output for the latter.
801 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
802 of the e-mail address in the DN (i.e., it will go into a certificate
803 extension only). The new configuration file option 'email_in_dn = no'
805 [Massimiliano Pala madwolf@openca.org]
807 *) Change all functions with names starting with des_ to be starting
808 with DES_ instead. Add wrappers that are compatible with libdes,
809 but are named _ossl_old_des_*. Finally, add macros that map the
810 des_* symbols to the corresponding _ossl_old_des_* if libdes
811 compatibility is desired. If OpenSSL 0.9.6c compatibility is
812 desired, the des_* symbols will be mapped to DES_*, with one
815 Since we provide two compatibility mappings, the user needs to
816 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
817 compatibility is desired. The default (i.e., when that macro
818 isn't defined) is OpenSSL 0.9.6c compatibility.
820 There are also macros that enable and disable the support of old
821 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
822 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
823 are defined, the default will apply: to support the old des routines.
825 In either case, one must include openssl/des.h to get the correct
826 definitions. Do not try to just include openssl/des_old.h, that
829 NOTE: This is a major break of an old API into a new one. Software
830 authors are encouraged to switch to the DES_ style functions. Some
831 time in the future, des_old.h and the libdes compatibility functions
832 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
833 default), and then completely removed.
836 *) Test for certificates which contain unsupported critical extensions.
837 If such a certificate is found during a verify operation it is
838 rejected by default: this behaviour can be overridden by either
839 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
840 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
841 X509_supported_extension() has also been added which returns 1 if a
842 particular extension is supported.
845 *) Modify the behaviour of EVP cipher functions in similar way to digests
846 to retain compatibility with existing code.
849 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
850 compatibility with existing code. In particular the 'ctx' parameter does
851 not have to be to be initialized before the call to EVP_DigestInit() and
852 it is tidied up after a call to EVP_DigestFinal(). New function
853 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
854 EVP_MD_CTX_copy() changed to not require the destination to be
855 initialized valid and new function EVP_MD_CTX_copy_ex() added which
856 requires the destination to be valid.
858 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
859 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
862 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
863 so that complete 'Handshake' protocol structures are kept in memory
864 instead of overwriting 'msg_type' and 'length' with 'body' data.
867 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
868 [Massimo Santin via Richard Levitte]
870 *) Major restructuring to the underlying ENGINE code. This includes
871 reduction of linker bloat, separation of pure "ENGINE" manipulation
872 (initialisation, etc) from functionality dealing with implementations
873 of specific crypto iterfaces. This change also introduces integrated
874 support for symmetric ciphers and digest implementations - so ENGINEs
875 can now accelerate these by providing EVP_CIPHER and EVP_MD
876 implementations of their own. This is detailed in crypto/engine/README
877 as it couldn't be adequately described here. However, there are a few
878 API changes worth noting - some RSA, DSA, DH, and RAND functions that
879 were changed in the original introduction of ENGINE code have now
880 reverted back - the hooking from this code to ENGINE is now a good
881 deal more passive and at run-time, operations deal directly with
882 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
883 dereferencing through an ENGINE pointer any more. Also, the ENGINE
884 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
885 they were not being used by the framework as there is no concept of a
886 BIGNUM_METHOD and they could not be generalised to the new
887 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
888 ENGINE_cpy() has been removed as it cannot be consistently defined in
892 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
895 *) Change mkdef.pl to sort symbols that get the same entry number,
896 and make sure the automatically generated functions ERR_load_*
897 become part of libeay.num as well.
900 *) New function SSL_renegotiate_pending(). This returns true once
901 renegotiation has been requested (either SSL_renegotiate() call
902 or HelloRequest/ClientHello receveived from the peer) and becomes
903 false once a handshake has been completed.
904 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
905 sends a HelloRequest, but does not ensure that a handshake takes
906 place. SSL_renegotiate_pending() is useful for checking if the
907 client has followed the request.)
910 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
911 By default, clients may request session resumption even during
912 renegotiation (if session ID contexts permit); with this option,
913 session resumption is possible only in the first handshake.
915 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
916 more bits available for options that should not be part of
917 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
920 *) Add some demos for certificate and certificate request creation.
923 *) Make maximum certificate chain size accepted from the peer application
924 settable (SSL*_get/set_max_cert_list()), as proposed by
925 "Douglas E. Engert" <deengert@anl.gov>.
928 *) Add support for shared libraries for Unixware-7
929 (Boyd Lynn Gerber <gerberb@zenez.com>).
932 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
933 be done prior to destruction. Use this to unload error strings from
934 ENGINEs that load their own error strings. NB: This adds two new API
935 functions to "get" and "set" this destroy handler in an ENGINE.
938 *) Alter all existing ENGINE implementations (except "openssl" and
939 "openbsd") to dynamically instantiate their own error strings. This
940 makes them more flexible to be built both as statically-linked ENGINEs
941 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
942 Also, add stub code to each that makes building them as self-contained
943 shared-libraries easier (see README.ENGINE).
946 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
947 implementations into applications that are completely implemented in
948 self-contained shared-libraries. The "dynamic" ENGINE exposes control
949 commands that can be used to configure what shared-library to load and
950 to control aspects of the way it is handled. Also, made an update to
951 the README.ENGINE file that brings its information up-to-date and
952 provides some information and instructions on the "dynamic" ENGINE
953 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
956 *) Make it possible to unload ranges of ERR strings with a new
957 "ERR_unload_strings" function.
960 *) Add a copy() function to EVP_MD.
963 *) Make EVP_MD routines take a context pointer instead of just the
964 md_data void pointer.
967 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
968 that the digest can only process a single chunk of data
969 (typically because it is provided by a piece of
970 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
971 is only going to provide a single chunk of data, and hence the
972 framework needn't accumulate the data for oneshot drivers.
975 *) As with "ERR", make it possible to replace the underlying "ex_data"
976 functions. This change also alters the storage and management of global
977 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
978 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
979 index counters. The API functions that use this state have been changed
980 to take a "class_index" rather than pointers to the class's local STACK
981 and counter, and there is now an API function to dynamically create new
982 classes. This centralisation allows us to (a) plug a lot of the
983 thread-safety problems that existed, and (b) makes it possible to clean
984 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
985 such data would previously have always leaked in application code and
986 workarounds were in place to make the memory debugging turn a blind eye
987 to it. Application code that doesn't use this new function will still
988 leak as before, but their memory debugging output will announce it now
989 rather than letting it slide.
991 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
992 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
993 has a return value to indicate success or failure.
996 *) Make it possible to replace the underlying "ERR" functions such that the
997 global state (2 LHASH tables and 2 locks) is only used by the "default"
998 implementation. This change also adds two functions to "get" and "set"
999 the implementation prior to it being automatically set the first time
1000 any other ERR function takes place. Ie. an application can call "get",
1001 pass the return value to a module it has just loaded, and that module
1002 can call its own "set" function using that value. This means the
1003 module's "ERR" operations will use (and modify) the error state in the
1004 application and not in its own statically linked copy of OpenSSL code.
1007 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1008 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1009 the operation, and provides a more encapsulated way for external code
1010 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1011 to use these functions rather than manually incrementing the counts.
1013 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1016 *) Add EVP test program.
1019 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1022 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1023 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1024 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1025 These allow a CRL to be built without having to access X509_CRL fields
1026 directly. Modify 'ca' application to use new functions.
1029 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1030 bug workarounds. Rollback attack detection is a security feature.
1031 The problem will only arise on OpenSSL servers when TLSv1 is not
1032 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1033 Software authors not wanting to support TLSv1 will have special reasons
1034 for their choice and can explicitly enable this option.
1035 [Bodo Moeller, Lutz Jaenicke]
1037 *) Rationalise EVP so it can be extended: don't include a union of
1038 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1039 (similar to those existing for EVP_CIPHER_CTX).
1044 EVP_MD_CTX_init(&md); /* new function call */
1045 EVP_DigestInit(&md, EVP_sha1());
1046 EVP_DigestUpdate(&md, in, len);
1047 EVP_DigestFinal(&md, out, NULL);
1048 EVP_MD_CTX_cleanup(&md); /* new function call */
1052 *) Make DES key schedule conform to the usual scheme, as well as
1053 correcting its structure. This means that calls to DES functions
1054 now have to pass a pointer to a des_key_schedule instead of a
1055 plain des_key_schedule (which was actually always a pointer
1058 des_key_schedule ks;
1060 des_set_key_checked(..., &ks);
1061 des_ncbc_encrypt(..., &ks, ...);
1063 (Note that a later change renames 'des_...' into 'DES_...'.)
1066 *) Initial reduction of linker bloat: the use of some functions, such as
1067 PEM causes large amounts of unused functions to be linked in due to
1068 poor organisation. For example pem_all.c contains every PEM function
1069 which has a knock on effect of linking in large amounts of (unused)
1070 ASN1 code. Grouping together similar functions and splitting unrelated
1071 functions prevents this.
1074 *) Cleanup of EVP macros.
1077 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1078 correct _ecb suffix.
1081 *) Add initial OCSP responder support to ocsp application. The
1082 revocation information is handled using the text based index
1083 use by the ca application. The responder can either handle
1084 requests generated internally, supplied in files (for example
1085 via a CGI script) or using an internal minimal server.
1088 *) Add configuration choices to get zlib compression for TLS.
1091 *) Changes to Kerberos SSL for RFC 2712 compliance:
1092 1. Implemented real KerberosWrapper, instead of just using
1093 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1094 2. Implemented optional authenticator field of KerberosWrapper.
1096 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1097 and authenticator structs; see crypto/krb5/.
1099 Generalized Kerberos calls to support multiple Kerberos libraries.
1100 [Vern Staats <staatsvr@asc.hpc.mil>,
1101 Jeffrey Altman <jaltman@columbia.edu>
1102 via Richard Levitte]
1104 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1105 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1106 values for each of the key sizes rather than having just
1107 parameters (and 'speed' generating keys each time).
1110 *) Speed up EVP routines.
1113 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1114 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1115 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1116 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1118 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1119 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1120 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1123 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1125 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1128 *) Added the OS2-EMX target.
1129 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1131 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1132 to support NCONF routines in extension code. New function CONF_set_nconf()
1133 to allow functions which take an NCONF to also handle the old LHASH
1134 structure: this means that the old CONF compatible routines can be
1135 retained (in particular wrt extensions) without having to duplicate the
1136 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1139 *) Enhance the general user interface with mechanisms for inner control
1140 and with possibilities to have yes/no kind of prompts.
1143 *) Change all calls to low level digest routines in the library and
1144 applications to use EVP. Add missing calls to HMAC_cleanup() and
1145 don't assume HMAC_CTX can be copied using memcpy().
1146 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1148 *) Add the possibility to control engines through control names but with
1149 arbitrary arguments instead of just a string.
1150 Change the key loaders to take a UI_METHOD instead of a callback
1151 function pointer. NOTE: this breaks binary compatibility with earlier
1152 versions of OpenSSL [engine].
1153 Adapt the nCipher code for these new conditions and add a card insertion
1157 *) Enhance the general user interface with mechanisms to better support
1158 dialog box interfaces, application-defined prompts, the possibility
1159 to use defaults (for example default passwords from somewhere else)
1160 and interrupts/cancellations.
1163 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1164 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1167 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1168 tidy up some unnecessarily weird code in 'sk_new()').
1169 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1171 *) Change the key loading routines for ENGINEs to use the same kind
1172 callback (pem_password_cb) as all other routines that need this
1176 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1177 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1178 than this minimum value is recommended.
1181 *) New random seeder for OpenVMS, using the system process statistics
1182 that are easily reachable.
1185 *) Windows apparently can't transparently handle global
1186 variables defined in DLLs. Initialisations such as:
1188 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1190 wont compile. This is used by the any applications that need to
1191 declare their own ASN1 modules. This was fixed by adding the option
1192 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1193 needed for static libraries under Win32.
1196 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1197 setting of purpose and trust fields. New X509_STORE trust and
1198 purpose functions and tidy up setting in other SSL functions.
1201 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1202 structure. These are inherited by X509_STORE_CTX when it is
1203 initialised. This allows various defaults to be set in the
1204 X509_STORE structure (such as flags for CRL checking and custom
1205 purpose or trust settings) for functions which only use X509_STORE_CTX
1206 internally such as S/MIME.
1208 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1209 trust settings if they are not set in X509_STORE. This allows X509_STORE
1210 purposes and trust (in S/MIME for example) to override any set by default.
1212 Add command line options for CRL checking to smime, s_client and s_server
1216 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1217 are set then the CRL is looked up in the X509_STORE structure and
1218 its validity and signature checked, then if the certificate is found
1219 in the CRL the verify fails with a revoked error.
1221 Various new CRL related callbacks added to X509_STORE_CTX structure.
1223 Command line options added to 'verify' application to support this.
1225 This needs some additional work, such as being able to handle multiple
1226 CRLs with different times, extension based lookup (rather than just
1227 by subject name) and ultimately more complete V2 CRL extension
1231 *) Add a general user interface API (crypto/ui/). This is designed
1232 to replace things like des_read_password and friends (backward
1233 compatibility functions using this new API are provided).
1234 The purpose is to remove prompting functions from the DES code
1235 section as well as provide for prompting through dialog boxes in
1236 a window system and the like.
1239 *) Add "ex_data" support to ENGINE so implementations can add state at a
1240 per-structure level rather than having to store it globally.
1243 *) Make it possible for ENGINE structures to be copied when retrieved by
1244 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1245 This causes the "original" ENGINE structure to act like a template,
1246 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1247 operational state can be localised to each ENGINE structure, despite the
1248 fact they all share the same "methods". New ENGINE structures returned in
1249 this case have no functional references and the return value is the single
1250 structural reference. This matches the single structural reference returned
1251 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1255 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1256 needs to match any other type at all we need to manually clear the
1260 *) Changes to the "openssl engine" utility to include;
1261 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1262 about an ENGINE's available control commands.
1263 - executing control commands from command line arguments using the
1264 '-pre' and '-post' switches. '-post' is only used if '-t' is
1265 specified and the ENGINE is successfully initialised. The syntax for
1266 the individual commands are colon-separated, for example;
1267 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1270 *) New dynamic control command support for ENGINEs. ENGINEs can now
1271 declare their own commands (numbers), names (strings), descriptions,
1272 and input types for run-time discovery by calling applications. A
1273 subset of these commands are implicitly classed as "executable"
1274 depending on their input type, and only these can be invoked through
1275 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1276 can be based on user input, config files, etc). The distinction is
1277 that "executable" commands cannot return anything other than a boolean
1278 result and can only support numeric or string input, whereas some
1279 discoverable commands may only be for direct use through
1280 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1281 pointers, or other custom uses. The "executable" commands are to
1282 support parameterisations of ENGINE behaviour that can be
1283 unambiguously defined by ENGINEs and used consistently across any
1284 OpenSSL-based application. Commands have been added to all the
1285 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1286 control over shared-library paths without source code alterations.
1289 *) Changed all ENGINE implementations to dynamically allocate their
1290 ENGINEs rather than declaring them statically. Apart from this being
1291 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1292 this also allows the implementations to compile without using the
1293 internal engine_int.h header.
1296 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1297 'const' value. Any code that should be able to modify a RAND_METHOD
1298 should already have non-const pointers to it (ie. they should only
1299 modify their own ones).
1302 *) Made a variety of little tweaks to the ENGINE code.
1303 - "atalla" and "ubsec" string definitions were moved from header files
1304 to C code. "nuron" string definitions were placed in variables
1305 rather than hard-coded - allowing parameterisation of these values
1306 later on via ctrl() commands.
1307 - Removed unused "#if 0"'d code.
1308 - Fixed engine list iteration code so it uses ENGINE_free() to release
1309 structural references.
1310 - Constified the RAND_METHOD element of ENGINE structures.
1311 - Constified various get/set functions as appropriate and added
1312 missing functions (including a catch-all ENGINE_cpy that duplicates
1313 all ENGINE values onto a new ENGINE except reference counts/state).
1314 - Removed NULL parameter checks in get/set functions. Setting a method
1315 or function to NULL is a way of cancelling out a previously set
1316 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1317 and doesn't justify the extra error symbols and code.
1318 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1319 flags from engine_int.h to engine.h.
1320 - Changed prototypes for ENGINE handler functions (init(), finish(),
1321 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1324 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1325 to the algorithm using long division. The binary algorithm can be
1326 used only if the modulus is odd. On 32-bit systems, it is faster
1327 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1328 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1329 up to 450 bits. In 64-bit environments, the binary algorithm
1330 appears to be advantageous for much longer moduli; here we use it
1331 for moduli up to 2048 bits.
1334 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1335 could not support the combine flag in choice fields.
1338 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1339 extensions from a certificate request to the certificate.
1342 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1343 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1344 file: this allows the display of the certificate about to be
1345 signed to be customised, to allow certain fields to be included
1346 or excluded and extension details. The old system didn't display
1347 multicharacter strings properly, omitted fields not in the policy
1348 and couldn't display additional details such as extensions.
1351 *) Function EC_POINTs_mul for multiple scalar multiplication
1352 of an arbitrary number of elliptic curve points
1353 \sum scalars[i]*points[i],
1354 optionally including the generator defined for the EC_GROUP:
1355 scalar*generator + \sum scalars[i]*points[i].
1357 EC_POINT_mul is a simple wrapper function for the typical case
1358 that the point list has just one item (besides the optional
1362 *) First EC_METHODs for curves over GF(p):
1364 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1365 operations and provides various method functions that can also
1366 operate with faster implementations of modular arithmetic.
1368 EC_GFp_mont_method() reuses most functions that are part of
1369 EC_GFp_simple_method, but uses Montgomery arithmetic.
1371 [Bodo Moeller; point addition and point doubling
1372 implementation directly derived from source code provided by
1373 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1375 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1376 crypto/ec/ec_lib.c):
1378 Curves are EC_GROUP objects (with an optional group generator)
1379 based on EC_METHODs that are built into the library.
1381 Points are EC_POINT objects based on EC_GROUP objects.
1383 Most of the framework would be able to handle curves over arbitrary
1384 finite fields, but as there are no obvious types for fields other
1385 than GF(p), some functions are limited to that for now.
1388 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1389 that the file contains a complete HTTP response.
1392 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1393 change the def and num file printf format specifier from "%-40sXXX"
1394 to "%-39s XXX". The latter will always guarantee a space after the
1395 field while the former will cause them to run together if the field
1396 is 40 of more characters long.
1399 *) Constify the cipher and digest 'method' functions and structures
1400 and modify related functions to take constant EVP_MD and EVP_CIPHER
1404 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1405 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1408 *) Modify EVP_Digest*() routines so they now return values. Although the
1409 internal software routines can never fail additional hardware versions
1413 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1415 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1416 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1419 ERR_R_NESTED_ASN1_ERROR
1421 ERR_R_MISSING_ASN1_EOS
1422 were 4 .. 9, conflicting with
1423 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1425 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1426 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1428 Add new error code 'ERR_R_INTERNAL_ERROR'.
1431 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1435 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1436 sets the subject name for a new request or supersedes the
1437 subject name in a given request. Formats that can be parsed are
1438 'CN=Some Name, OU=myOU, C=IT'
1440 'CN=Some Name/OU=myOU/C=IT'.
1442 Add options '-batch' and '-verbose' to 'openssl req'.
1443 [Massimiliano Pala <madwolf@hackmasters.net>]
1445 *) Introduce the possibility to access global variables through
1446 functions on platform were that's the best way to handle exporting
1447 global variables in shared libraries. To enable this functionality,
1448 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1449 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1450 is normally done by Configure or something similar).
1452 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1453 in the source file (foo.c) like this:
1455 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1456 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1458 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1459 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1461 OPENSSL_DECLARE_GLOBAL(int,foo);
1462 #define foo OPENSSL_GLOBAL_REF(foo)
1463 OPENSSL_DECLARE_GLOBAL(double,bar);
1464 #define bar OPENSSL_GLOBAL_REF(bar)
1466 The #defines are very important, and therefore so is including the
1467 header file everywhere where the defined globals are used.
1469 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1470 of ASN.1 items, but that structure is a bit different.
1472 The largest change is in util/mkdef.pl which has been enhanced with
1473 better and easier to understand logic to choose which symbols should
1474 go into the Windows .def files as well as a number of fixes and code
1475 cleanup (among others, algorithm keywords are now sorted
1476 lexicographically to avoid constant rewrites).
1479 *) In BN_div() keep a copy of the sign of 'num' before writing the
1480 result to 'rm' because if rm==num the value will be overwritten
1481 and produce the wrong result if 'num' is negative: this caused
1482 problems with BN_mod() and BN_nnmod().
1485 *) Function OCSP_request_verify(). This checks the signature on an
1486 OCSP request and verifies the signer certificate. The signer
1487 certificate is just checked for a generic purpose and OCSP request
1491 *) Add OCSP_check_validity() function to check the validity of OCSP
1492 responses. OCSP responses are prepared in real time and may only
1493 be a few seconds old. Simply checking that the current time lies
1494 between thisUpdate and nextUpdate max reject otherwise valid responses
1495 caused by either OCSP responder or client clock inaccuracy. Instead
1496 we allow thisUpdate and nextUpdate to fall within a certain period of
1497 the current time. The age of the response can also optionally be
1498 checked. Two new options -validity_period and -status_age added to
1502 *) If signature or public key algorithm is unrecognized print out its
1503 OID rather that just UNKNOWN.
1506 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1507 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1508 ID to be generated from the issuer certificate alone which can then be
1509 passed to OCSP_id_issuer_cmp().
1512 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1513 ASN1 modules to export functions returning ASN1_ITEM pointers
1514 instead of the ASN1_ITEM structures themselves. This adds several
1515 new macros which allow the underlying ASN1 function/structure to
1516 be accessed transparently. As a result code should not use ASN1_ITEM
1517 references directly (such as &X509_it) but instead use the relevant
1518 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1519 use of the new ASN1 code on platforms where exporting structures
1520 is problematical (for example in shared libraries) but exporting
1521 functions returning pointers to structures is not.
1524 *) Add support for overriding the generation of SSL/TLS session IDs.
1525 These callbacks can be registered either in an SSL_CTX or per SSL.
1526 The purpose of this is to allow applications to control, if they wish,
1527 the arbitrary values chosen for use as session IDs, particularly as it
1528 can be useful for session caching in multiple-server environments. A
1529 command-line switch for testing this (and any client code that wishes
1530 to use such a feature) has been added to "s_server".
1531 [Geoff Thorpe, Lutz Jaenicke]
1533 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1534 of the form '#if defined(...) || defined(...) || ...' and
1535 '#if !defined(...) && !defined(...) && ...'. This also avoids
1536 the growing number of special cases it was previously handling.
1539 *) Make all configuration macros available for application by making
1540 sure they are available in opensslconf.h, by giving them names starting
1541 with "OPENSSL_" to avoid conflicts with other packages and by making
1542 sure e_os2.h will cover all platform-specific cases together with
1544 Additionally, it is now possible to define configuration/platform-
1545 specific names (called "system identities"). In the C code, these
1546 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1547 macro with the name beginning with "OPENSSL_SYS_", which is determined
1548 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1552 *) New option -set_serial to 'req' and 'x509' this allows the serial
1553 number to use to be specified on the command line. Previously self
1554 signed certificates were hard coded with serial number 0 and the
1555 CA options of 'x509' had to use a serial number in a file which was
1559 *) New options to 'ca' utility to support V2 CRL entry extensions.
1560 Currently CRL reason, invalidity date and hold instruction are
1561 supported. Add new CRL extensions to V3 code and some new objects.
1564 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1565 disable standard block padding (aka PKCS#5 padding) in the EVP
1566 API, which was previously mandatory. This means that the data is
1567 not padded in any way and so the total length much be a multiple
1568 of the block size, otherwise an error occurs.
1571 *) Initial (incomplete) OCSP SSL support.
1574 *) New function OCSP_parse_url(). This splits up a URL into its host,
1575 port and path components: primarily to parse OCSP URLs. New -url
1576 option to ocsp utility.
1579 *) New nonce behavior. The return value of OCSP_check_nonce() now
1580 reflects the various checks performed. Applications can decide
1581 whether to tolerate certain situations such as an absent nonce
1582 in a response when one was present in a request: the ocsp application
1583 just prints out a warning. New function OCSP_add1_basic_nonce()
1584 this is to allow responders to include a nonce in a response even if
1585 the request is nonce-less.
1588 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1589 skipped when using openssl x509 multiple times on a single input file,
1590 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1593 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1594 set string type: to handle setting ASN1_TIME structures. Fix ca
1595 utility to correctly initialize revocation date of CRLs.
1598 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1599 the clients preferred ciphersuites and rather use its own preferences.
1600 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1601 Internet Explorer by ensuring unchanged hash method during stepup.
1602 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1605 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1606 to aes and add a new 'exist' option to print out symbols that don't
1610 *) Additional options to ocsp utility to allow flags to be set and
1611 additional certificates supplied.
1614 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1615 OCSP client a number of certificate to only verify the response
1619 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1620 handle the new API. Currently only ECB, CBC modes supported. Add new
1623 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1624 Encryption Standard (AES) Ciphersuites for Transport Layer
1625 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1626 not enabled by default and were not part of the "ALL" ciphersuite
1627 alias because they were not yet official; they could be
1628 explicitly requested by specifying the "AESdraft" ciphersuite
1629 group alias. In the final release of OpenSSL 0.9.7, the group
1630 alias is called "AES" and is part of "ALL".)
1631 [Ben Laurie, Steve Henson, Bodo Moeller]
1633 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1634 request to response.
1637 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1638 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1639 extract information from a certificate request. OCSP_response_create()
1640 creates a response and optionally adds a basic response structure.
1641 OCSP_basic_add1_status() adds a complete single response to a basic
1642 response and returns the OCSP_SINGLERESP structure just added (to allow
1643 extensions to be included for example). OCSP_basic_add1_cert() adds a
1644 certificate to a basic response and OCSP_basic_sign() signs a basic
1645 response with various flags. New helper functions ASN1_TIME_check()
1646 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1647 (converts ASN1_TIME to GeneralizedTime).
1650 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1651 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1652 structure from a certificate. X509_pubkey_digest() digests the public_key
1653 contents: this is used in various key identifiers.
1656 *) Make sk_sort() tolerate a NULL argument.
1657 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1659 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1660 passed by the function are trusted implicitly. If any of them signed the
1661 response then it is assumed to be valid and is not verified.
1664 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1665 to data. This was previously part of the PKCS7 ASN1 code. This
1666 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1667 [Steve Henson, reported by Kenneth R. Robinette
1668 <support@securenetterm.com>]
1670 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1671 routines: without these tracing memory leaks is very painful.
1672 Fix leaks in PKCS12 and PKCS7 routines.
1675 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1676 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1677 effectively meant GeneralizedTime would never be used. Now it
1678 is initialised to -1 but X509_time_adj() now has to check the value
1679 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1680 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1681 [Steve Henson, reported by Kenneth R. Robinette
1682 <support@securenetterm.com>]
1684 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1685 result in a zero length in the ASN1_INTEGER structure which was
1686 not consistent with the structure when d2i_ASN1_INTEGER() was used
1687 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1688 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1689 where it did not print out a minus for negative ASN1_INTEGER.
1692 *) Add summary printout to ocsp utility. The various functions which
1693 convert status values to strings have been renamed to:
1694 OCSP_response_status_str(), OCSP_cert_status_str() and
1695 OCSP_crl_reason_str() and are no longer static. New options
1696 to verify nonce values and to disable verification. OCSP response
1697 printout format cleaned up.
1700 *) Add additional OCSP certificate checks. These are those specified
1701 in RFC2560. This consists of two separate checks: the CA of the
1702 certificate being checked must either be the OCSP signer certificate
1703 or the issuer of the OCSP signer certificate. In the latter case the
1704 OCSP signer certificate must contain the OCSP signing extended key
1705 usage. This check is performed by attempting to match the OCSP
1706 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1707 in the OCSP_CERTID structures of the response.
1710 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1711 and related routines. This uses the standard OpenSSL certificate
1712 verify routines to perform initial checks (just CA validity) and
1713 to obtain the certificate chain. Then additional checks will be
1714 performed on the chain. Currently the root CA is checked to see
1715 if it is explicitly trusted for OCSP signing. This is used to set
1716 a root CA as a global signing root: that is any certificate that
1717 chains to that CA is an acceptable OCSP signing certificate.
1720 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1721 extensions from a separate configuration file.
1722 As when reading extensions from the main configuration file,
1723 the '-extensions ...' option may be used for specifying the
1725 [Massimiliano Pala <madwolf@comune.modena.it>]
1727 *) New OCSP utility. Allows OCSP requests to be generated or
1728 read. The request can be sent to a responder and the output
1729 parsed, outputed or printed in text form. Not complete yet:
1730 still needs to check the OCSP response validity.
1733 *) New subcommands for 'openssl ca':
1734 'openssl ca -status <serial>' prints the status of the cert with
1735 the given serial number (according to the index file).
1736 'openssl ca -updatedb' updates the expiry status of certificates
1738 [Massimiliano Pala <madwolf@comune.modena.it>]
1740 *) New '-newreq-nodes' command option to CA.pl. This is like
1741 '-newreq', but calls 'openssl req' with the '-nodes' option
1742 so that the resulting key is not encrypted.
1743 [Damien Miller <djm@mindrot.org>]
1745 *) New configuration for the GNU Hurd.
1746 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1748 *) Initial code to implement OCSP basic response verify. This
1749 is currently incomplete. Currently just finds the signer's
1750 certificate and verifies the signature on the response.
1753 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1754 value of OPENSSLDIR. This is available via the new '-d' option
1755 to 'openssl version', and is also included in 'openssl version -a'.
1758 *) Allowing defining memory allocation callbacks that will be given
1759 file name and line number information in additional arguments
1760 (a const char* and an int). The basic functionality remains, as
1761 well as the original possibility to just replace malloc(),
1762 realloc() and free() by functions that do not know about these
1763 additional arguments. To register and find out the current
1764 settings for extended allocation functions, the following
1765 functions are provided:
1767 CRYPTO_set_mem_ex_functions
1768 CRYPTO_set_locked_mem_ex_functions
1769 CRYPTO_get_mem_ex_functions
1770 CRYPTO_get_locked_mem_ex_functions
1772 These work the same way as CRYPTO_set_mem_functions and friends.
1773 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1774 extended allocation function is enabled.
1775 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1776 a conventional allocation function is enabled.
1777 [Richard Levitte, Bodo Moeller]
1779 *) Finish off removing the remaining LHASH function pointer casts.
1780 There should no longer be any prototype-casting required when using
1781 the LHASH abstraction, and any casts that remain are "bugs". See
1782 the callback types and macros at the head of lhash.h for details
1783 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1786 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1787 If /dev/[u]random devices are not available or do not return enough
1788 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1790 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1791 /etc/entropy will be queried once each in this sequence, quering stops
1792 when enough entropy was collected without querying more sockets.
1795 *) Change the Unix RAND_poll() variant to be able to poll several
1796 random devices, as specified by DEVRANDOM, until a sufficient amount
1797 of data has been collected. We spend at most 10 ms on each file
1798 (select timeout) and read in non-blocking mode. DEVRANDOM now
1799 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1800 (previously it was just the string "/dev/urandom"), so on typical
1801 platforms the 10 ms delay will never occur.
1802 Also separate out the Unix variant to its own file, rand_unix.c.
1803 For VMS, there's a currently-empty rand_vms.c.
1806 *) Move OCSP client related routines to ocsp_cl.c. These
1807 provide utility functions which an application needing
1808 to issue a request to an OCSP responder and analyse the
1809 response will typically need: as opposed to those which an
1810 OCSP responder itself would need which will be added later.
1812 OCSP_request_sign() signs an OCSP request with an API similar
1813 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1814 response. OCSP_response_get1_basic() extracts basic response
1815 from response. OCSP_resp_find_status(): finds and extracts status
1816 information from an OCSP_CERTID structure (which will be created
1817 when the request structure is built). These are built from lower
1818 level functions which work on OCSP_SINGLERESP structures but
1819 wont normally be used unless the application wishes to examine
1820 extensions in the OCSP response for example.
1822 Replace nonce routines with a pair of functions.
1823 OCSP_request_add1_nonce() adds a nonce value and optionally
1824 generates a random value. OCSP_check_nonce() checks the
1825 validity of the nonce in an OCSP response.
1828 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1829 This doesn't copy the supplied OCSP_CERTID and avoids the
1830 need to free up the newly created id. Change return type
1831 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1832 This can then be used to add extensions to the request.
1833 Deleted OCSP_request_new(), since most of its functionality
1834 is now in OCSP_REQUEST_new() (and the case insensitive name
1835 clash) apart from the ability to set the request name which
1836 will be added elsewhere.
1839 *) Update OCSP API. Remove obsolete extensions argument from
1840 various functions. Extensions are now handled using the new
1841 OCSP extension code. New simple OCSP HTTP function which
1842 can be used to send requests and parse the response.
1845 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1846 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1847 uses the special reorder version of SET OF to sort the attributes
1848 and reorder them to match the encoded order. This resolves a long
1849 standing problem: a verify on a PKCS7 structure just after signing
1850 it used to fail because the attribute order did not match the
1851 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1852 it uses the received order. This is necessary to tolerate some broken
1853 software that does not order SET OF. This is handled by encoding
1854 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1855 to produce the required SET OF.
1858 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1859 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1860 files to get correct declarations of the ASN.1 item variables.
1863 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1864 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1865 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1866 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1867 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1868 ASN1_ITEM and no wrapper functions.
1871 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1872 replace the old function pointer based I/O routines. Change most of
1873 the *_d2i_bio() and *_d2i_fp() functions to use these.
1876 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1877 lines, recognice more "algorithms" that can be deselected, and make
1878 it complain about algorithm deselection that isn't recognised.
1881 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1882 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1883 to use new functions. Add NO_ASN1_OLD which can be set to remove
1884 some old style ASN1 functions: this can be used to determine if old
1885 code will still work when these eventually go away.
1888 *) New extension functions for OCSP structures, these follow the
1889 same conventions as certificates and CRLs.
1892 *) New function X509V3_add1_i2d(). This automatically encodes and
1893 adds an extension. Its behaviour can be customised with various
1894 flags to append, replace or delete. Various wrappers added for
1895 certifcates and CRLs.
1898 *) Fix to avoid calling the underlying ASN1 print routine when
1899 an extension cannot be parsed. Correct a typo in the
1900 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1903 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1904 entries for variables.
1907 *) Add functionality to apps/openssl.c for detecting locking
1908 problems: As the program is single-threaded, all we have
1909 to do is register a locking callback using an array for
1910 storing which locks are currently held by the program.
1913 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1914 SSL_get_ex_data_X509_STORE_idx(), which is used in
1915 ssl_verify_cert_chain() and thus can be called at any time
1916 during TLS/SSL handshakes so that thread-safety is essential.
1917 Unfortunately, the ex_data design is not at all suited
1918 for multi-threaded use, so it probably should be abolished.
1921 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1922 [Broadcom, tweaked and integrated by Geoff Thorpe]
1924 *) Move common extension printing code to new function
1925 X509V3_print_extensions(). Reorganise OCSP print routines and
1926 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1929 *) New function X509_signature_print() to remove duplication in some
1933 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1934 set (this was treated exactly the same as SET OF previously). This
1935 is used to reorder the STACK representing the structure to match the
1936 encoding. This will be used to get round a problem where a PKCS7
1937 structure which was signed could not be verified because the STACK
1938 order did not reflect the encoded order.
1941 *) Reimplement the OCSP ASN1 module using the new code.
1944 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1945 for its ASN1 operations. The old style function pointers still exist
1946 for now but they will eventually go away.
1949 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1950 completely replaces the old ASN1 functionality with a table driven
1951 encoder and decoder which interprets an ASN1_ITEM structure describing
1952 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1953 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1954 has also been converted to the new form.
1957 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1958 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1959 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1960 for negative moduli.
1963 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1964 of not touching the result's sign bit.
1967 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1971 *) Changed the LHASH code to use prototypes for callbacks, and created
1972 macros to declare and implement thin (optionally static) functions
1973 that provide type-safety and avoid function pointer casting for the
1974 type-specific callbacks.
1977 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1979 [Veers Staats <staatsvr@asc.hpc.mil>,
1980 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1982 *) Reformat the FAQ so the different questions and answers can be divided
1983 in sections depending on the subject.
1986 *) Have the zlib compression code load ZLIB.DLL dynamically under
1990 *) New function BN_mod_sqrt for computing square roots modulo a prime
1991 (using the probabilistic Tonelli-Shanks algorithm unless
1992 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1993 be handled deterministically).
1994 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1996 *) Make BN_mod_inverse faster by explicitly handling small quotients
1997 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1998 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2001 *) New function BN_kronecker.
2004 *) Fix BN_gcd so that it works on negative inputs; the result is
2005 positive unless both parameters are zero.
2006 Previously something reasonably close to an infinite loop was
2007 possible because numbers could be growing instead of shrinking
2008 in the implementation of Euclid's algorithm.
2011 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2012 sign of the number in question.
2014 Fix BN_is_word(a,w) to work correctly for w == 0.
2016 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2017 because its test if the absolute value of 'a' equals 'w'.
2018 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2019 it exists mostly for use in the implementations of BN_is_zero(),
2020 BN_is_one(), and BN_is_word().
2023 *) New function BN_swap.
2026 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2027 the exponentiation functions are more likely to produce reasonable
2028 results on negative inputs.
2031 *) Change BN_mod_mul so that the result is always non-negative.
2032 Previously, it could be negative if one of the factors was negative;
2033 I don't think anyone really wanted that behaviour.
2036 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2037 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2038 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2039 and add new functions:
2048 BN_mod_lshift1_quick
2052 These functions always generate non-negative results.
2054 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2055 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2057 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2058 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2059 be reduced modulo m.
2060 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2062 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2063 was actually never needed) and in BN_mul(). The removal in BN_mul()
2064 required a small change in bn_mul_part_recursive() and the addition
2065 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2066 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2067 bn_sub_words() and bn_add_words() except they take arrays with
2071 *) In 'openssl passwd', verify passwords read from the terminal
2072 unless the '-salt' option is used (which usually means that
2073 verification would just waste user's time since the resulting
2074 hash is going to be compared with some given password hash)
2075 or the new '-noverify' option is used.
2077 This is an incompatible change, but it does not affect
2078 non-interactive use of 'openssl passwd' (passwords on the command
2079 line, '-stdin' option, '-in ...' option) and thus should not
2083 *) Remove all references to RSAref, since there's no more need for it.
2086 *) Make DSO load along a path given through an environment variable
2087 (SHLIB_PATH) with shl_load().
2090 *) Constify the ENGINE code as a result of BIGNUM constification.
2091 Also constify the RSA code and most things related to it. In a
2092 few places, most notable in the depth of the ASN.1 code, ugly
2093 casts back to non-const were required (to be solved at a later
2097 *) Make it so the openssl application has all engines loaded by default.
2100 *) Constify the BIGNUM routines a little more.
2103 *) Add the following functions:
2105 ENGINE_load_cswift()
2107 ENGINE_load_atalla()
2109 ENGINE_load_builtin_engines()
2111 That way, an application can itself choose if external engines that
2112 are built-in in OpenSSL shall ever be used or not. The benefit is
2113 that applications won't have to be linked with libdl or other dso
2114 libraries unless it's really needed.
2116 Changed 'openssl engine' to load all engines on demand.
2117 Changed the engine header files to avoid the duplication of some
2118 declarations (they differed!).
2121 *) 'openssl engine' can now list capabilities.
2124 *) Better error reporting in 'openssl engine'.
2127 *) Never call load_dh_param(NULL) in s_server.
2130 *) Add engine application. It can currently list engines by name and
2131 identity, and test if they are actually available.
2134 *) Improve RPM specification file by forcing symbolic linking and making
2135 sure the installed documentation is also owned by root.root.
2136 [Damien Miller <djm@mindrot.org>]
2138 *) Give the OpenSSL applications more possibilities to make use of
2139 keys (public as well as private) handled by engines.
2142 *) Add OCSP code that comes from CertCo.
2145 *) Add VMS support for the Rijndael code.
2148 *) Added untested support for Nuron crypto accelerator.
2151 *) Add support for external cryptographic devices. This code was
2152 previously distributed separately as the "engine" branch.
2153 [Geoff Thorpe, Richard Levitte]
2155 *) Rework the filename-translation in the DSO code. It is now possible to
2156 have far greater control over how a "name" is turned into a filename
2157 depending on the operating environment and any oddities about the
2158 different shared library filenames on each system.
2161 *) Support threads on FreeBSD-elf in Configure.
2164 *) Fix for SHA1 assembly problem with MASM: it produces
2165 warnings about corrupt line number information when assembling
2166 with debugging information. This is caused by the overlapping
2168 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2171 NCONF_get_number() has no error checking at all. As a replacement,
2172 NCONF_get_number_e() is defined (_e for "error checking") and is
2173 promoted strongly. The old NCONF_get_number is kept around for
2174 binary backward compatibility.
2175 Make it possible for methods to load from something other than a BIO,
2176 by providing a function pointer that is given a name instead of a BIO.
2177 For example, this could be used to load configuration data from an
2181 *) Fix for non blocking accept BIOs. Added new I/O special reason
2182 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2183 with non blocking I/O was not possible because no retry code was
2184 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2188 *) Added the beginnings of Rijndael support.
2191 *) Fix for bug in DirectoryString mask setting. Add support for
2192 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2193 to allow certificate printing to more controllable, additional
2194 'certopt' option to 'x509' to allow new printing options to be
2198 *) Clean old EAY MD5 hack from e_os.h.
2201 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2203 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2204 memory from it's contents. This is done with a counter that will
2205 place alternating values in each byte. This can be used to solve
2206 two issues: 1) the removal of calls to memset() by highly optimizing
2207 compilers, and 2) cleansing with other values than 0, since those can
2208 be read through on certain media, for example a swap space on disk.
2211 *) Bugfix: client side session caching did not work with external caching,
2212 because the session->cipher setting was not restored when reloading
2213 from the external cache. This problem was masked, when
2214 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2215 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2218 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2219 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2220 [Zeev Lieber <zeev-l@yahoo.com>]
2222 *) Undo an undocumented change introduced in 0.9.6e which caused
2223 repeated calls to OpenSSL_add_all_ciphers() and
2224 OpenSSL_add_all_digests() to be ignored, even after calling
2228 *) Change the default configuration reader to deal with last line not
2229 being properly terminated.
2232 *) Change X509_NAME_cmp() so it applies the special rules on handling
2233 DN values that are of type PrintableString, as well as RDNs of type
2234 emailAddress where the value has the type ia5String.
2235 [stefank@valicert.com via Richard Levitte]
2237 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2238 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2239 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2240 the bitwise-OR of the two for use by the majority of applications
2241 wanting this behaviour, and update the docs. The documented
2242 behaviour and actual behaviour were inconsistent and had been
2243 changing anyway, so this is more a bug-fix than a behavioural
2245 [Geoff Thorpe, diagnosed by Nadav Har'El]
2247 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2248 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2251 *) Fix initialization code race conditions in
2252 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2253 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2254 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2255 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2256 ssl2_get_cipher_by_char(),
2257 ssl3_get_cipher_by_char().
2258 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2260 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2261 the cached sessions are flushed, as the remove_cb() might use ex_data
2262 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2263 (see [openssl.org #212]).
2264 [Geoff Thorpe, Lutz Jaenicke]
2266 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2267 length, instead of the encoding length to d2i_ASN1_OBJECT.
2270 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2272 *) [In 0.9.6g-engine release:]
2273 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2274 [Lynn Gazis <lgazis@rainbow.com>]
2276 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2278 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2279 and get fix the header length calculation.
2280 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2281 Alon Kantor <alonk@checkpoint.com> (and others),
2284 *) Use proper error handling instead of 'assertions' in buffer
2285 overflow checks added in 0.9.6e. This prevents DoS (the
2286 assertions could call abort()).
2287 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2289 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2291 *) Add various sanity checks to asn1_get_length() to reject
2292 the ASN1 length bytes if they exceed sizeof(long), will appear
2293 negative or the content length exceeds the length of the
2295 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2297 *) Fix cipher selection routines: ciphers without encryption had no flags
2298 for the cipher strength set and where therefore not handled correctly
2299 by the selection routines (PR #130).
2302 *) Fix EVP_dsa_sha macro.
2306 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2307 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2308 that was added in OpenSSL 0.9.6d.
2310 As the countermeasure turned out to be incompatible with some
2311 broken SSL implementations, the new option is part of SSL_OP_ALL.
2312 SSL_OP_ALL is usually employed when compatibility with weird SSL
2313 implementations is desired (e.g. '-bugs' option to 's_client' and
2314 's_server'), so the new option is automatically set in many
2318 *) Changes in security patch:
2320 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2321 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2322 Air Force Materiel Command, USAF, under agreement number
2325 *) Add various sanity checks to asn1_get_length() to reject
2326 the ASN1 length bytes if they exceed sizeof(long), will appear
2327 negative or the content length exceeds the length of the
2328 supplied buffer. (CAN-2002-0659)
2329 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2331 *) Assertions for various potential buffer overflows, not known to
2333 [Ben Laurie (CHATS)]
2335 *) Various temporary buffers to hold ASCII versions of integers were
2336 too small for 64 bit platforms. (CAN-2002-0655)
2337 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2339 *) Remote buffer overflow in SSL3 protocol - an attacker could
2340 supply an oversized session ID to a client. (CAN-2002-0656)
2341 [Ben Laurie (CHATS)]
2343 *) Remote buffer overflow in SSL2 protocol - an attacker could
2344 supply an oversized client master key. (CAN-2002-0656)
2345 [Ben Laurie (CHATS)]
2347 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2349 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2350 encoded as NULL) with id-dsa-with-sha1.
2351 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2353 *) Check various X509_...() return values in apps/req.c.
2354 [Nils Larsch <nla@trustcenter.de>]
2356 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2357 an end-of-file condition would erronously be flagged, when the CRLF
2358 was just at the end of a processed block. The bug was discovered when
2359 processing data through a buffering memory BIO handing the data to a
2360 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2361 <ptsekov@syntrex.com> and Nedelcho Stanev.
2364 *) Implement a countermeasure against a vulnerability recently found
2365 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2366 before application data chunks to avoid the use of known IVs
2367 with data potentially chosen by the attacker.
2370 *) Fix length checks in ssl3_get_client_hello().
2373 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2374 to prevent ssl3_read_internal() from incorrectly assuming that
2375 ssl3_read_bytes() found application data while handshake
2376 processing was enabled when in fact s->s3->in_read_app_data was
2377 merely automatically cleared during the initial handshake.
2378 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2380 *) Fix object definitions for Private and Enterprise: they were not
2381 recognized in their shortname (=lowercase) representation. Extend
2382 obj_dat.pl to issue an error when using undefined keywords instead
2383 of silently ignoring the problem (Svenning Sorensen
2384 <sss@sss.dnsalias.net>).
2387 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2388 generators, i.e. generators other than 2 and 5. (Previously, the
2389 code did not properly initialise the 'add' and 'rem' values to
2390 BN_generate_prime().)
2392 In the new general case, we do not insist that 'generator' is
2393 actually a primitive root: This requirement is rather pointless;
2394 a generator of the order-q subgroup is just as good, if not
2398 *) Map new X509 verification errors to alerts. Discovered and submitted by
2399 Tom Wu <tom@arcot.com>.
2402 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2403 returning non-zero before the data has been completely received
2404 when using non-blocking I/O.
2405 [Bodo Moeller; problem pointed out by John Hughes]
2407 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2408 [Ben Laurie, Lutz Jaenicke]
2410 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2411 Yoram Zahavi <YoramZ@gilian.com>).
2414 *) Add information about CygWin 1.3 and on, and preserve proper
2415 configuration for the versions before that.
2416 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2418 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2419 check whether we deal with a copy of a session and do not delete from
2420 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2421 <izhar@checkpoint.com>.
2424 *) Do not store session data into the internal session cache, if it
2425 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2426 flag is set). Proposed by Aslam <aslam@funk.com>.
2429 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2433 *) [In 0.9.6d-engine release:]
2434 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2435 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2437 *) Add the configuration target linux-s390x.
2438 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2440 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2441 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2442 variable as an indication that a ClientHello message has been
2443 received. As the flag value will be lost between multiple
2444 invocations of ssl3_accept when using non-blocking I/O, the
2445 function may not be aware that a handshake has actually taken
2446 place, thus preventing a new session from being added to the
2449 To avoid this problem, we now set s->new_session to 2 instead of
2450 using a local variable.
2451 [Lutz Jaenicke, Bodo Moeller]
2453 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
2454 if the SSL_R_LENGTH_MISMATCH error is detected.
2455 [Geoff Thorpe, Bodo Moeller]
2457 *) New 'shared_ldflag' column in Configure platform table.
2460 *) Fix EVP_CIPHER_mode macro.
2461 ["Dan S. Camper" <dan@bti.net>]
2463 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
2464 type, we must throw them away by setting rr->length to 0.
2465 [D P Chang <dpc@qualys.com>]
2467 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
2469 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
2470 <Dominikus.Scherkl@biodata.com>. (The previous implementation
2471 worked incorrectly for those cases where range = 10..._2 and
2472 3*range is two bits longer than range.)
2475 *) Only add signing time to PKCS7 structures if it is not already
2479 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
2480 OBJ_ld_ce should be OBJ_id_ce.
2481 Also some ip-pda OIDs in crypto/objects/objects.txt were
2482 incorrect (cf. RFC 3039).
2483 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
2485 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
2486 returns early because it has nothing to do.
2487 [Andy Schneider <andy.schneider@bjss.co.uk>]
2489 *) [In 0.9.6c-engine release:]
2490 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
2491 [Andy Schneider <andy.schneider@bjss.co.uk>]
2493 *) [In 0.9.6c-engine release:]
2494 Add support for Cryptographic Appliance's keyserver technology.
2495 (Use engine 'keyclient')
2496 [Cryptographic Appliances and Geoff Thorpe]
2498 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
2499 is called via tools/c89.sh because arguments have to be
2500 rearranged (all '-L' options must appear before the first object
2502 [Richard Shapiro <rshapiro@abinitio.com>]
2504 *) [In 0.9.6c-engine release:]
2505 Add support for Broadcom crypto accelerator cards, backported
2507 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
2509 *) [In 0.9.6c-engine release:]
2510 Add support for SureWare crypto accelerator cards from
2511 Baltimore Technologies. (Use engine 'sureware')
2512 [Baltimore Technologies and Mark Cox]
2514 *) [In 0.9.6c-engine release:]
2515 Add support for crypto accelerator cards from Accelerated
2516 Encryption Processing, www.aep.ie. (Use engine 'aep')
2517 [AEP Inc. and Mark Cox]
2519 *) Add a configuration entry for gcc on UnixWare.
2520 [Gary Benson <gbenson@redhat.com>]
2522 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
2523 messages are stored in a single piece (fixed-length part and
2524 variable-length part combined) and fix various bugs found on the way.
2527 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
2528 instead. BIO_gethostbyname() does not know what timeouts are
2529 appropriate, so entries would stay in cache even when they have
2531 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
2533 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
2534 faced with a pathologically small ClientHello fragment that does
2535 not contain client_version: Instead of aborting with an error,
2536 simply choose the highest available protocol version (i.e.,
2537 TLS 1.0 unless it is disabled). In practice, ClientHello
2538 messages are never sent like this, but this change gives us
2539 strictly correct behaviour at least for TLS.
2542 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
2543 never resets s->method to s->ctx->method when called from within
2544 one of the SSL handshake functions.
2545 [Bodo Moeller; problem pointed out by Niko Baric]
2547 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
2548 (sent using the client's version number) if client_version is
2549 smaller than the protocol version in use. Also change
2550 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
2551 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
2552 the client will at least see that alert.
2555 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
2559 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
2560 client receives HelloRequest while in a handshake.
2561 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2563 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2564 should end in 'break', not 'goto end' which circuments various
2565 cleanups done in state SSL_ST_OK. But session related stuff
2566 must be disabled for SSL_ST_OK in the case that we just sent a
2569 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2570 before just sending a HelloRequest.
2571 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2573 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2574 reveal whether illegal block cipher padding was found or a MAC
2575 verification error occured. (Neither SSLerr() codes nor alerts
2576 are directly visible to potential attackers, but the information
2577 may leak via logfiles.)
2579 Similar changes are not required for the SSL 2.0 implementation
2580 because the number of padding bytes is sent in clear for SSL 2.0,
2581 and the extra bytes are just ignored. However ssl/s2_pkt.c
2582 failed to verify that the purported number of padding bytes is in
2586 *) Add OpenUNIX-8 support including shared libraries
2587 (Boyd Lynn Gerber <gerberb@zenez.com>).
2590 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2591 'wristwatch attack' using huge encoding parameters (cf.
2592 James H. Manger's CRYPTO 2001 paper). Note that the
2593 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2594 encoding parameters and hence was not vulnerable.
2597 *) BN_sqr() bug fix.
2598 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2600 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2601 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2602 followed by modular reduction.
2603 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2605 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2606 equivalent based on BN_pseudo_rand() instead of BN_rand().
2609 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2610 This function was broken, as the check for a new client hello message
2611 to handle SGC did not allow these large messages.
2612 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2615 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2618 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2619 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2622 *) Rework the configuration and shared library support for Tru64 Unix.
2623 The configuration part makes use of modern compiler features and
2624 still retains old compiler behavior for those that run older versions
2625 of the OS. The shared library support part includes a variant that
2626 uses the RPATH feature, and is available through the special
2627 configuration target "alpha-cc-rpath", which will never be selected
2629 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2631 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2632 with the same message size as in ssl3_get_certificate_request().
2633 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2634 messages might inadvertently be reject as too long.
2635 [Petr Lampa <lampa@fee.vutbr.cz>]
2637 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2640 *) Modified SSL library such that the verify_callback that has been set
2641 specificly for an SSL object with SSL_set_verify() is actually being
2642 used. Before the change, a verify_callback set with this function was
2643 ignored and the verify_callback() set in the SSL_CTX at the time of
2644 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2645 to allow the necessary settings.
2648 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2649 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2650 done automatically (in contradiction to the requirements of the C
2651 standard). This made problems when used from OpenSSH.
2654 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2655 dh->length and always used
2657 BN_rand_range(priv_key, dh->p).
2659 BN_rand_range() is not necessary for Diffie-Hellman, and this
2660 specific range makes Diffie-Hellman unnecessarily inefficient if
2661 dh->length (recommended exponent length) is much smaller than the
2662 length of dh->p. We could use BN_rand_range() if the order of
2663 the subgroup was stored in the DH structure, but we only have
2668 BN_rand(priv_key, l, ...)
2670 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2676 RSA_eay_public_encrypt
2677 RSA_eay_private_decrypt
2678 RSA_eay_private_encrypt (signing)
2679 RSA_eay_public_decrypt (signature verification)
2681 (default implementations for RSA_public_encrypt,
2682 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2683 always reject numbers >= n.
2686 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2687 to synchronize access to 'locking_thread'. This is necessary on
2688 systems where access to 'locking_thread' (an 'unsigned long'
2689 variable) is not atomic.
2692 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2693 *before* setting the 'crypto_lock_rand' flag. The previous code had
2694 a race condition if 0 is a valid thread ID.
2695 [Travis Vitek <vitek@roguewave.com>]
2697 *) Add support for shared libraries under Irix.
2698 [Albert Chin-A-Young <china@thewrittenword.com>]
2700 *) Add configuration option to build on Linux on both big-endian and
2702 [Ralf Baechle <ralf@uni-koblenz.de>]
2704 *) Add the possibility to create shared libraries on HP-UX.
2707 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2709 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2710 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2711 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2712 PRNG state recovery was possible based on the output of
2713 one PRNG request appropriately sized to gain knowledge on
2714 'md' followed by enough consecutive 1-byte PRNG requests
2715 to traverse all of 'state'.
2717 1. When updating 'md_local' (the current thread's copy of 'md')
2718 during PRNG output generation, hash all of the previous
2719 'md_local' value, not just the half used for PRNG output.
2721 2. Make the number of bytes from 'state' included into the hash
2722 independent from the number of PRNG bytes requested.
2724 The first measure alone would be sufficient to avoid
2725 Markku-Juhani's attack. (Actually it had never occurred
2726 to me that the half of 'md_local' used for chaining was the
2727 half from which PRNG output bytes were taken -- I had always
2728 assumed that the secret half would be used.) The second
2729 measure makes sure that additional data from 'state' is never
2730 mixed into 'md_local' in small portions; this heuristically
2731 further strengthens the PRNG.
2734 *) Fix crypto/bn/asm/mips3.s.
2737 *) When only the key is given to "enc", the IV is undefined. Print out
2738 an error message in this case.
2741 *) Handle special case when X509_NAME is empty in X509 printing routines.
2744 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2745 positive and less than q.
2748 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2749 used: it isn't thread safe and the add_lock_callback should handle
2751 [Paul Rose <Paul.Rose@bridge.com>]
2753 *) Verify that incoming data obeys the block size in
2754 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2758 [Ulf Möller, Bodo Möller]
2760 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2761 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2762 when fixing the server behaviour for backwards-compatible 'client
2763 hello' messages. (Note that the attack is impractical against
2764 SSL 3.0 and TLS 1.0 anyway because length and version checking
2765 means that the probability of guessing a valid ciphertext is
2766 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2769 Before 0.9.5, the countermeasure (hide the error by generating a
2770 random 'decryption result') did not work properly because
2771 ERR_clear_error() was missing, meaning that SSL_get_error() would
2772 detect the supposedly ignored error.
2774 Both problems are now fixed.
2777 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2778 (previously it was 1024).
2781 *) Fix for compatibility mode trust settings: ignore trust settings
2782 unless some valid trust or reject settings are present.
2785 *) Fix for blowfish EVP: its a variable length cipher.
2788 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2789 parameters in DSA public key structures and return an error in the
2790 DSA routines if parameters are absent.
2793 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2794 in the current directory if neither $RANDFILE nor $HOME was set.
2795 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2796 caused some confusion to Windows users who haven't defined $HOME.
2797 Thus RAND_file_name() is changed again: e_os.h can define a
2798 DEFAULT_HOME, which will be used if $HOME is not set.
2799 For Windows, we use "C:"; on other platforms, we still require
2800 environment variables.
2802 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2803 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2804 having multiple threads call RAND_poll() concurrently.
2807 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2808 combination of a flag and a thread ID variable.
2809 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2810 flag), *other* threads can enter ssleay_add_bytes without obeying
2811 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2812 that they do not hold after the first thread unsets add_do_not_lock).
2815 *) Change bctest again: '-x' expressions are not available in all
2819 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2821 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2822 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2824 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2825 the default extension for executables, if any. Also, make the perl
2826 scripts that use symlink() to test if it really exists and use "cp"
2827 if it doesn't. All this made OpenSSL compilable and installable in
2831 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2832 If SEQUENCE is length is indefinite just set c->slen to the total
2833 amount of data available.
2834 [Steve Henson, reported by shige@FreeBSD.org]
2835 [This change does not apply to 0.9.7.]
2837 *) Change bctest to avoid here-documents inside command substitution
2838 (workaround for FreeBSD /bin/sh bug).
2839 For compatibility with Ultrix, avoid shell functions (introduced
2840 in the bctest version that searches along $PATH).
2843 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2844 with des_encrypt() defined on some operating systems, like Solaris
2848 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2849 On the Importance of Eliminating Errors in Cryptographic
2850 Computations, J. Cryptology 14 (2001) 2, 101-119,
2851 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2854 *) MIPS assembler BIGNUM division bug fix.
2857 *) Disabled incorrect Alpha assembler code.
2860 *) Fix PKCS#7 decode routines so they correctly update the length
2861 after reading an EOC for the EXPLICIT tag.
2863 [This change does not apply to 0.9.7.]
2865 *) Fix bug in PKCS#12 key generation routines. This was triggered
2866 if a 3DES key was generated with a 0 initial byte. Include
2867 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2868 (but broken) behaviour.
2871 *) Enhance bctest to search for a working bc along $PATH and print
2873 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2875 *) Fix memory leaks in err.c: free err_data string if necessary;
2876 don't write to the wrong index in ERR_set_error_data.
2879 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2883 *) Replace rdtsc with _emit statements for VC++ version 5.
2884 [Jeremy Cooper <jeremy@baymoo.org>]
2886 *) Make it possible to reuse SSLv2 sessions.
2889 *) In copy_email() check for >= 0 as a return value for
2890 X509_NAME_get_index_by_NID() since 0 is a valid index.
2891 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2893 *) Avoid coredump with unsupported or invalid public keys by checking if
2894 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2895 PKCS7_verify() fails with non detached data.
2898 *) Don't use getenv in library functions when run as setuid/setgid.
2899 New function OPENSSL_issetugid().
2902 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2903 due to incorrect handling of multi-threading:
2905 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2907 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2909 3. Count how many times MemCheck_off() has been called so that
2910 nested use can be treated correctly. This also avoids
2911 inband-signalling in the previous code (which relied on the
2912 assumption that thread ID 0 is impossible).
2915 *) Add "-rand" option also to s_client and s_server.
2918 *) Fix CPU detection on Irix 6.x.
2919 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2920 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2922 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2925 [This change does not apply to 0.9.7.]
2927 *) Use the cached encoding of an X509_NAME structure rather than
2928 copying it. This is apparently the reason for the libsafe "errors"
2929 but the code is actually correct.
2932 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2933 Bleichenbacher's DSA attack.
2934 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2935 to be set and top=0 forces the highest bit to be set; top=-1 is new
2936 and leaves the highest bit random.
2937 [Ulf Moeller, Bodo Moeller]
2939 *) In the NCONF_...-based implementations for CONF_... queries
2940 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2941 a temporary CONF structure with the data component set to NULL
2942 (which gives segmentation faults in lh_retrieve).
2943 Instead, use NULL for the CONF pointer in CONF_get_string and
2944 CONF_get_number (which may use environment variables) and directly
2945 return NULL from CONF_get_section.
2948 *) Fix potential buffer overrun for EBCDIC.
2951 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2952 keyUsage if basicConstraints absent for a CA.
2955 *) Make SMIME_write_PKCS7() write mail header values with a format that
2956 is more generally accepted (no spaces before the semicolon), since
2957 some programs can't parse those values properly otherwise. Also make
2958 sure BIO's that break lines after each write do not create invalid
2962 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2963 macros previously used would not encode an empty SEQUENCE OF
2964 and break the signature.
2966 [This change does not apply to 0.9.7.]
2968 *) Zero the premaster secret after deriving the master secret in
2972 *) Add some EVP_add_digest_alias registrations (as found in
2973 OpenSSL_add_all_digests()) to SSL_library_init()
2974 aka OpenSSL_add_ssl_algorithms(). This provides improved
2975 compatibility with peers using X.509 certificates
2976 with unconventional AlgorithmIdentifier OIDs.
2979 *) Fix for Irix with NO_ASM.
2980 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2982 *) ./config script fixes.
2983 [Ulf Moeller, Richard Levitte]
2985 *) Fix 'openssl passwd -1'.
2988 *) Change PKCS12_key_gen_asc() so it can cope with non null
2989 terminated strings whose length is passed in the passlen
2990 parameter, for example from PEM callbacks. This was done
2991 by adding an extra length parameter to asc2uni().
2992 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2994 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2995 call failed, free the DSA structure.
2998 *) Fix to uni2asc() to cope with zero length Unicode strings.
2999 These are present in some PKCS#12 files.
3002 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3003 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3004 when writing a 32767 byte record.
3005 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3007 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3008 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3010 (RSA objects have a reference count access to which is protected
3011 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3012 so they are meant to be shared between threads.)
3013 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3014 "Reddie, Steven" <Steven.Reddie@ca.com>]
3016 *) Fix a deadlock in CRYPTO_mem_leaks().
3019 *) Use better test patterns in bntest.
3022 *) rand_win.c fix for Borland C.
3025 *) BN_rshift bugfix for n == 0.
3028 *) Add a 'bctest' script that checks for some known 'bc' bugs
3029 so that 'make test' does not abort just because 'bc' is broken.
3032 *) Store verify_result within SSL_SESSION also for client side to
3033 avoid potential security hole. (Re-used sessions on the client side
3034 always resulted in verify_result==X509_V_OK, not using the original
3035 result of the server certificate verification.)
3038 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3039 SSL3_RT_APPLICATION_DATA, return 0.
3040 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3044 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3045 releases, have been re-implemented by renaming the previous
3046 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3047 and ssl3_read_internal, respectively, and adding 'peek' parameters
3048 to them. The new ssl[23]_{read,peek} functions are calls to
3049 ssl[23]_read_internal with the 'peek' flag set appropriately.
3050 A 'peek' parameter has also been added to ssl3_read_bytes, which
3051 does the actual work for ssl3_read_internal.
3054 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3055 the method-specific "init()" handler. Also clean up ex_data after
3056 calling the method-specific "finish()" handler. Previously, this was
3057 happening the other way round.
3060 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3061 The previous value, 12, was not always sufficient for BN_mod_exp().
3064 *) Make sure that shared libraries get the internal name engine with
3065 the full version number and not just 0. This should mark the
3066 shared libraries as not backward compatible. Of course, this should
3067 be changed again when we can guarantee backward binary compatibility.
3070 *) Fix typo in get_cert_by_subject() in by_dir.c
3071 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3073 *) Rework the system to generate shared libraries:
3075 - Make note of the expected extension for the shared libraries and
3076 if there is a need for symbolic links from for example libcrypto.so.0
3077 to libcrypto.so.0.9.7. There is extended info in Configure for
3080 - Make as few rebuilds of the shared libraries as possible.
3082 - Still avoid linking the OpenSSL programs with the shared libraries.
3084 - When installing, install the shared libraries separately from the
3088 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3090 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3091 and not in SSL_clear because the latter is also used by the
3092 accept/connect functions; previously, the settings made by
3093 SSL_set_read_ahead would be lost during the handshake.
3094 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3096 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3097 Previously, it would create entries for disableed algorithms no
3101 *) Added several new manual pages for SSL_* function.
3104 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3106 *) In ssl23_get_client_hello, generate an error message when faced
3107 with an initial SSL 3.0/TLS record that is too small to contain the
3108 first two bytes of the ClientHello message, i.e. client_version.
3109 (Note that this is a pathologic case that probably has never happened
3110 in real life.) The previous approach was to use the version number
3111 from the record header as a substitute; but our protocol choice
3112 should not depend on that one because it is not authenticated
3113 by the Finished messages.
3116 *) More robust randomness gathering functions for Windows.
3117 [Jeffrey Altman <jaltman@columbia.edu>]
3119 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3120 not set then we don't setup the error code for issuer check errors
3121 to avoid possibly overwriting other errors which the callback does
3122 handle. If an application does set the flag then we assume it knows
3123 what it is doing and can handle the new informational codes
3127 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3128 a general "ANY" type, as such it should be able to decode anything
3129 including tagged types. However it didn't check the class so it would
3130 wrongly interpret tagged types in the same way as their universal
3131 counterpart and unknown types were just rejected. Changed so that the
3132 tagged and unknown types are handled in the same way as a SEQUENCE:
3133 that is the encoding is stored intact. There is also a new type
3134 "V_ASN1_OTHER" which is used when the class is not universal, in this
3135 case we have no idea what the actual type is so we just lump them all
3139 *) On VMS, stdout may very well lead to a file that is written to
3140 in a record-oriented fashion. That means that every write() will
3141 write a separate record, which will be read separately by the
3142 programs trying to read from it. This can be very confusing.
3144 The solution is to put a BIO filter in the way that will buffer
3145 text until a linefeed is reached, and then write everything a
3146 line at a time, so every record written will be an actual line,
3147 not chunks of lines and not (usually doesn't happen, but I've
3148 seen it once) several lines in one record. BIO_f_linebuffer() is
3151 Currently, it's a VMS-only method, because that's where it has
3152 been tested well enough.
3155 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3156 it can return incorrect results.
3157 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3158 but it was in 0.9.6-beta[12].)
3161 *) Disable the check for content being present when verifying detached
3162 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3163 include zero length content when signing messages.
3166 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3167 BIO_ctrl (for BIO pairs).
3170 *) Add DSO method for VMS.
3173 *) Bug fix: Montgomery multiplication could produce results with the
3177 *) Add RPM specification openssl.spec and modify it to build three
3178 packages. The default package contains applications, application
3179 documentation and run-time libraries. The devel package contains
3180 include files, static libraries and function documentation. The
3181 doc package contains the contents of the doc directory. The original
3182 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3185 *) Add a large number of documentation files for many SSL routines.
3186 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3188 *) Add a configuration entry for Sony News 4.
3189 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3191 *) Don't set the two most significant bits to one when generating a
3192 random number < q in the DSA library.
3195 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3196 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3197 the underlying transport is blocking) if a handshake took place.
3198 (The default behaviour is needed by applications such as s_client
3199 and s_server that use select() to determine when to use SSL_read;
3200 but for applications that know in advance when to expect data, it
3201 just makes things more complicated.)
3204 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3208 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3209 work better on such systems.
3210 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3212 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3213 Update PKCS12_parse() so it copies the friendlyName and the
3214 keyid to the certificates aux info.
3217 *) Fix bug in PKCS7_verify() which caused an infinite loop
3218 if there was more than one signature.
3219 [Sven Uszpelkat <su@celocom.de>]
3221 *) Major change in util/mkdef.pl to include extra information
3222 about each symbol, as well as presentig variables as well
3223 as functions. This change means that there's n more need
3224 to rebuild the .num files when some algorithms are excluded.
3227 *) Allow the verify time to be set by an application,
3228 rather than always using the current time.
3231 *) Phase 2 verify code reorganisation. The certificate
3232 verify code now looks up an issuer certificate by a
3233 number of criteria: subject name, authority key id
3234 and key usage. It also verifies self signed certificates
3235 by the same criteria. The main comparison function is
3236 X509_check_issued() which performs these checks.
3238 Lot of changes were necessary in order to support this
3239 without completely rewriting the lookup code.
3241 Authority and subject key identifier are now cached.
3243 The LHASH 'certs' is X509_STORE has now been replaced
3244 by a STACK_OF(X509_OBJECT). This is mainly because an
3245 LHASH can't store or retrieve multiple objects with
3246 the same hash value.
3248 As a result various functions (which were all internal
3249 use only) have changed to handle the new X509_STORE
3250 structure. This will break anything that messed round
3251 with X509_STORE internally.
3253 The functions X509_STORE_add_cert() now checks for an
3254 exact match, rather than just subject name.
3256 The X509_STORE API doesn't directly support the retrieval
3257 of multiple certificates matching a given criteria, however
3258 this can be worked round by performing a lookup first
3259 (which will fill the cache with candidate certificates)
3260 and then examining the cache for matches. This is probably
3261 the best we can do without throwing out X509_LOOKUP
3262 entirely (maybe later...).
3264 The X509_VERIFY_CTX structure has been enhanced considerably.
3266 All certificate lookup operations now go via a get_issuer()
3267 callback. Although this currently uses an X509_STORE it
3268 can be replaced by custom lookups. This is a simple way
3269 to bypass the X509_STORE hackery necessary to make this
3270 work and makes it possible to use more efficient techniques
3271 in future. A very simple version which uses a simple
3272 STACK for its trusted certificate store is also provided
3273 using X509_STORE_CTX_trusted_stack().
3275 The verify_cb() and verify() callbacks now have equivalents
3276 in the X509_STORE_CTX structure.
3278 X509_STORE_CTX also has a 'flags' field which can be used
3279 to customise the verify behaviour.
3282 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3283 excludes S/MIME capabilities.
3286 *) When a certificate request is read in keep a copy of the
3287 original encoding of the signed data and use it when outputing
3288 again. Signatures then use the original encoding rather than
3289 a decoded, encoded version which may cause problems if the
3290 request is improperly encoded.
3293 *) For consistency with other BIO_puts implementations, call
3294 buffer_write(b, ...) directly in buffer_puts instead of calling
3297 In BIO_puts, increment b->num_write as in BIO_write.
3298 [Peter.Sylvester@EdelWeb.fr]
3300 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3301 BN_zero, we may not return a BIGNUM with an array consisting of
3305 *) Avoid calling abort() from within the library when problems are
3306 detected, except if preprocessor symbols have been defined
3307 (such as REF_CHECK, BN_DEBUG etc.).
3310 *) New openssl application 'rsautl'. This utility can be
3311 used for low level RSA operations. DER public key
3312 BIO/fp routines also added.
3315 *) New Configure entry and patches for compiling on QNX 4.
3316 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3318 *) A demo state-machine implementation was sponsored by
3319 Nuron (http://www.nuron.com/) and is now available in
3320 demos/state_machine.
3323 *) New options added to the 'dgst' utility for signature
3324 generation and verification.
3327 *) Unrecognized PKCS#7 content types are now handled via a
3328 catch all ASN1_TYPE structure. This allows unsupported
3329 types to be stored as a "blob" and an application can
3330 encode and decode it manually.
3333 *) Fix various signed/unsigned issues to make a_strex.c
3335 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3337 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3338 length if passed a buffer. ASN1_INTEGER_to_BN failed
3339 if passed a NULL BN and its argument was negative.
3340 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3342 *) Modification to PKCS#7 encoding routines to output definite
3343 length encoding. Since currently the whole structures are in
3344 memory there's not real point in using indefinite length
3345 constructed encoding. However if OpenSSL is compiled with
3346 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3349 *) Added BIO_vprintf() and BIO_vsnprintf().
3352 *) Added more prefixes to parse for in the the strings written
3353 through a logging bio, to cover all the levels that are available
3354 through syslog. The prefixes are now:
3356 PANIC, EMERG, EMR => LOG_EMERG
3357 ALERT, ALR => LOG_ALERT
3358 CRIT, CRI => LOG_CRIT
3359 ERROR, ERR => LOG_ERR
3360 WARNING, WARN, WAR => LOG_WARNING
3361 NOTICE, NOTE, NOT => LOG_NOTICE
3362 INFO, INF => LOG_INFO
3363 DEBUG, DBG => LOG_DEBUG
3365 and as before, if none of those prefixes are present at the
3366 beginning of the string, LOG_ERR is chosen.
3368 On Win32, the LOG_* levels are mapped according to this:
3370 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3371 LOG_WARNING => EVENTLOG_WARNING_TYPE
3372 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3376 *) Made it possible to reconfigure with just the configuration
3377 argument "reconf" or "reconfigure". The command line arguments
3378 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3379 and are retrieved from there when reconfiguring.
3383 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3385 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3388 *) The obj_dat.pl script was messing up the sorting of object
3389 names. The reason was that it compared the quoted version
3390 of strings as a result "OCSP" > "OCSP Signing" because
3391 " > SPACE. Changed script to store unquoted versions of
3392 names and add quotes on output. It was also omitting some
3393 names from the lookup table if they were given a default
3394 value (that is if SN is missing it is given the same
3395 value as LN and vice versa), these are now added on the
3396 grounds that if an object has a name we should be able to
3397 look it up. Finally added warning output when duplicate
3398 short or long names are found.
3401 *) Changes needed for Tandem NSK.
3402 [Scott Uroff <scott@xypro.com>]
3404 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3405 RSA_padding_check_SSLv23(), special padding was never detected
3406 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3407 version rollback attacks was not effective.
3409 In s23_clnt.c, don't use special rollback-attack detection padding
3410 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3411 client; similarly, in s23_srvr.c, don't do the rollback check if
3412 SSL 2.0 is the only protocol enabled in the server.
3415 *) Make it possible to get hexdumps of unprintable data with 'openssl
3416 asn1parse'. By implication, the functions ASN1_parse_dump() and
3417 BIO_dump_indent() are added.
3420 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3421 these print out strings and name structures based on various
3422 flags including RFC2253 support and proper handling of
3423 multibyte characters. Added options to the 'x509' utility
3424 to allow the various flags to be set.
3427 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3428 Also change the functions X509_cmp_current_time() and
3429 X509_gmtime_adj() work with an ASN1_TIME structure,
3430 this will enable certificates using GeneralizedTime in validity
3431 dates to be checked.
3434 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3435 negative public key encodings) on by default,
3436 NO_NEG_PUBKEY_BUG can be set to disable it.
3439 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3440 content octets. An i2c_ASN1_OBJECT is unnecessary because
3441 the encoding can be trivially obtained from the structure.
3444 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3445 not read locks (CRYPTO_r_[un]lock).
3448 *) A first attempt at creating official support for shared
3449 libraries through configuration. I've kept it so the
3450 default is static libraries only, and the OpenSSL programs
3451 are always statically linked for now, but there are
3452 preparations for dynamic linking in place.
3453 This has been tested on Linux and Tru64.
3456 *) Randomness polling function for Win9x, as described in:
3457 Peter Gutmann, Software Generation of Practically Strong
3461 *) Fix so PRNG is seeded in req if using an already existing
3465 *) New options to smime application. -inform and -outform
3466 allow alternative formats for the S/MIME message including
3467 PEM and DER. The -content option allows the content to be
3468 specified separately. This should allow things like Netscape
3469 form signing output easier to verify.
3472 *) Fix the ASN1 encoding of tags using the 'long form'.
3475 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
3476 STRING types. These convert content octets to and from the
3477 underlying type. The actual tag and length octets are
3478 already assumed to have been read in and checked. These
3479 are needed because all other string types have virtually
3480 identical handling apart from the tag. By having versions
3481 of the ASN1 functions that just operate on content octets
3482 IMPLICIT tagging can be handled properly. It also allows
3483 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
3484 and ASN1_INTEGER are identical apart from the tag.
3487 *) Change the handling of OID objects as follows:
3489 - New object identifiers are inserted in objects.txt, following
3490 the syntax given in objects.README.
3491 - objects.pl is used to process obj_mac.num and create a new
3493 - obj_dat.pl is used to create a new obj_dat.h, using the data in
3496 This is currently kind of a hack, and the perl code in objects.pl
3497 isn't very elegant, but it works as I intended. The simplest way
3498 to check that it worked correctly is to look in obj_dat.h and
3499 check the array nid_objs and make sure the objects haven't moved
3500 around (this is important!). Additions are OK, as well as
3501 consistent name changes.
3504 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
3507 *) Addition of the command line parameter '-rand file' to 'openssl req'.
3508 The given file adds to whatever has already been seeded into the
3509 random pool through the RANDFILE configuration file option or
3510 environment variable, or the default random state file.
3513 *) mkstack.pl now sorts each macro group into lexical order.
3514 Previously the output order depended on the order the files
3515 appeared in the directory, resulting in needless rewriting
3519 *) Patches to make OpenSSL compile under Win32 again. Mostly
3520 work arounds for the VC++ problem that it treats func() as
3521 func(void). Also stripped out the parts of mkdef.pl that
3522 added extra typesafe functions: these no longer exist.
3525 *) Reorganisation of the stack code. The macros are now all
3526 collected in safestack.h . Each macro is defined in terms of
3527 a "stack macro" of the form SKM_<name>(type, a, b). The
3528 DEBUG_SAFESTACK is now handled in terms of function casts,
3529 this has the advantage of retaining type safety without the
3530 use of additional functions. If DEBUG_SAFESTACK is not defined
3531 then the non typesafe macros are used instead. Also modified the
3532 mkstack.pl script to handle the new form. Needs testing to see
3533 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
3534 the default if no major problems. Similar behaviour for ASN1_SET_OF
3535 and PKCS12_STACK_OF.
3538 *) When some versions of IIS use the 'NET' form of private key the
3539 key derivation algorithm is different. Normally MD5(password) is
3540 used as a 128 bit RC4 key. In the modified case
3541 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
3542 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
3543 as the old Netscape_RSA functions except they have an additional
3544 'sgckey' parameter which uses the modified algorithm. Also added
3545 an -sgckey command line option to the rsa utility. Thanks to
3546 Adrian Peck <bertie@ncipher.com> for posting details of the modified
3547 algorithm to openssl-dev.
3550 *) The evp_local.h macros were using 'c.##kname' which resulted in
3551 invalid expansion on some systems (SCO 5.0.5 for example).
3552 Corrected to 'c.kname'.
3553 [Phillip Porch <root@theporch.com>]
3555 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
3556 a STACK of email addresses from a certificate or request, these look
3557 in the subject name and the subject alternative name extensions and
3558 omit any duplicate addresses.
3561 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3562 This makes DSA verification about 2 % faster.
3565 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3566 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3567 plus overhead for 1024 bit moduli).
3568 This makes exponentiations about 0.5 % faster for 1024 bit
3569 exponents (as measured by "openssl speed rsa2048").
3572 *) Rename memory handling macros to avoid conflicts with other
3574 Malloc => OPENSSL_malloc
3575 Malloc_locked => OPENSSL_malloc_locked
3576 Realloc => OPENSSL_realloc
3577 Free => OPENSSL_free
3580 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3581 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3584 *) CygWin32 support.
3585 [John Jarvie <jjarvie@newsguy.com>]
3587 *) The type-safe stack code has been rejigged. It is now only compiled
3588 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3589 by default all type-specific stack functions are "#define"d back to
3590 standard stack functions. This results in more streamlined output
3591 but retains the type-safety checking possibilities of the original
3595 *) The STACK code has been cleaned up, and certain type declarations
3596 that didn't make a lot of sense have been brought in line. This has
3597 also involved a cleanup of sorts in safestack.h to more correctly
3598 map type-safe stack functions onto their plain stack counterparts.
3599 This work has also resulted in a variety of "const"ifications of
3600 lots of the code, especially "_cmp" operations which should normally
3601 be prototyped with "const" parameters anyway.
3604 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3605 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3606 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3607 where all of 'md' is used each time the PRNG is used, but 'state'
3608 is used only indexed by a cyclic counter. As entropy may not be
3609 well distributed from the beginning, 'md' is important as a
3610 chaining variable. However, the output function chains only half
3611 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3612 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3613 in all of 'state' being rewritten, with the new values depending
3614 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3617 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3618 the handshake is continued after ssl_verify_cert_chain();
3619 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3620 can lead to 'unexplainable' connection aborts later.
3621 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3623 *) Major EVP API cipher revision.
3624 Add hooks for extra EVP features. This allows various cipher
3625 parameters to be set in the EVP interface. Support added for variable
3626 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3627 setting of RC2 and RC5 parameters.
3629 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3632 Remove lots of duplicated code from the EVP library. For example *every*
3633 cipher init() function handles the 'iv' in the same way according to the
3634 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3635 for CFB and OFB modes they zero ctx->num.
3637 New functionality allows removal of S/MIME code RC2 hack.
3639 Most of the routines have the same form and so can be declared in terms
3642 By shifting this to the top level EVP_CipherInit() it can be removed from
3643 all individual ciphers. If the cipher wants to handle IVs or keys
3644 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3647 Change lots of functions like EVP_EncryptUpdate() to now return a
3648 value: although software versions of the algorithms cannot fail
3649 any installed hardware versions can.
3652 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3653 this option is set, tolerate broken clients that send the negotiated
3654 protocol version number instead of the requested protocol version
3658 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3659 i.e. non-zero for export ciphersuites, zero otherwise.
3660 Previous versions had this flag inverted, inconsistent with
3661 rsa_tmp_cb (..._TMP_RSA_CB).
3662 [Bodo Moeller; problem reported by Amit Chopra]
3664 *) Add missing DSA library text string. Work around for some IIS
3665 key files with invalid SEQUENCE encoding.
3668 *) Add a document (doc/standards.txt) that list all kinds of standards
3669 and so on that are implemented in OpenSSL.
3672 *) Enhance c_rehash script. Old version would mishandle certificates
3673 with the same subject name hash and wouldn't handle CRLs at all.
3674 Added -fingerprint option to crl utility, to support new c_rehash
3678 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3681 *) Fix for SSL server purpose checking. Server checking was
3682 rejecting certificates which had extended key usage present
3683 but no ssl client purpose.
3684 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3686 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3687 is a little unclear about how a blank password is handled.
3688 Since the password in encoded as a BMPString with terminating
3689 double NULL a zero length password would end up as just the
3690 double NULL. However no password at all is different and is
3691 handled differently in the PKCS#12 key generation code. NS
3692 treats a blank password as zero length. MSIE treats it as no
3693 password on export: but it will try both on import. We now do
3694 the same: PKCS12_parse() tries zero length and no password if
3695 the password is set to "" or NULL (NULL is now a valid password:
3696 it wasn't before) as does the pkcs12 application.
3699 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3700 perror when PEM_read_bio_X509_REQ fails, the error message must
3701 be obtained from the error queue.
3704 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3705 it in ERR_remove_state if appropriate, and change ERR_get_state
3706 accordingly to avoid race conditions (this is necessary because
3707 thread_hash is no longer constant once set).
3710 *) Bugfix for linux-elf makefile.one.
3713 *) RSA_get_default_method() will now cause a default
3714 RSA_METHOD to be chosen if one doesn't exist already.
3715 Previously this was only set during a call to RSA_new()
3716 or RSA_new_method(NULL) meaning it was possible for
3717 RSA_get_default_method() to return NULL.
3720 *) Added native name translation to the existing DSO code
3721 that will convert (if the flag to do so is set) filenames
3722 that are sufficiently small and have no path information
3723 into a canonical native form. Eg. "blah" converted to
3724 "libblah.so" or "blah.dll" etc.
3727 *) New function ERR_error_string_n(e, buf, len) which is like
3728 ERR_error_string(e, buf), but writes at most 'len' bytes
3729 including the 0 terminator. For ERR_error_string_n, 'buf'
3731 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3733 *) CONF library reworked to become more general. A new CONF
3734 configuration file reader "class" is implemented as well as a
3735 new functions (NCONF_*, for "New CONF") to handle it. The now
3736 old CONF_* functions are still there, but are reimplemented to
3737 work in terms of the new functions. Also, a set of functions
3738 to handle the internal storage of the configuration data is
3739 provided to make it easier to write new configuration file
3740 reader "classes" (I can definitely see something reading a
3741 configuration file in XML format, for example), called _CONF_*,
3742 or "the configuration storage API"...
3744 The new configuration file reading functions are:
3746 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3747 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3749 NCONF_default, NCONF_WIN32
3751 NCONF_dump_fp, NCONF_dump_bio
3753 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3754 NCONF_new creates a new CONF object. This works in the same way
3755 as other interfaces in OpenSSL, like the BIO interface.
3756 NCONF_dump_* dump the internal storage of the configuration file,
3757 which is useful for debugging. All other functions take the same
3758 arguments as the old CONF_* functions wth the exception of the
3759 first that must be a `CONF *' instead of a `LHASH *'.
3761 To make it easer to use the new classes with the old CONF_* functions,
3762 the function CONF_set_default_method is provided.
3765 *) Add '-tls1' option to 'openssl ciphers', which was already
3766 mentioned in the documentation but had not been implemented.
3767 (This option is not yet really useful because even the additional
3768 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3771 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3772 OpenSSL-based applications) load shared libraries and bind to
3773 them in a portable way.
3774 [Geoff Thorpe, with contributions from Richard Levitte]
3776 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3778 *) Make sure _lrotl and _lrotr are only used with MSVC.
3780 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3781 (the default implementation of RAND_status).
3783 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3784 to '-clrext' (= clear extensions), as intended and documented.
3785 [Bodo Moeller; inconsistency pointed out by Michael Attili
3788 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3789 was larger than the MD block size.
3790 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3792 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3793 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3794 using the passed key: if the passed key was a private key the result
3795 of X509_print(), for example, would be to print out all the private key
3799 *) des_quad_cksum() byte order bug fix.
3800 [Ulf Möller, using the problem description in krb4-0.9.7, where
3801 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3803 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3805 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3807 *) For easily testing in shell scripts whether some command
3808 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3809 returns with exit code 0 iff no command of the given name is available.
3810 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3811 the output goes to stdout and nothing is printed to stderr.
3812 Additional arguments are always ignored.
3814 Since for each cipher there is a command of the same name,
3815 the 'no-cipher' compilation switches can be tested this way.
3817 ('openssl no-XXX' is not able to detect pseudo-commands such
3818 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3821 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3824 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3825 is set; it will be thrown away anyway because each handshake creates
3827 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3828 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3829 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3830 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3833 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3834 'Q' and 'R' lose their special meanings (quit/renegotiate).
3835 This is part of what -quiet does; unlike -quiet, -ign_eof
3836 does not suppress any output.
3839 *) Add compatibility options to the purpose and trust code. The
3840 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3841 accepts a certificate or CA, this was the previous behaviour,
3842 with all the associated security issues.
3844 X509_TRUST_COMPAT is the old trust behaviour: only and
3845 automatically trust self signed roots in certificate store. A
3846 new trust setting X509_TRUST_DEFAULT is used to specify that
3847 a purpose has no associated trust setting and it should instead
3848 use the value in the default purpose.
3851 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3852 and fix a memory leak.
3855 *) In util/mkerr.pl (which implements 'make errors'), preserve
3856 reason strings from the previous version of the .c file, as
3857 the default to have only downcase letters (and digits) in
3858 automatically generated reasons codes is not always appropriate.
3861 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3862 using strerror. Previously, ERR_reason_error_string() returned
3863 library names as reason strings for SYSerr; but SYSerr is a special
3864 case where small numbers are errno values, not library numbers.
3867 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3868 converts DSA parameters into DH parameters. (When creating parameters,
3869 DSA_generate_parameters is used.)
3872 *) Include 'length' (recommended exponent length) in C code generated
3873 by 'openssl dhparam -C'.
3876 *) The second argument to set_label in perlasm was already being used
3877 so couldn't be used as a "file scope" flag. Moved to third argument
3881 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3882 instead of RAND_bytes for encryption IVs and salts.
3885 *) Include RAND_status() into RAND_METHOD instead of implementing
3886 it only for md_rand.c Otherwise replacing the PRNG by calling
3887 RAND_set_rand_method would be impossible.
3890 *) Don't let DSA_generate_key() enter an infinite loop if the random
3891 number generation fails.
3894 *) New 'rand' application for creating pseudo-random output.
3897 *) Added configuration support for Linux/IA64
3898 [Rolf Haberrecker <rolf@suse.de>]
3900 *) Assembler module support for Mingw32.
3903 *) Shared library support for HPUX (in shlib/).
3904 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3906 *) Shared library support for Solaris gcc.
3907 [Lutz Behnke <behnke@trustcenter.de>]
3909 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3911 *) PKCS7_encrypt() was adding text MIME headers twice because they
3912 were added manually and by SMIME_crlf_copy().
3915 *) In bntest.c don't call BN_rand with zero bits argument.
3916 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3918 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3919 case was implemented. This caused BN_div_recp() to fail occasionally.
3922 *) Add an optional second argument to the set_label() in the perl
3923 assembly language builder. If this argument exists and is set
3924 to 1 it signals that the assembler should use a symbol whose
3925 scope is the entire file, not just the current function. This
3926 is needed with MASM which uses the format label:: for this scope.
3927 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3929 *) Change the ASN1 types so they are typedefs by default. Before
3930 almost all types were #define'd to ASN1_STRING which was causing
3931 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3935 *) Change names of new functions to the new get1/get0 naming
3936 convention: After 'get1', the caller owns a reference count
3937 and has to call ..._free; 'get0' returns a pointer to some
3938 data structure without incrementing reference counters.
3939 (Some of the existing 'get' functions increment a reference
3940 counter, some don't.)
3941 Similarly, 'set1' and 'add1' functions increase reference
3942 counters or duplicate objects.
3945 *) Allow for the possibility of temp RSA key generation failure:
3946 the code used to assume it always worked and crashed on failure.
3949 *) Fix potential buffer overrun problem in BIO_printf().
3950 [Ulf Möller, using public domain code by Patrick Powell; problem
3951 pointed out by David Sacerdote <das33@cornell.edu>]
3953 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3954 RAND_egd() and RAND_status(). In the command line application,
3955 the EGD socket can be specified like a seed file using RANDFILE
3959 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3960 Some CAs (e.g. Verisign) distribute certificates in this form.
3963 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3964 list to exclude them. This means that no special compilation option
3965 is needed to use anonymous DH: it just needs to be included in the
3969 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3970 EVP_MD_type. The old functionality is available in a new macro called
3971 EVP_MD_md(). Change code that uses it and update docs.
3974 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3975 where the 'void *' argument is replaced by a function pointer argument.
3976 Previously 'void *' was abused to point to functions, which works on
3977 many platforms, but is not correct. As these functions are usually
3978 called by macros defined in OpenSSL header files, most source code
3979 should work without changes.
3982 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3983 sections with information on -D... compiler switches used for
3984 compiling the library so that applications can see them. To enable
3985 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3986 must be defined. E.g.,
3987 #define OPENSSL_ALGORITHM_DEFINES
3988 #include <openssl/opensslconf.h>
3989 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3990 [Richard Levitte, Ulf and Bodo Möller]
3992 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3996 *) Change the 'other' type in certificate aux info to a STACK_OF
3997 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3998 the required ASN1 format: arbitrary types determined by an OID.
4001 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4002 argument to 'req'. This is not because the function is newer or
4003 better than others it just uses the work 'NEW' in the certificate
4004 request header lines. Some software needs this.
4007 *) Reorganise password command line arguments: now passwords can be
4008 obtained from various sources. Delete the PEM_cb function and make
4009 it the default behaviour: i.e. if the callback is NULL and the
4010 usrdata argument is not NULL interpret it as a null terminated pass
4011 phrase. If usrdata and the callback are NULL then the pass phrase
4012 is prompted for as usual.
4015 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4016 the support is automatically enabled. The resulting binaries will
4017 autodetect the card and use it if present.
4018 [Ben Laurie and Compaq Inc.]
4020 *) Work around for Netscape hang bug. This sends certificate request
4021 and server done in one record. Since this is perfectly legal in the
4022 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4023 the bugs/SSLv3 entry for more info.
4026 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4029 *) Add -rand argument to smime and pkcs12 applications and read/write
4033 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4036 *) Add command line password options to the remaining applications.
4039 *) Bug fix for BN_div_recp() for numerators with an even number of
4043 *) More tests in bntest.c, and changed test_bn output.
4046 *) ./config recognizes MacOS X now.
4049 *) Bug fix for BN_div() when the first words of num and divsor are
4050 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4053 *) Add support for various broken PKCS#8 formats, and command line
4054 options to produce them.
4057 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4058 get temporary BIGNUMs from a BN_CTX.
4061 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4065 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4066 include a #define from the old name to the new. The original intent
4067 was that statically linked binaries could for example just call
4068 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4069 link with digests. This never worked becayse SSLeay_add_all_digests()
4070 and SSLeay_add_all_ciphers() were in the same source file so calling
4071 one would link with the other. They are now in separate source files.
4074 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4077 *) Use a less unusual form of the Miller-Rabin primality test (it used
4078 a binary algorithm for exponentiation integrated into the Miller-Rabin
4079 loop, our standard modexp algorithms are faster).
4082 *) Support for the EBCDIC character set completed.
4083 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4085 *) Source code cleanups: use const where appropriate, eliminate casts,
4086 use void * instead of char * in lhash.
4089 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4090 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4091 this the server could overwrite ephemeral keys that the client
4095 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4096 using 50 iterations of the Rabin-Miller test.
4098 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4099 iterations of the Rabin-Miller test as required by the appendix
4100 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4101 As BN_is_prime_fasttest includes trial division, DSA parameter
4102 generation becomes much faster.
4104 This implies a change for the callback functions in DSA_is_prime
4105 and DSA_generate_parameters: The callback function is called once
4106 for each positive witness in the Rabin-Miller test, not just
4107 occasionally in the inner loop; and the parameters to the
4108 callback function now provide an iteration count for the outer
4109 loop rather than for the current invocation of the inner loop.
4110 DSA_generate_parameters additionally can call the callback
4111 function with an 'iteration count' of -1, meaning that a
4112 candidate has passed the trial division test (when q is generated
4113 from an application-provided seed, trial division is skipped).
4116 *) New function BN_is_prime_fasttest that optionally does trial
4117 division before starting the Rabin-Miller test and has
4118 an additional BN_CTX * argument (whereas BN_is_prime always
4119 has to allocate at least one BN_CTX).
4120 'callback(1, -1, cb_arg)' is called when a number has passed the
4121 trial division stage.
4124 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4128 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4131 *) New function BN_pseudo_rand().
4134 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4135 bignum version of BN_from_montgomery() with the working code from
4136 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4140 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4141 made it impossible to use the same SSL_SESSION data structure in
4142 SSL2 clients in multiple threads.
4145 *) The return value of RAND_load_file() no longer counts bytes obtained
4146 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4147 to seed the PRNG (previously an explicit byte count was required).
4148 [Ulf Möller, Bodo Möller]
4150 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4151 used (char *) instead of (void *) and had casts all over the place.
4154 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4157 *) Retain source code compatibility for BN_prime_checks macro:
4158 BN_is_prime(..., BN_prime_checks, ...) now uses
4159 BN_prime_checks_for_size to determine the appropriate number of
4160 Rabin-Miller iterations.
4163 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4164 DH_CHECK_P_NOT_SAFE_PRIME.
4165 (Check if this is true? OpenPGP calls them "strong".)
4168 *) Merge the functionality of "dh" and "gendh" programs into a new program
4169 "dhparam". The old programs are retained for now but will handle DH keys
4170 (instead of parameters) in future.
4173 *) Make the ciphers, s_server and s_client programs check the return values
4174 when a new cipher list is set.
4177 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4178 ciphers. Before when the 56bit ciphers were enabled the sorting was
4181 The syntax for the cipher sorting has been extended to support sorting by
4182 cipher-strength (using the strength_bits hard coded in the tables).
4183 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4185 Fix a bug in the cipher-command parser: when supplying a cipher command
4186 string with an "undefined" symbol (neither command nor alphanumeric
4187 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4188 an error is flagged.
4190 Due to the strength-sorting extension, the code of the
4191 ssl_create_cipher_list() function was completely rearranged. I hope that
4192 the readability was also increased :-)
4193 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4195 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4196 for the first serial number and places 2 in the serial number file. This
4197 avoids problems when the root CA is created with serial number zero and
4198 the first user certificate has the same issuer name and serial number
4202 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4203 the new code. Add documentation for this stuff.
4206 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4207 X509_*() to X509at_*() on the grounds that they don't handle X509
4208 structures and behave in an analagous way to the X509v3 functions:
4209 they shouldn't be called directly but wrapper functions should be used
4212 So we also now have some wrapper functions that call the X509at functions
4213 when passed certificate requests. (TO DO: similar things can be done with
4214 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4215 things. Some of these need some d2i or i2d and print functionality
4216 because they handle more complex structures.)
4219 *) Add missing #ifndefs that caused missing symbols when building libssl
4220 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4221 NO_RSA in ssl/s2*.c.
4222 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4224 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4225 has a return value which indicates the quality of the random data
4226 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4227 error queue. New function RAND_pseudo_bytes() generates output that is
4228 guaranteed to be unique but not unpredictable. RAND_add is like
4229 RAND_seed, but takes an extra argument for an entropy estimate
4230 (RAND_seed always assumes full entropy).
4233 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4234 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4235 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4236 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4237 false-positive rate of at most 2^-80 for random input.
4240 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4243 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4244 in the 0.9.5 release), this returns the chain
4245 from an X509_CTX structure with a dup of the stack and all
4246 the X509 reference counts upped: so the stack will exist
4247 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4250 Also make SSL_SESSION_print() print out the verify return
4254 *) Add manpage for the pkcs12 command. Also change the default
4255 behaviour so MAC iteration counts are used unless the new
4256 -nomaciter option is used. This improves file security and
4257 only older versions of MSIE (4.0 for example) need it.
4260 *) Honor the no-xxx Configure options when creating .DEF files.
4263 *) Add PKCS#10 attributes to field table: challengePassword,
4264 unstructuredName and unstructuredAddress. These are taken from
4265 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4266 international characters are used.
4268 More changes to X509_ATTRIBUTE code: allow the setting of types
4269 based on strings. Remove the 'loc' parameter when adding
4270 attributes because these will be a SET OF encoding which is sorted
4274 *) Initial changes to the 'req' utility to allow request generation
4275 automation. This will allow an application to just generate a template
4276 file containing all the field values and have req construct the
4279 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4280 used all over the place including certificate requests and PKCS#7
4281 structures. They are currently handled manually where necessary with
4282 some primitive wrappers for PKCS#7. The new functions behave in a
4283 manner analogous to the X509 extension functions: they allow
4284 attributes to be looked up by NID and added.
4286 Later something similar to the X509V3 code would be desirable to
4287 automatically handle the encoding, decoding and printing of the
4288 more complex types. The string types like challengePassword can
4289 be handled by the string table functions.
4291 Also modified the multi byte string table handling. Now there is
4292 a 'global mask' which masks out certain types. The table itself
4293 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4294 is useful when for example there is only one permissible type
4295 (as in countryName) and using the mask might result in no valid
4299 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4300 SSL_get_peer_finished to allow applications to obtain the latest
4301 Finished messages sent to the peer or expected from the peer,
4302 respectively. (SSL_get_peer_finished is usually the Finished message
4303 actually received from the peer, otherwise the protocol will be aborted.)
4305 As the Finished message are message digests of the complete handshake
4306 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4307 be used for external authentication procedures when the authentication
4308 provided by SSL/TLS is not desired or is not enough.
4311 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4312 the host supports BWX extension and if Compaq C is present on the
4313 $PATH. Just exploiting of the BWX extension results in 20-30%
4314 performance kick for some algorithms, e.g. DES and RC4 to mention
4315 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4319 *) Add support for MS "fast SGC". This is arguably a violation of the
4320 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4321 weak crypto and after checking the certificate is SGC a second one
4322 with strong crypto. MS SGC stops the first handshake after receiving
4323 the server certificate message and sends a second client hello. Since
4324 a server will typically do all the time consuming operations before
4325 expecting any further messages from the client (server key exchange
4326 is the most expensive) there is little difference between the two.
4328 To get OpenSSL to support MS SGC we have to permit a second client
4329 hello message after we have sent server done. In addition we have to
4330 reset the MAC if we do get this second client hello.
4333 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4334 if a DER encoded private key is RSA or DSA traditional format. Changed
4335 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4336 format DER encoded private key. Newer code should use PKCS#8 format which
4337 has the key type encoded in the ASN1 structure. Added DER private key
4338 support to pkcs8 application.
4341 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4342 ciphersuites has been selected (as required by the SSL 3/TLS 1
4343 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4344 is set, we interpret this as a request to violate the specification
4345 (the worst that can happen is a handshake failure, and 'correct'
4346 behaviour would result in a handshake failure anyway).
4349 *) In SSL_CTX_add_session, take into account that there might be multiple
4350 SSL_SESSION structures with the same session ID (e.g. when two threads
4351 concurrently obtain them from an external cache).
4352 The internal cache can handle only one SSL_SESSION with a given ID,
4353 so if there's a conflict, we now throw out the old one to achieve
4357 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4358 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4359 some routines that use cipher OIDs: some ciphers do not have OIDs
4360 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4364 *) Simplify the trust setting structure and code. Now we just have
4365 two sequences of OIDs for trusted and rejected settings. These will
4366 typically have values the same as the extended key usage extension
4367 and any application specific purposes.
4369 The trust checking code now has a default behaviour: it will just
4370 check for an object with the same NID as the passed id. Functions can
4371 be provided to override either the default behaviour or the behaviour
4372 for a given id. SSL client, server and email already have functions
4373 in place for compatibility: they check the NID and also return "trusted"
4374 if the certificate is self signed.
4377 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4378 traditional format into an EVP_PKEY structure.
4381 *) Add a password callback function PEM_cb() which either prompts for
4382 a password if usr_data is NULL or otherwise assumes it is a null
4383 terminated password. Allow passwords to be passed on command line
4384 environment or config files in a few more utilities.
4387 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4388 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4389 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4390 Update documentation.
4393 *) Support for ASN1 "NULL" type. This could be handled before by using
4394 ASN1_TYPE but there wasn't any function that would try to read a NULL
4395 and produce an error if it couldn't. For compatibility we also have
4396 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4397 don't allocate anything because they don't need to.
4400 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4402 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4404 *) Rebuild of the memory allocation routines used by OpenSSL code and
4405 possibly others as well. The purpose is to make an interface that
4406 provide hooks so anyone can build a separate set of allocation and
4407 deallocation routines to be used by OpenSSL, for example memory
4408 pool implementations, or something else, which was previously hard
4409 since Malloc(), Realloc() and Free() were defined as macros having
4410 the values malloc, realloc and free, respectively (except for Win32
4411 compilations). The same is provided for memory debugging code.
4412 OpenSSL already comes with functionality to find memory leaks, but
4413 this gives people a chance to debug other memory problems.
4415 With these changes, a new set of functions and macros have appeared:
4417 CRYPTO_set_mem_debug_functions() [F]
4418 CRYPTO_get_mem_debug_functions() [F]
4419 CRYPTO_dbg_set_options() [F]
4420 CRYPTO_dbg_get_options() [F]
4421 CRYPTO_malloc_debug_init() [M]
4423 The memory debug functions are NULL by default, unless the library
4424 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4425 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4426 gives the standard debugging functions that come with OpenSSL) or
4427 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4428 provided by the library user) must be used. When the standard
4429 debugging functions are used, CRYPTO_dbg_set_options can be used to
4430 request additional information:
4431 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4432 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4434 Also, things like CRYPTO_set_mem_functions will always give the
4435 expected result (the new set of functions is used for allocation
4436 and deallocation) at all times, regardless of platform and compiler
4439 To finish it up, some functions that were never use in any other
4440 way than through macros have a new API and new semantic:
4443 CRYPTO_dbg_realloc()
4446 All macros of value have retained their old syntax.
4447 [Richard Levitte and Bodo Moeller]
4449 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4450 ordering of SMIMECapabilities wasn't in "strength order" and there
4451 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
4455 *) Some ASN1 types with illegal zero length encoding (INTEGER,
4456 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
4457 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
4459 *) Merge in my S/MIME library for OpenSSL. This provides a simple
4460 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
4461 functionality to handle multipart/signed properly) and a utility
4462 called 'smime' to call all this stuff. This is based on code I
4463 originally wrote for Celo who have kindly allowed it to be
4464 included in OpenSSL.
4467 *) Add variants des_set_key_checked and des_set_key_unchecked of
4468 des_set_key (aka des_key_sched). Global variable des_check_key
4469 decides which of these is called by des_set_key; this way
4470 des_check_key behaves as it always did, but applications and
4471 the library itself, which was buggy for des_check_key == 1,
4472 have a cleaner way to pick the version they need.
4475 *) New function PKCS12_newpass() which changes the password of a
4479 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
4480 dynamic mix. In both cases the ids can be used as an index into the
4481 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
4482 functions so they accept a list of the field values and the
4483 application doesn't need to directly manipulate the X509_TRUST
4487 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
4491 *) Modify the way the V3 extension code looks up extensions. This now
4492 works in a similar way to the object code: we have some "standard"
4493 extensions in a static table which is searched with OBJ_bsearch()
4494 and the application can add dynamic ones if needed. The file
4495 crypto/x509v3/ext_dat.h now has the info: this file needs to be
4496 updated whenever a new extension is added to the core code and kept
4497 in ext_nid order. There is a simple program 'tabtest.c' which checks
4498 this. New extensions are not added too often so this file can readily
4499 be maintained manually.
4501 There are two big advantages in doing things this way. The extensions
4502 can be looked up immediately and no longer need to be "added" using
4503 X509V3_add_standard_extensions(): this function now does nothing.
4504 [Side note: I get *lots* of email saying the extension code doesn't
4505 work because people forget to call this function]
4506 Also no dynamic allocation is done unless new extensions are added:
4507 so if we don't add custom extensions there is no need to call
4508 X509V3_EXT_cleanup().
4511 *) Modify enc utility's salting as follows: make salting the default. Add a
4512 magic header, so unsalted files fail gracefully instead of just decrypting
4513 to garbage. This is because not salting is a big security hole, so people
4514 should be discouraged from doing it.
4517 *) Fixes and enhancements to the 'x509' utility. It allowed a message
4518 digest to be passed on the command line but it only used this
4519 parameter when signing a certificate. Modified so all relevant
4520 operations are affected by the digest parameter including the
4521 -fingerprint and -x509toreq options. Also -x509toreq choked if a
4522 DSA key was used because it didn't fix the digest.
4525 *) Initial certificate chain verify code. Currently tests the untrusted
4526 certificates for consistency with the verify purpose (which is set
4527 when the X509_STORE_CTX structure is set up) and checks the pathlength.
4529 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
4530 this is because it will reject chains with invalid extensions whereas
4531 every previous version of OpenSSL and SSLeay made no checks at all.
4533 Trust code: checks the root CA for the relevant trust settings. Trust
4534 settings have an initial value consistent with the verify purpose: e.g.
4535 if the verify purpose is for SSL client use it expects the CA to be
4536 trusted for SSL client use. However the default value can be changed to
4537 permit custom trust settings: one example of this would be to only trust
4538 certificates from a specific "secure" set of CAs.
4540 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
4541 which should be used for version portability: especially since the
4542 verify structure is likely to change more often now.
4544 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
4545 to set them. If not set then assume SSL clients will verify SSL servers
4548 Two new options to the verify program: -untrusted allows a set of
4549 untrusted certificates to be passed in and -purpose which sets the
4550 intended purpose of the certificate. If a purpose is set then the
4551 new chain verify code is used to check extension consistency.
4554 *) Support for the authority information access extension.
4557 *) Modify RSA and DSA PEM read routines to transparently handle
4558 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
4559 public keys in a format compatible with certificate
4560 SubjectPublicKeyInfo structures. Unfortunately there were already
4561 functions called *_PublicKey_* which used various odd formats so
4562 these are retained for compatibility: however the DSA variants were
4563 never in a public release so they have been deleted. Changed dsa/rsa
4564 utilities to handle the new format: note no releases ever handled public
4565 keys so we should be OK.
4567 The primary motivation for this change is to avoid the same fiasco
4568 that dogs private keys: there are several incompatible private key
4569 formats some of which are standard and some OpenSSL specific and
4570 require various evil hacks to allow partial transparent handling and
4571 even then it doesn't work with DER formats. Given the option anything
4572 other than PKCS#8 should be dumped: but the other formats have to
4573 stay in the name of compatibility.
4575 With public keys and the benefit of hindsight one standard format
4576 is used which works with EVP_PKEY, RSA or DSA structures: though
4577 it clearly returns an error if you try to read the wrong kind of key.
4579 Added a -pubkey option to the 'x509' utility to output the public key.
4580 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4581 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4582 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4583 that do the same as the EVP_PKEY_assign_*() except they up the
4584 reference count of the added key (they don't "swallow" the
4588 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4589 CRLs would fail if the file contained no certificates or no CRLs:
4590 added a new function to read in both types and return the number
4591 read: this means that if none are read it will be an error. The
4592 DER versions of the certificate and CRL reader would always fail
4593 because it isn't possible to mix certificates and CRLs in DER format
4594 without choking one or the other routine. Changed this to just read
4595 a certificate: this is the best we can do. Also modified the code
4596 in apps/verify.c to take notice of return codes: it was previously
4597 attempting to read in certificates from NULL pointers and ignoring
4598 any errors: this is one reason why the cert and CRL reader seemed
4599 to work. It doesn't check return codes from the default certificate
4600 routines: these may well fail if the certificates aren't installed.
4603 *) Code to support otherName option in GeneralName.
4606 *) First update to verify code. Change the verify utility
4607 so it warns if it is passed a self signed certificate:
4608 for consistency with the normal behaviour. X509_verify
4609 has been modified to it will now verify a self signed
4610 certificate if *exactly* the same certificate appears
4611 in the store: it was previously impossible to trust a
4612 single self signed certificate. This means that:
4613 openssl verify ss.pem
4614 now gives a warning about a self signed certificate but
4615 openssl verify -CAfile ss.pem ss.pem
4619 *) For servers, store verify_result in SSL_SESSION data structure
4620 (and add it to external session representation).
4621 This is needed when client certificate verifications fails,
4622 but an application-provided verification callback (set by
4623 SSL_CTX_set_cert_verify_callback) allows accepting the session
4624 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4625 but returns 1): When the session is reused, we have to set
4626 ssl->verify_result to the appropriate error code to avoid
4628 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4630 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4631 case in PKCS7_dataInit() where the signed PKCS7 structure
4632 didn't contain any existing data because it was being created.
4633 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4635 *) Add a salt to the key derivation routines in enc.c. This
4636 forms the first 8 bytes of the encrypted file. Also add a
4637 -S option to allow a salt to be input on the command line.
4640 *) New function X509_cmp(). Oddly enough there wasn't a function
4641 to compare two certificates. We do this by working out the SHA1
4642 hash and comparing that. X509_cmp() will be needed by the trust
4646 *) SSL_get1_session() is like SSL_get_session(), but increments
4647 the reference count in the SSL_SESSION returned.
4648 [Geoff Thorpe <geoff@eu.c2.net>]
4650 *) Fix for 'req': it was adding a null to request attributes.
4651 Also change the X509_LOOKUP and X509_INFO code to handle
4652 certificate auxiliary information.
4655 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4659 *) Add the possibility to add extra information to the memory leak
4660 detecting output, to form tracebacks, showing from where each
4661 allocation was originated: CRYPTO_push_info("constant string") adds
4662 the string plus current file name and line number to a per-thread
4663 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4664 is like calling CYRPTO_pop_info() until the stack is empty.
4665 Also updated memory leak detection code to be multi-thread-safe.
4668 *) Add options -text and -noout to pkcs7 utility and delete the
4669 encryption options which never did anything. Update docs.
4672 *) Add options to some of the utilities to allow the pass phrase
4673 to be included on either the command line (not recommended on
4674 OSes like Unix) or read from the environment. Update the
4675 manpages and fix a few bugs.
4678 *) Add a few manpages for some of the openssl commands.
4681 *) Fix the -revoke option in ca. It was freeing up memory twice,
4682 leaking and not finding already revoked certificates.
4685 *) Extensive changes to support certificate auxiliary information.
4686 This involves the use of X509_CERT_AUX structure and X509_AUX
4687 functions. An X509_AUX function such as PEM_read_X509_AUX()
4688 can still read in a certificate file in the usual way but it
4689 will also read in any additional "auxiliary information". By
4690 doing things this way a fair degree of compatibility can be
4691 retained: existing certificates can have this information added
4692 using the new 'x509' options.
4694 Current auxiliary information includes an "alias" and some trust
4695 settings. The trust settings will ultimately be used in enhanced
4696 certificate chain verification routines: currently a certificate
4697 can only be trusted if it is self signed and then it is trusted
4701 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4702 The problem was that one of the replacement routines had not been working
4703 since SSLeay releases. For now the offending routine has been replaced
4704 with non-optimised assembler. Even so, this now gives around 95%
4705 performance improvement for 1024 bit RSA signs.
4708 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4709 handling. Most clients have the effective key size in bits equal to
4710 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4711 A few however don't do this and instead use the size of the decrypted key
4712 to determine the RC2 key length and the AlgorithmIdentifier to determine
4713 the effective key length. In this case the effective key length can still
4714 be 40 bits but the key length can be 168 bits for example. This is fixed
4715 by manually forcing an RC2 key into the EVP_PKEY structure because the
4716 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4717 the key length and effective key length are equal.
4720 *) Add a bunch of functions that should simplify the creation of
4721 X509_NAME structures. Now you should be able to do:
4722 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4723 and have it automatically work out the correct field type and fill in
4724 the structures. The more adventurous can try:
4725 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4726 and it will (hopefully) work out the correct multibyte encoding.
4729 *) Change the 'req' utility to use the new field handling and multibyte
4730 copy routines. Before the DN field creation was handled in an ad hoc
4731 way in req, ca, and x509 which was rather broken and didn't support
4732 BMPStrings or UTF8Strings. Since some software doesn't implement
4733 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4734 using the dirstring_type option. See the new comment in the default
4735 openssl.cnf for more info.
4738 *) Make crypto/rand/md_rand.c more robust:
4739 - Assure unique random numbers after fork().
4740 - Make sure that concurrent threads access the global counter and
4741 md serializably so that we never lose entropy in them
4742 or use exactly the same state in multiple threads.
4743 Access to the large state is not always serializable because
4744 the additional locking could be a performance killer, and
4745 md should be large enough anyway.
4748 *) New file apps/app_rand.c with commonly needed functionality
4749 for handling the random seed file.
4751 Use the random seed file in some applications that previously did not:
4753 dsaparam -genkey (which also ignored its '-rand' option),
4756 x509 (when signing).
4757 Except on systems with /dev/urandom, it is crucial to have a random
4758 seed file at least for key creation, DSA signing, and for DH exchanges;
4759 for RSA signatures we could do without one.
4761 gendh and gendsa (unlike genrsa) used to read only the first byte
4762 of each file listed in the '-rand' option. The function as previously
4763 found in genrsa is now in app_rand.c and is used by all programs
4764 that support '-rand'.
4767 *) In RAND_write_file, use mode 0600 for creating files;
4768 don't just chmod when it may be too late.
4771 *) Report an error from X509_STORE_load_locations
4772 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4775 *) New function ASN1_mbstring_copy() this copies a string in either
4776 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4777 into an ASN1_STRING type. A mask of permissible types is passed
4778 and it chooses the "minimal" type to use or an error if not type
4782 *) Add function equivalents to the various macros in asn1.h. The old
4783 macros are retained with an M_ prefix. Code inside the library can
4784 use the M_ macros. External code (including the openssl utility)
4785 should *NOT* in order to be "shared library friendly".
4788 *) Add various functions that can check a certificate's extensions
4789 to see if it usable for various purposes such as SSL client,
4790 server or S/MIME and CAs of these types. This is currently
4791 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4792 verification. Also added a -purpose flag to x509 utility to
4793 print out all the purposes.
4796 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4800 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4801 for, obtain and decode and extension and obtain its critical flag.
4802 This allows all the necessary extension code to be handled in a
4803 single function call.
4806 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4807 platforms. See crypto/rc4/rc4_enc.c for further details.
4810 *) New -noout option to asn1parse. This causes no output to be produced
4811 its main use is when combined with -strparse and -out to extract data
4812 from a file (which may not be in ASN.1 format).
4815 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4816 when producing the local key id.
4817 [Richard Levitte <levitte@stacken.kth.se>]
4819 *) New option -dhparam in s_server. This allows a DH parameter file to be
4820 stated explicitly. If it is not stated then it tries the first server
4821 certificate file. The previous behaviour hard coded the filename
4825 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4826 a public key to be input or output. For example:
4827 openssl rsa -in key.pem -pubout -out pubkey.pem
4828 Also added necessary DSA public key functions to handle this.
4831 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4832 in the message. This was handled by allowing
4833 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4834 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4836 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4837 to the end of the strings whereas this didn't. This would cause problems
4838 if strings read with d2i_ASN1_bytes() were later modified.
4839 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4841 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4842 data and it contains EOF it will end up returning an error. This is
4843 caused by input 46 bytes long. The cause is due to the way base64
4844 BIOs find the start of base64 encoded data. They do this by trying a
4845 trial decode on each line until they find one that works. When they
4846 do a flag is set and it starts again knowing it can pass all the
4847 data directly through the decoder. Unfortunately it doesn't reset
4848 the context it uses. This means that if EOF is reached an attempt
4849 is made to pass two EOFs through the context and this causes the
4850 resulting error. This can also cause other problems as well. As is
4851 usual with these problems it takes *ages* to find and the fix is
4852 trivial: move one line.
4853 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4855 *) Ugly workaround to get s_client and s_server working under Windows. The
4856 old code wouldn't work because it needed to select() on sockets and the
4857 tty (for keypresses and to see if data could be written). Win32 only
4858 supports select() on sockets so we select() with a 1s timeout on the
4859 sockets and then see if any characters are waiting to be read, if none
4860 are present then we retry, we also assume we can always write data to
4861 the tty. This isn't nice because the code then blocks until we've
4862 received a complete line of data and it is effectively polling the
4863 keyboard at 1s intervals: however it's quite a bit better than not
4864 working at all :-) A dedicated Windows application might handle this
4865 with an event loop for example.
4868 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4869 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4870 will be called when RSA_sign() and RSA_verify() are used. This is useful
4871 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4872 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4873 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4874 This necessitated the support of an extra signature type NID_md5_sha1
4875 for SSL signatures and modifications to the SSL library to use it instead
4876 of calling RSA_public_decrypt() and RSA_private_encrypt().
4879 *) Add new -verify -CAfile and -CApath options to the crl program, these
4880 will lookup a CRL issuers certificate and verify the signature in a
4881 similar way to the verify program. Tidy up the crl program so it
4882 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4883 less strict. It will now permit CRL extensions even if it is not
4884 a V2 CRL: this will allow it to tolerate some broken CRLs.
4887 *) Initialize all non-automatic variables each time one of the openssl
4888 sub-programs is started (this is necessary as they may be started
4889 multiple times from the "OpenSSL>" prompt).
4890 [Lennart Bang, Bodo Moeller]
4892 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4893 removing all other RSA functionality (this is what NO_RSA does). This
4894 is so (for example) those in the US can disable those operations covered
4895 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4899 *) Non-copying interface to BIO pairs.
4900 (still largely untested)
4903 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4904 ASCII string. This was handled independently in various places before.
4907 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4908 UTF8 strings a character at a time.
4911 *) Use client_version from client hello to select the protocol
4912 (s23_srvr.c) and for RSA client key exchange verification
4913 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4916 *) Add various utility functions to handle SPKACs, these were previously
4917 handled by poking round in the structure internals. Added new function
4918 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4919 print, verify and generate SPKACs. Based on an original idea from
4920 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4923 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4926 *) Allow the config file extension section to be overwritten on the
4927 command line. Based on an original idea from Massimiliano Pala
4928 <madwolf@comune.modena.it>. The new option is called -extensions
4929 and can be applied to ca, req and x509. Also -reqexts to override
4930 the request extensions in req and -crlexts to override the crl extensions
4934 *) Add new feature to the SPKAC handling in ca. Now you can include
4935 the same field multiple times by preceding it by "XXXX." for example:
4938 this is the same syntax as used in the req config file.
4941 *) Allow certificate extensions to be added to certificate requests. These
4942 are specified in a 'req_extensions' option of the req section of the
4943 config file. They can be printed out with the -text option to req but
4944 are otherwise ignored at present.
4947 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4948 data read consists of only the final block it would not decrypted because
4949 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4950 A misplaced 'break' also meant the decrypted final block might not be
4951 copied until the next read.
4954 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4955 a few extra parameters to the DH structure: these will be useful if
4956 for example we want the value of 'q' or implement X9.42 DH.
4959 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4960 provides hooks that allow the default DSA functions or functions on a
4961 "per key" basis to be replaced. This allows hardware acceleration and
4962 hardware key storage to be handled without major modification to the
4963 library. Also added low level modexp hooks and CRYPTO_EX structure and
4964 associated functions.
4967 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4968 as "read only": it can't be written to and the buffer it points to will
4969 not be freed. Reading from a read only BIO is much more efficient than
4970 a normal memory BIO. This was added because there are several times when
4971 an area of memory needs to be read from a BIO. The previous method was
4972 to create a memory BIO and write the data to it, this results in two
4973 copies of the data and an O(n^2) reading algorithm. There is a new
4974 function BIO_new_mem_buf() which creates a read only memory BIO from
4975 an area of memory. Also modified the PKCS#7 routines to use read only
4979 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4980 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4981 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4982 but a retry condition occured while trying to read the rest.
4985 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4986 NID_pkcs7_encrypted by default: this was wrong since this should almost
4987 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4988 the encrypted data type: this is a more sensible place to put it and it
4989 allows the PKCS#12 code to be tidied up that duplicated this
4993 *) Changed obj_dat.pl script so it takes its input and output files on
4994 the command line. This should avoid shell escape redirection problems
4998 *) Initial support for certificate extension requests, these are included
4999 in things like Xenroll certificate requests. Included functions to allow
5000 extensions to be obtained and added.
5003 *) -crlf option to s_client and s_server for sending newlines as
5004 CRLF (as required by many protocols).
5007 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5009 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5010 [Ralf S. Engelschall]
5012 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5013 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5015 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5019 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5020 DH parameters/keys (q is lost during that conversion, but the resulting
5021 DH parameters contain its length).
5023 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5024 much faster than DH_generate_parameters (which creates parameters
5025 where p = 2*q + 1), and also the smaller q makes DH computations
5026 much more efficient (160-bit exponentiation instead of 1024-bit
5027 exponentiation); so this provides a convenient way to support DHE
5028 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5029 utter importance to use
5030 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5032 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5033 when such DH parameters are used, because otherwise small subgroup
5034 attacks may become possible!
5037 *) Avoid memory leak in i2d_DHparams.
5040 *) Allow the -k option to be used more than once in the enc program:
5041 this allows the same encrypted message to be read by multiple recipients.
5044 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5045 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5046 it will always use the numerical form of the OID, even if it has a short
5050 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5051 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5052 otherwise bn_mod_exp was called. In the case of hardware keys for example
5053 no private key components need be present and it might store extra data
5054 in the RSA structure, which cannot be accessed from bn_mod_exp.
5055 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5056 private key operations.
5059 *) Added support for SPARC Linux.
5062 *) pem_password_cb function type incompatibly changed from
5063 typedef int pem_password_cb(char *buf, int size, int rwflag);
5065 ....(char *buf, int size, int rwflag, void *userdata);
5066 so that applications can pass data to their callbacks:
5067 The PEM[_ASN1]_{read,write}... functions and macros now take an
5068 additional void * argument, which is just handed through whenever
5069 the password callback is called.
5070 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5072 New function SSL_CTX_set_default_passwd_cb_userdata.
5074 Compatibility note: As many C implementations push function arguments
5075 onto the stack in reverse order, the new library version is likely to
5076 interoperate with programs that have been compiled with the old
5077 pem_password_cb definition (PEM_whatever takes some data that
5078 happens to be on the stack as its last argument, and the callback
5079 just ignores this garbage); but there is no guarantee whatsoever that
5082 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5083 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5084 problems not only on Windows, but also on some Unix platforms.
5085 To avoid problematic command lines, these definitions are now in an
5086 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5087 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5090 *) MIPS III/IV assembler module is reimplemented.
5093 *) More DES library cleanups: remove references to srand/rand and
5094 delete an unused file.
5097 *) Add support for the the free Netwide assembler (NASM) under Win32,
5098 since not many people have MASM (ml) and it can be hard to obtain.
5099 This is currently experimental but it seems to work OK and pass all
5100 the tests. Check out INSTALL.W32 for info.
5103 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5104 without temporary keys kept an extra copy of the server key,
5105 and connections with temporary keys did not free everything in case
5109 *) New function RSA_check_key and new openssl rsa option -check
5110 for verifying the consistency of RSA keys.
5111 [Ulf Moeller, Bodo Moeller]
5113 *) Various changes to make Win32 compile work:
5114 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5115 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5116 comparison" warnings.
5117 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5120 *) Add a debugging option to PKCS#5 v2 key generation function: when
5121 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5122 derived keys are printed to stderr.
5125 *) Copy the flags in ASN1_STRING_dup().
5126 [Roman E. Pavlov <pre@mo.msk.ru>]
5128 *) The x509 application mishandled signing requests containing DSA
5129 keys when the signing key was also DSA and the parameters didn't match.
5131 It was supposed to omit the parameters when they matched the signing key:
5132 the verifying software was then supposed to automatically use the CA's
5133 parameters if they were absent from the end user certificate.
5135 Omitting parameters is no longer recommended. The test was also
5136 the wrong way round! This was probably due to unusual behaviour in
5137 EVP_cmp_parameters() which returns 1 if the parameters match.
5138 This meant that parameters were omitted when they *didn't* match and
5139 the certificate was useless. Certificates signed with 'ca' didn't have
5141 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5143 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5144 The interface is as follows:
5145 Applications can use
5146 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5147 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5148 "off" is now the default.
5149 The library internally uses
5150 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5151 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5152 to disable memory-checking temporarily.
5154 Some inconsistent states that previously were possible (and were
5155 even the default) are now avoided.
5157 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5158 with each memory chunk allocated; this is occasionally more helpful
5159 than just having a counter.
5161 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5163 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5167 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5168 which largely parallels "options", but is for changing API behaviour,
5169 whereas "options" are about protocol behaviour.
5170 Initial "mode" flags are:
5172 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5173 a single record has been written.
5174 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5175 retries use the same buffer location.
5176 (But all of the contents must be
5180 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5183 *) Fix problems with no-hmac etc.
5184 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5186 *) New functions RSA_get_default_method(), RSA_set_method() and
5187 RSA_get_method(). These allows replacement of RSA_METHODs without having
5188 to mess around with the internals of an RSA structure.
5191 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5192 Also really enable memory leak checks in openssl.c and in some
5194 [Chad C. Mulligan, Bodo Moeller]
5196 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5197 up the length of negative integers. This has now been simplified to just
5198 store the length when it is first determined and use it later, rather
5199 than trying to keep track of where data is copied and updating it to
5201 [Steve Henson, reported by Brien Wheeler
5202 <bwheeler@authentica-security.com>]
5204 *) Add a new function PKCS7_signatureVerify. This allows the verification
5205 of a PKCS#7 signature but with the signing certificate passed to the
5206 function itself. This contrasts with PKCS7_dataVerify which assumes the
5207 certificate is present in the PKCS#7 structure. This isn't always the
5208 case: certificates can be omitted from a PKCS#7 structure and be
5209 distributed by "out of band" means (such as a certificate database).
5212 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5213 function prototypes in pem.h, also change util/mkdef.pl to add the
5214 necessary function names.
5217 *) mk1mf.pl (used by Windows builds) did not properly read the
5218 options set by Configure in the top level Makefile, and Configure
5219 was not even able to write more than one option correctly.
5220 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5223 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5224 file to be loaded from a BIO or FILE pointer. The BIO version will
5225 for example allow memory BIOs to contain config info.
5228 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5229 Whoever hopes to achieve shared-library compatibility across versions
5230 must use this, not the compile-time macro.
5231 (Exercise 0.9.4: Which is the minimum library version required by
5233 Note: All this applies only to multi-threaded programs, others don't
5237 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5238 through a BIO pair triggered the default case, i.e.
5239 SSLerr(...,SSL_R_UNKNOWN_STATE).
5242 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5243 can use the SSL library even if none of the specific BIOs is
5247 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5248 for the encoded length.
5249 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5251 *) Add initial documentation of the X509V3 functions.
5254 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5255 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5256 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5257 secure PKCS#8 private key format with a high iteration count.
5260 *) Fix determination of Perl interpreter: A perl or perl5
5261 _directory_ in $PATH was also accepted as the interpreter.
5262 [Ralf S. Engelschall]
5264 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5265 wrong with it but it was very old and did things like calling
5266 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5270 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5271 to use the new extension code.
5274 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5275 with macros. This should make it easier to change their form, add extra
5276 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5280 *) Add to configuration table a new entry that can specify an alternative
5281 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5282 according to Mark Crispin <MRC@Panda.COM>.
5286 *) DES CBC did not update the IV. Weird.
5289 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5290 Changing the behaviour of the former might break existing programs --
5291 where IV updating is needed, des_ncbc_encrypt can be used.
5294 *) When bntest is run from "make test" it drives bc to check its
5295 calculations, as well as internally checking them. If an internal check
5296 fails, it needs to cause bc to give a non-zero result or make test carries
5297 on without noticing the failure. Fixed.
5300 *) DES library cleanups.
5303 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5304 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5305 ciphers. NOTE: although the key derivation function has been verified
5306 against some published test vectors it has not been extensively tested
5307 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5311 *) Instead of "mkdir -p", which is not fully portable, use new
5312 Perl script "util/mkdir-p.pl".
5315 *) Rewrite the way password based encryption (PBE) is handled. It used to
5316 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5317 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5318 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5319 the 'parameter' field of the AlgorithmIdentifier is passed to the
5320 underlying key generation function so it must do its own ASN1 parsing.
5321 This has also changed the EVP_PBE_CipherInit() function which now has a
5322 'parameter' argument instead of literal salt and iteration count values
5323 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5326 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5327 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5328 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5329 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5330 value was just used as a "magic string" and not used directly its
5331 value doesn't matter.
5334 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5338 *) "linux-sparc64" configuration (ultrapenguin).
5339 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5340 "linux-sparc" configuration.
5341 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5343 *) config now generates no-xxx options for missing ciphers.
5346 *) Support the EBCDIC character set (work in progress).
5347 File ebcdic.c not yet included because it has a different license.
5348 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5350 *) Support BS2000/OSD-POSIX.
5351 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5353 *) Make callbacks for key generation use void * instead of char *.
5356 *) Make S/MIME samples compile (not yet tested).
5359 *) Additional typesafe stacks.
5362 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5366 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5368 *) New configuration variant "sco5-gcc".
5370 *) Updated some demos.
5371 [Sean O Riordain, Wade Scholine]
5373 *) Add missing BIO_free at exit of pkcs12 application.
5376 *) Fix memory leak in conf.c.
5379 *) Updates for Win32 to assembler version of MD5.
5382 *) Set #! path to perl in apps/der_chop to where we found it
5383 instead of using a fixed path.
5386 *) SHA library changes for irix64-mips4-cc.
5389 *) Improvements for VMS support.
5393 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5395 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5396 This also avoids the problems with SC4.2 and unpatched SC5.
5397 [Andy Polyakov <appro@fy.chalmers.se>]
5399 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5400 These are required because of the typesafe stack would otherwise break
5401 existing code. If old code used a structure member which used to be STACK
5402 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5403 sk_num or sk_value it would produce an error because the num, data members
5404 are not present in STACK_OF. Now it just produces a warning. sk_set
5405 replaces the old method of assigning a value to sk_value
5406 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5407 that does this will no longer work (and should use sk_set instead) but
5408 this could be regarded as a "questionable" behaviour anyway.
5411 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5412 correctly handle encrypted S/MIME data.
5415 *) Change type of various DES function arguments from des_cblock
5416 (which means, in function argument declarations, pointer to char)
5417 to des_cblock * (meaning pointer to array with 8 char elements),
5418 which allows the compiler to do more typechecking; it was like
5419 that back in SSLeay, but with lots of ugly casts.
5421 Introduce new type const_des_cblock.
5424 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5425 problems: find RecipientInfo structure that matches recipient certificate
5426 and initialise the ASN1 structures properly based on passed cipher.
5429 *) Belatedly make the BN tests actually check the results.
5432 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5433 to and from BNs: it was completely broken. New compilation option
5434 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5435 key elements as negative integers.
5438 *) Reorganize and speed up MD5.
5439 [Andy Polyakov <appro@fy.chalmers.se>]
5442 [Richard Levitte <richard@levitte.org>]
5444 *) New option -out to asn1parse to allow the parsed structure to be
5445 output to a file. This is most useful when combined with the -strparse
5446 option to examine the output of things like OCTET STRINGS.
5449 *) Make SSL library a little more fool-proof by not requiring any longer
5450 that SSL_set_{accept,connect}_state be called before
5451 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5452 in many applications because usually everything *appeared* to work as
5453 intended anyway -- now it really works as intended).
5456 *) Move openssl.cnf out of lib/.
5459 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
5460 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
5461 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
5462 [Ralf S. Engelschall]
5464 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
5465 handle PKCS#7 enveloped data properly.
5466 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
5468 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
5469 copying pointers. The cert_st handling is changed by this in
5470 various ways (and thus what used to be known as ctx->default_cert
5471 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
5472 any longer when s->cert does not give us what we need).
5473 ssl_cert_instantiate becomes obsolete by this change.
5474 As soon as we've got the new code right (possibly it already is?),
5475 we have solved a couple of bugs of the earlier code where s->cert
5476 was used as if it could not have been shared with other SSL structures.
5478 Note that using the SSL API in certain dirty ways now will result
5479 in different behaviour than observed with earlier library versions:
5480 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
5481 does not influence s as it used to.
5483 In order to clean up things more thoroughly, inside SSL_SESSION
5484 we don't use CERT any longer, but a new structure SESS_CERT
5485 that holds per-session data (if available); currently, this is
5486 the peer's certificate chain and, for clients, the server's certificate
5487 and temporary key. CERT holds only those values that can have
5488 meaningful defaults in an SSL_CTX.
5491 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
5492 from the internal representation. Various PKCS#7 fixes: remove some
5493 evil casts and set the enc_dig_alg field properly based on the signing
5497 *) Allow PKCS#12 password to be set from the command line or the
5498 environment. Let 'ca' get its config file name from the environment
5499 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
5503 *) Allow certificate policies extension to use an IA5STRING for the
5504 organization field. This is contrary to the PKIX definition but
5505 VeriSign uses it and IE5 only recognises this form. Document 'x509'
5509 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
5510 without disallowing inline assembler and the like for non-pedantic builds.
5513 *) Support Borland C++ builder.
5514 [Janez Jere <jj@void.si>, modified by Ulf Möller]
5519 *) SHA-1 cleanups and performance enhancements.
5520 [Andy Polyakov <appro@fy.chalmers.se>]
5522 *) Sparc v8plus assembler for the bignum library.
5523 [Andy Polyakov <appro@fy.chalmers.se>]
5525 *) Accept any -xxx and +xxx compiler options in Configure.
5528 *) Update HPUX configuration.
5531 *) Add missing sk_<type>_unshift() function to safestack.h
5532 [Ralf S. Engelschall]
5534 *) New function SSL_CTX_use_certificate_chain_file that sets the
5535 "extra_cert"s in addition to the certificate. (This makes sense
5536 only for "PEM" format files, as chains as a whole are not
5540 *) Support verify_depth from the SSL API.
5541 x509_vfy.c had what can be considered an off-by-one-error:
5542 Its depth (which was not part of the external interface)
5543 was actually counting the number of certificates in a chain;
5544 now it really counts the depth.
5547 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
5548 instead of X509err, which often resulted in confusing error
5549 messages since the error codes are not globally unique
5550 (e.g. an alleged error in ssl3_accept when a certificate
5551 didn't match the private key).
5553 *) New function SSL_CTX_set_session_id_context that allows to set a default
5554 value (so that you don't need SSL_set_session_id_context for each
5555 connection using the SSL_CTX).
5558 *) OAEP decoding bug fix.
5561 *) Support INSTALL_PREFIX for package builders, as proposed by
5565 *) New Configure options "threads" and "no-threads". For systems
5566 where the proper compiler options are known (currently Solaris
5567 and Linux), "threads" is the default.
5570 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5573 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5574 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5575 such as /usr/local/bin.
5578 *) "make linux-shared" to build shared libraries.
5579 [Niels Poppe <niels@netbox.org>]
5581 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5584 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5585 extension adding in x509 utility.
5588 *) Remove NOPROTO sections and error code comments.
5591 *) Partial rewrite of the DEF file generator to now parse the ANSI
5595 *) New Configure options --prefix=DIR and --openssldir=DIR.
5598 *) Complete rewrite of the error code script(s). It is all now handled
5599 by one script at the top level which handles error code gathering,
5600 header rewriting and C source file generation. It should be much better
5601 than the old method: it now uses a modified version of Ulf's parser to
5602 read the ANSI prototypes in all header files (thus the old K&R definitions
5603 aren't needed for error creation any more) and do a better job of
5604 translating function codes into names. The old 'ASN1 error code imbedded
5605 in a comment' is no longer necessary and it doesn't use .err files which
5606 have now been deleted. Also the error code call doesn't have to appear all
5607 on one line (which resulted in some large lines...).
5610 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5613 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5614 0 (which usually indicates a closed connection), but continue reading.
5617 *) Fix some race conditions.
5620 *) Add support for CRL distribution points extension. Add Certificate
5621 Policies and CRL distribution points documentation.
5624 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5627 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5628 8 of keying material. Merlin has also confirmed interop with this fix
5629 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5630 [Merlin Hughes <merlin@baltimore.ie>]
5632 *) Fix lots of warnings.
5633 [Richard Levitte <levitte@stacken.kth.se>]
5635 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5636 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5637 [Richard Levitte <levitte@stacken.kth.se>]
5639 *) Fix problems with sizeof(long) == 8.
5640 [Andy Polyakov <appro@fy.chalmers.se>]
5642 *) Change functions to ANSI C.
5645 *) Fix typos in error codes.
5646 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5648 *) Remove defunct assembler files from Configure.
5651 *) SPARC v8 assembler BIGNUM implementation.
5652 [Andy Polyakov <appro@fy.chalmers.se>]
5654 *) Support for Certificate Policies extension: both print and set.
5655 Various additions to support the r2i method this uses.
5658 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5659 return a const string when you are expecting an allocated buffer.
5662 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5663 types DirectoryString and DisplayText.
5666 *) Add code to allow r2i extensions to access the configuration database,
5667 add an LHASH database driver and add several ctx helper functions.
5670 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5671 fail when they extended the size of a BIGNUM.
5674 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5675 support typesafe stack.
5678 *) Fix typo in SSL_[gs]et_options().
5679 [Nils Frostberg <nils@medcom.se>]
5681 *) Delete various functions and files that belonged to the (now obsolete)
5682 old X509V3 handling code.
5685 *) New Configure option "rsaref".
5688 *) Don't auto-generate pem.h.
5691 *) Introduce type-safe ASN.1 SETs.
5694 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5695 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5697 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5698 that links with OpenSSL (well at least cause lots of warnings), but fear
5699 not: the conversion is trivial, and it eliminates loads of evil casts. A
5700 few STACKed things have been converted already. Feel free to convert more.
5701 In the fullness of time, I'll do away with the STACK type altogether.
5704 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5705 specified in <certfile> by updating the entry in the index.txt file.
5706 This way one no longer has to edit the index.txt file manually for
5707 revoking a certificate. The -revoke option does the gory details now.
5708 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5710 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5711 `-text' option at all and this way the `-noout -text' combination was
5712 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5713 [Ralf S. Engelschall]
5715 *) Make sure a corresponding plain text error message exists for the
5716 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5717 verify callback function determined that a certificate was revoked.
5718 [Ralf S. Engelschall]
5720 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5721 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5722 all available cipers including rc5, which was forgotten until now.
5723 In order to let the testing shell script know which algorithms
5724 are available, a new (up to now undocumented) command
5725 "openssl list-cipher-commands" is used.
5728 *) Bugfix: s_client occasionally would sleep in select() when
5729 it should have checked SSL_pending() first.
5732 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5733 the raw DSA values prior to ASN.1 encoding.
5736 *) Tweaks to Configure
5737 [Niels Poppe <niels@netbox.org>]
5739 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5743 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5746 *) New config option to avoid instructions that are illegal on the 80386.
5747 The default code is faster, but requires at least a 486.
5750 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5751 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5752 same as SSL2_VERSION anyway.
5755 *) New "-showcerts" option for s_client.
5758 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5759 application. Various cleanups and fixes.
5762 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5763 modify error routines to work internally. Add error codes and PBE init
5764 to library startup routines.
5767 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5768 packing functions to asn1 and evp. Changed function names and error
5769 codes along the way.
5772 *) PKCS12 integration: and so it begins... First of several patches to
5773 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5774 objects to objects.h
5777 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5778 and display support for Thawte strong extranet extension.
5781 *) Add LinuxPPC support.
5782 [Jeff Dubrule <igor@pobox.org>]
5784 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5785 bn_div_words in alpha.s.
5786 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5788 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5789 OAEP isn't supported when OpenSSL is built with RSAref.
5790 [Ulf Moeller <ulf@fitug.de>]
5792 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5793 so they no longer are missing under -DNOPROTO.
5794 [Soren S. Jorvang <soren@t.dk>]
5797 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5799 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5800 doesn't work when the session is reused. Coming soon!
5803 *) Fix a security hole, that allows sessions to be reused in the wrong
5804 context thus bypassing client cert protection! All software that uses
5805 client certs and session caches in multiple contexts NEEDS PATCHING to
5806 allow session reuse! A fuller solution is in the works.
5807 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5809 *) Some more source tree cleanups (removed obsolete files
5810 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5811 permission on "config" script to be executable) and a fix for the INSTALL
5813 [Ulf Moeller <ulf@fitug.de>]
5815 *) Remove some legacy and erroneous uses of malloc, free instead of
5817 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5819 *) Make rsa_oaep_test return non-zero on error.
5820 [Ulf Moeller <ulf@fitug.de>]
5822 *) Add support for native Solaris shared libraries. Configure
5823 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5824 if someone would make that last step automatic.
5825 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5827 *) ctx_size was not built with the right compiler during "make links". Fixed.
5830 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5831 except NULL ciphers". This means the default cipher list will no longer
5832 enable NULL ciphers. They need to be specifically enabled e.g. with
5833 the string "DEFAULT:eNULL".
5836 *) Fix to RSA private encryption routines: if p < q then it would
5837 occasionally produce an invalid result. This will only happen with
5838 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5841 *) Be less restrictive and allow also `perl util/perlpath.pl
5842 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5843 because this way one can also use an interpreter named `perl5' (which is
5844 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5845 installed as `perl').
5846 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5848 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5849 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5851 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5852 advapi32.lib to Win32 build and change the pem test comparision
5853 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5854 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5855 and crypto/des/ede_cbcm_enc.c.
5858 *) DES quad checksum was broken on big-endian architectures. Fixed.
5861 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5862 Win32 test batch file so it (might) work again. The Win32 test batch file
5863 is horrible: I feel ill....
5866 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5867 in e_os.h. Audit of header files to check ANSI and non ANSI
5868 sections: 10 functions were absent from non ANSI section and not exported
5869 from Windows DLLs. Fixed up libeay.num for new functions.
5872 *) Make `openssl version' output lines consistent.
5873 [Ralf S. Engelschall]
5875 *) Fix Win32 symbol export lists for BIO functions: Added
5876 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5877 to ms/libeay{16,32}.def.
5878 [Ralf S. Engelschall]
5880 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5881 fine under Unix and passes some trivial tests I've now added. But the
5882 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5883 added to make sure no one expects that this stuff really works in the
5884 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5885 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5887 [Ralf S. Engelschall]
5889 *) Fix the generation of two part addresses in perl.
5890 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5892 *) Add config entry for Linux on MIPS.
5893 [John Tobey <jtobey@channel1.com>]
5895 *) Make links whenever Configure is run, unless we are on Windoze.
5898 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5899 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5903 *) Add a useful kludge to allow package maintainers to specify compiler and
5904 other platforms details on the command line without having to patch the
5905 Configure script everytime: One now can use ``perl Configure
5906 <id>:<details>'', i.e. platform ids are allowed to have details appended
5907 to them (seperated by colons). This is treated as there would be a static
5908 pre-configured entry in Configure's %table under key <id> with value
5909 <details> and ``perl Configure <id>'' is called. So, when you want to
5910 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5911 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5912 now, which overrides the FreeBSD-elf entry on-the-fly.
5913 [Ralf S. Engelschall]
5915 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5918 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5919 on the `perl Configure ...' command line. This way one can compile
5920 OpenSSL libraries with Position Independent Code (PIC) which is needed
5921 for linking it into DSOs.
5922 [Ralf S. Engelschall]
5924 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5928 *) Cleaned up the LICENSE document: The official contact for any license
5929 questions now is the OpenSSL core team under openssl-core@openssl.org.
5930 And add a paragraph about the dual-license situation to make sure people
5931 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5932 to the OpenSSL toolkit.
5933 [Ralf S. Engelschall]
5935 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5936 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5937 Additonally cleaned up the `make links' target: Remove unnecessary
5938 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5939 to speed processing and no longer clutter the display with confusing
5940 stuff. Instead only the actually done links are displayed.
5941 [Ralf S. Engelschall]
5943 *) Permit null encryption ciphersuites, used for authentication only. It used
5944 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5945 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5949 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5950 signed attributes when verifying signatures (this would break them),
5951 the detached data encoding was wrong and public keys obtained using
5952 X509_get_pubkey() weren't freed.
5955 *) Add text documentation for the BUFFER functions. Also added a work around
5956 to a Win95 console bug. This was triggered by the password read stuff: the
5957 last character typed gets carried over to the next fread(). If you were
5958 generating a new cert request using 'req' for example then the last
5959 character of the passphrase would be CR which would then enter the first
5963 *) Added the new `Includes OpenSSL Cryptography Software' button as
5964 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5965 button and can be used by applications based on OpenSSL to show the
5966 relationship to the OpenSSL project.
5967 [Ralf S. Engelschall]
5969 *) Remove confusing variables in function signatures in files
5970 ssl/ssl_lib.c and ssl/ssl.h.
5971 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5973 *) Don't install bss_file.c under PREFIX/include/
5974 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5976 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5977 functions that return function pointers and has support for NT specific
5978 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5979 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5980 unsigned to signed types: this was killing the Win32 compile.
5983 *) Add new certificate file to stack functions,
5984 SSL_add_dir_cert_subjects_to_stack() and
5985 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5986 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5987 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5988 This means that Apache-SSL and similar packages don't have to mess around
5989 to add as many CAs as they want to the preferred list.
5992 *) Experiment with doxygen documentation. Currently only partially applied to
5994 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5995 openssl.doxy as the configuration file.
5998 *) Get rid of remaining C++-style comments which strict C compilers hate.
5999 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6001 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6002 compiled in by default: it has problems with large keys.
6005 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6006 DH private keys and/or callback functions which directly correspond to
6007 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6008 is needed for applications which have to configure certificates on a
6009 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6011 For the RSA certificate situation is makes no difference, but
6012 for the DSA certificate situation this fixes the "no shared cipher"
6013 problem where the OpenSSL cipher selection procedure failed because the
6014 temporary keys were not overtaken from the context and the API provided
6015 no way to reconfigure them.
6016 The new functions now let applications reconfigure the stuff and they
6017 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6018 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6019 non-public-API function ssl_cert_instantiate() is used as a helper
6020 function and also to reduce code redundancy inside ssl_rsa.c.
6021 [Ralf S. Engelschall]
6023 *) Move s_server -dcert and -dkey options out of the undocumented feature
6024 area because they are useful for the DSA situation and should be
6025 recognized by the users.
6026 [Ralf S. Engelschall]
6028 *) Fix the cipher decision scheme for export ciphers: the export bits are
6029 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6030 SSL_EXP_MASK. So, the original variable has to be used instead of the
6031 already masked variable.
6032 [Richard Levitte <levitte@stacken.kth.se>]
6034 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6035 [Richard Levitte <levitte@stacken.kth.se>]
6037 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6038 from `int' to `unsigned int' because it's a length and initialized by
6039 EVP_DigestFinal() which expects an `unsigned int *'.
6040 [Richard Levitte <levitte@stacken.kth.se>]
6042 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6043 script. Instead use the usual Shell->Perl transition trick.
6044 [Ralf S. Engelschall]
6046 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6047 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6048 -noout -modulus' as it's already the case for `openssl rsa -noout
6049 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6050 currently the public key is printed (a decision which was already done by
6051 `openssl dsa -modulus' in the past) which serves a similar purpose.
6052 Additionally the NO_RSA no longer completely removes the whole -modulus
6053 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6055 [Ralf S. Engelschall]
6057 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6058 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6059 [Arne Ansper <arne@ats.cyber.ee>]
6061 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6062 to be added. Now both 'req' and 'ca' can use new objects defined in the
6066 *) Add cool BIO that does syslog (or event log on NT).
6067 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6069 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6070 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6071 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6072 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6075 *) Add preliminary config info for new extension code.
6078 *) Make RSA_NO_PADDING really use no padding.
6079 [Ulf Moeller <ulf@fitug.de>]
6081 *) Generate errors when private/public key check is done.
6084 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6085 for some CRL extensions and new objects added.
6088 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6089 key usage extension and fuller support for authority key id.
6092 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6093 padding method for RSA, which is recommended for new applications in PKCS
6094 #1 v2.0 (RFC 2437, October 1998).
6095 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6096 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6097 against Bleichbacher's attack on RSA.
6098 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6101 *) Updates to the new SSL compression code
6102 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6104 *) Fix so that the version number in the master secret, when passed
6105 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6106 (because the server will not accept higher), that the version number
6107 is 0x03,0x01, not 0x03,0x00
6108 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6110 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6111 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6112 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6115 *) Support for RAW extensions where an arbitrary extension can be
6116 created by including its DER encoding. See apps/openssl.cnf for
6120 *) Make sure latest Perl versions don't interpret some generated C array
6121 code as Perl array code in the crypto/err/err_genc.pl script.
6122 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6124 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6125 not many people have the assembler. Various Win32 compilation fixes and
6126 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6130 *) Modify configure script 'Configure' to automatically create crypto/date.h
6131 file under Win32 and also build pem.h from pem.org. New script
6132 util/mkfiles.pl to create the MINFO file on environments that can't do a
6133 'make files': perl util/mkfiles.pl >MINFO should work.
6136 *) Major rework of DES function declarations, in the pursuit of correctness
6137 and purity. As a result, many evil casts evaporated, and some weirdness,
6138 too. You may find this causes warnings in your code. Zapping your evil
6139 casts will probably fix them. Mostly.
6142 *) Fix for a typo in asn1.h. Bug fix to object creation script
6143 obj_dat.pl. It considered a zero in an object definition to mean
6144 "end of object": none of the objects in objects.h have any zeros
6145 so it wasn't spotted.
6146 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6148 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6149 Masking (CBCM). In the absence of test vectors, the best I have been able
6150 to do is check that the decrypt undoes the encrypt, so far. Send me test
6151 vectors if you have them.
6154 *) Correct calculation of key length for export ciphers (too much space was
6155 allocated for null ciphers). This has not been tested!
6158 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6159 message is now correct (it understands "crypto" and "ssl" on its
6160 command line). There is also now an "update" option. This will update
6161 the util/ssleay.num and util/libeay.num files with any new functions.
6163 perl util/mkdef.pl crypto ssl update
6164 it will update them.
6167 *) Overhauled the Perl interface (perl/*):
6168 - ported BN stuff to OpenSSL's different BN library
6169 - made the perl/ source tree CVS-aware
6170 - renamed the package from SSLeay to OpenSSL (the files still contain
6171 their history because I've copied them in the repository)
6172 - removed obsolete files (the test scripts will be replaced
6173 by better Test::Harness variants in the future)
6174 [Ralf S. Engelschall]
6176 *) First cut for a very conservative source tree cleanup:
6177 1. merge various obsolete readme texts into doc/ssleay.txt
6178 where we collect the old documents and readme texts.
6179 2. remove the first part of files where I'm already sure that we no
6180 longer need them because of three reasons: either they are just temporary
6181 files which were left by Eric or they are preserved original files where
6182 I've verified that the diff is also available in the CVS via "cvs diff
6183 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6184 the crypto/md/ stuff).
6185 [Ralf S. Engelschall]
6187 *) More extension code. Incomplete support for subject and issuer alt
6188 name, issuer and authority key id. Change the i2v function parameters
6189 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6190 what that's for :-) Fix to ASN1 macro which messed up
6191 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6194 *) Preliminary support for ENUMERATED type. This is largely copied from the
6198 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6199 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6201 *) Make sure `make rehash' target really finds the `openssl' program.
6202 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6204 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6205 like to hear about it if this slows down other processors.
6208 *) Add CygWin32 platform information to Configure script.
6209 [Alan Batie <batie@aahz.jf.intel.com>]
6211 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6212 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6214 *) New program nseq to manipulate netscape certificate sequences
6217 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6221 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6222 but the BN code had some problems that would cause failures when
6223 doing certificate verification and some other functions.
6224 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6226 *) Add ASN1 and PEM code to support netscape certificate sequences.
6229 *) Add ASN1 and PEM code to support netscape certificate sequences.
6232 *) Add several PKIX and private extended key usage OIDs.
6235 *) Modify the 'ca' program to handle the new extension code. Modify
6236 openssl.cnf for new extension format, add comments.
6239 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6240 and add a sample to openssl.cnf so req -x509 now adds appropriate
6244 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6245 error code, add initial support to X509_print() and x509 application.
6248 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6249 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6250 stuff is currently isolated and isn't even compiled yet.
6253 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6254 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6255 Removed the versions check from X509 routines when loading extensions:
6256 this allows certain broken certificates that don't set the version
6257 properly to be processed.
6260 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6261 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6262 can still be regenerated with "make depend".
6265 *) Spelling mistake in C version of CAST-128.
6266 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6268 *) Changes to the error generation code. The perl script err-code.pl
6269 now reads in the old error codes and retains the old numbers, only
6270 adding new ones if necessary. It also only changes the .err files if new
6271 codes are added. The makefiles have been modified to only insert errors
6272 when needed (to avoid needlessly modifying header files). This is done
6273 by only inserting errors if the .err file is newer than the auto generated
6274 C file. To rebuild all the error codes from scratch (the old behaviour)
6275 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6276 or delete all the .err files.
6279 *) CAST-128 was incorrectly implemented for short keys. The C version has
6280 been fixed, but is untested. The assembler versions are also fixed, but
6281 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6282 to regenerate it if needed.
6283 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6284 Hagino <itojun@kame.net>]
6286 *) File was opened incorrectly in randfile.c.
6287 [Ulf Möller <ulf@fitug.de>]
6289 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6290 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6291 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6292 al: it's just almost always a UTCTime. Note this patch adds new error
6293 codes so do a "make errors" if there are problems.
6296 *) Correct Linux 1 recognition in config.
6297 [Ulf Möller <ulf@fitug.de>]
6299 *) Remove pointless MD5 hash when using DSA keys in ca.
6300 [Anonymous <nobody@replay.com>]
6302 *) Generate an error if given an empty string as a cert directory. Also
6303 generate an error if handed NULL (previously returned 0 to indicate an
6304 error, but didn't set one).
6305 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6307 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6310 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6311 parameters. This was causing a warning which killed off the Win32 compile.
6314 *) Remove C++ style comments from crypto/bn/bn_local.h.
6315 [Neil Costigan <neil.costigan@celocom.com>]
6317 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6318 based on a text string, looking up short and long names and finally
6319 "dot" format. The "dot" format stuff didn't work. Added new function
6320 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6321 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6322 OID is not part of the table.
6325 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6326 X509_LOOKUP_by_alias().
6329 *) Sort openssl functions by name.
6332 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6333 encryption from sample DSA keys (in case anyone is interested the password
6337 *) Make _all_ *_free functions accept a NULL pointer.
6338 [Frans Heymans <fheymans@isaserver.be>]
6340 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6342 [Anonymous <nobody@replay.com>]
6344 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6345 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6347 *) Don't blow it for numeric -newkey arguments to apps/req.
6348 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6350 *) Temp key "for export" tests were wrong in s3_srvr.c.
6351 [Anonymous <nobody@replay.com>]
6353 *) Add prototype for temp key callback functions
6354 SSL_CTX_set_tmp_{rsa,dh}_callback().
6357 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6358 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6361 *) X509_name_add_entry() freed the wrong thing after an error.
6362 [Arne Ansper <arne@ats.cyber.ee>]
6364 *) rsa_eay.c would attempt to free a NULL context.
6365 [Arne Ansper <arne@ats.cyber.ee>]
6367 *) BIO_s_socket() had a broken should_retry() on Windoze.
6368 [Arne Ansper <arne@ats.cyber.ee>]
6370 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6371 [Arne Ansper <arne@ats.cyber.ee>]
6373 *) Make sure the already existing X509_STORE->depth variable is initialized
6374 in X509_STORE_new(), but document the fact that this variable is still
6375 unused in the certificate verification process.
6376 [Ralf S. Engelschall]
6378 *) Fix the various library and apps files to free up pkeys obtained from
6379 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6382 *) Fix reference counting in X509_PUBKEY_get(). This makes
6383 demos/maurice/example2.c work, amongst others, probably.
6384 [Steve Henson and Ben Laurie]
6386 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6387 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6388 are no longer created. This way we have a single and consistent command
6389 line interface `openssl <command>', similar to `cvs <command>'.
6390 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6392 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6393 BIT STRING wrapper always have zero unused bits.
6396 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6399 *) Make the top-level INSTALL documentation easier to understand.
6402 *) Makefiles updated to exit if an error occurs in a sub-directory
6403 make (including if user presses ^C) [Paul Sutton]
6405 *) Make Montgomery context stuff explicit in RSA data structure.
6408 *) Fix build order of pem and err to allow for generated pem.h.
6411 *) Fix renumbering bug in X509_NAME_delete_entry().
6414 *) Enhanced the err-ins.pl script so it makes the error library number
6415 global and can add a library name. This is needed for external ASN1 and
6416 other error libraries.
6419 *) Fixed sk_insert which never worked properly.
6422 *) Fix ASN1 macros so they can handle indefinite length construted
6423 EXPLICIT tags. Some non standard certificates use these: they can now
6427 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6428 into a single doc/ssleay.txt bundle. This way the information is still
6429 preserved but no longer messes up this directory. Now it's new room for
6430 the new set of documenation files.
6431 [Ralf S. Engelschall]
6433 *) SETs were incorrectly DER encoded. This was a major pain, because they
6434 shared code with SEQUENCEs, which aren't coded the same. This means that
6435 almost everything to do with SETs or SEQUENCEs has either changed name or
6436 number of arguments.
6437 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6439 *) Fix test data to work with the above.
6442 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6443 was already fixed by Eric for 0.9.1 it seems.
6444 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6446 *) Autodetect FreeBSD3.
6449 *) Fix various bugs in Configure. This affects the following platforms:
6453 unixware-2.0-pentium
6457 *) Eliminate generated files from CVS. Reorder tests to regenerate files
6458 before they are needed.
6461 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
6465 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
6467 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
6468 changed SSLeay to OpenSSL in version strings.
6469 [Ralf S. Engelschall]
6471 *) Some fixups to the top-level documents.
6474 *) Fixed the nasty bug where rsaref.h was not found under compile-time
6475 because the symlink to include/ was missing.
6476 [Ralf S. Engelschall]
6478 *) Incorporated the popular no-RSA/DSA-only patches
6479 which allow to compile a RSA-free SSLeay.
6480 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
6482 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
6483 when "ssleay" is still not found.
6484 [Ralf S. Engelschall]
6486 *) Added more platforms to Configure: Cray T3E, HPUX 11,
6487 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
6489 *) Updated the README file.
6490 [Ralf S. Engelschall]
6492 *) Added various .cvsignore files in the CVS repository subdirs
6493 to make a "cvs update" really silent.
6494 [Ralf S. Engelschall]
6496 *) Recompiled the error-definition header files and added
6497 missing symbols to the Win32 linker tables.
6498 [Ralf S. Engelschall]
6500 *) Cleaned up the top-level documents;
6501 o new files: CHANGES and LICENSE
6502 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
6503 o merged COPYRIGHT into LICENSE
6504 o removed obsolete TODO file
6505 o renamed MICROSOFT to INSTALL.W32
6506 [Ralf S. Engelschall]
6508 *) Removed dummy files from the 0.9.1b source tree:
6509 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
6510 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
6511 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
6512 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
6513 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
6514 [Ralf S. Engelschall]
6516 *) Added various platform portability fixes.
6519 *) The Genesis of the OpenSSL rpject:
6520 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
6521 Young and Tim J. Hudson created while they were working for C2Net until
6523 [The OpenSSL Project]
6526 Changes between 0.9.0b and 0.9.1b [not released]
6528 *) Updated a few CA certificates under certs/
6531 *) Changed some BIGNUM api stuff.
6534 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
6535 DGUX x86, Linux Alpha, etc.
6538 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
6539 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
6543 *) Add -strparse option to asn1pars program which parses nested
6545 [Dr Stephen Henson <shenson@bigfoot.com>]
6547 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
6550 *) DSA fix for "ca" program.
6553 *) Added "-genkey" option to "dsaparam" program.
6556 *) Added RIPE MD160 (rmd160) message digest.
6559 *) Added -a (all) option to "ssleay version" command.
6562 *) Added PLATFORM define which is the id given to Configure.
6565 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6568 *) Extended the ASN.1 parser routines.
6571 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6574 *) Added a BN_CTX to the BN library.
6577 *) Fixed the weak key values in DES library
6580 *) Changed API in EVP library for cipher aliases.
6583 *) Added support for RC2/64bit cipher.
6586 *) Converted the lhash library to the crypto/mem.c functions.
6589 *) Added more recognized ASN.1 object ids.
6592 *) Added more RSA padding checks for SSL/TLS.
6595 *) Added BIO proxy/filter functionality.
6598 *) Added extra_certs to SSL_CTX which can be used
6599 send extra CA certificates to the client in the CA cert chain sending
6600 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6603 *) Now Fortezza is denied in the authentication phase because
6604 this is key exchange mechanism is not supported by SSLeay at all.
6607 *) Additional PKCS1 checks.
6610 *) Support the string "TLSv1" for all TLS v1 ciphers.
6613 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6614 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6617 *) Fixed a few memory leaks.
6620 *) Fixed various code and comment typos.
6623 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6624 bytes sent in the client random.
6625 [Edward Bishop <ebishop@spyglass.com>]