5 Changes between 0.9.8 and 0.9.9 [xx XXX xxxx]
9 Changes between 0.9.7h and 0.9.8 [xx XXX xxxx]
11 *) Add attribute functions to EVP_PKEY structure. Modify
12 PKCS12_create() to recognize a CSP name attribute and
13 use it. Make -CSP option work again in pkcs12 utility.
16 *) Add new functionality to the bn blinding code:
17 - automatic re-creation of the BN_BLINDING parameters after
18 a fixed number of uses (currently 32)
19 - add new function for parameter creation
20 - introduce flags to control the update behaviour of the
21 BN_BLINDING parameters
22 - hide BN_BLINDING structure
23 Add a second BN_BLINDING slot to the RSA structure to improve
24 performance when a single RSA object is shared among several
28 *) Add support for DTLS.
29 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
31 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
32 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
35 *) Remove buggy and incompletet DH cert support from
36 ssl/ssl_rsa.c and ssl/s3_both.c
39 *) Use SHA-1 instead of MD5 as the default digest algorithm for
40 the apps/openssl applications.
43 *) Compile clean with "-Wall -Wmissing-prototypes
44 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
45 DEBUG_SAFESTACK must also be set.
48 *) Change ./Configure so that certain algorithms can be disabled by default.
49 The new counterpiece to "no-xxx" is "enable-xxx".
51 The patented RC5 and MDC2 algorithms will now be disabled unless
52 "enable-rc5" and "enable-mdc2", respectively, are specified.
54 (IDEA remains enabled despite being patented. This is because IDEA
55 is frequently required for interoperability, and there is no license
56 fee for non-commercial use. As before, "no-idea" can be used to
57 avoid this algorithm.)
61 *) Add processing of proxy certificates (see RFC 3820). This work was
62 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
63 EGEE (Enabling Grids for E-science in Europe).
66 *) RC4 performance overhaul on modern architectures/implementations, such
67 as Intel P4, IA-64 and AMD64.
70 *) New utility extract-section.pl. This can be used specify an alternative
71 section number in a pod file instead of having to treat each file as
72 a separate case in Makefile. This can be done by adding two lines to the
75 =for comment openssl_section:XXX
77 The blank line is mandatory.
81 *) New arguments -certform, -keyform and -pass for s_client and s_server
82 to allow alternative format key and certificate files and passphrase
86 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
87 update associated structures and add various utility functions.
89 Add new policy related verify parameters, include policy checking in
90 standard verify code. Enhance 'smime' application with extra parameters
91 to support policy checking and print out.
94 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
95 Nehemiah processors. These extensions support AES encryption in hardware
96 as well as RNG (though RNG support is currently disabled).
97 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
99 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
102 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
103 [Andy Polyakov and a number of other people]
105 *) Improved PowerPC platform support. Most notably BIGNUM assembler
106 implementation contributed by IBM.
107 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
109 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
110 exponent rather than 'unsigned long'. There is a corresponding change to
111 the new 'rsa_keygen' element of the RSA_METHOD structure.
112 [Jelte Jansen, Geoff Thorpe]
114 *) Functionality for creating the initial serial number file is now
115 moved from CA.pl to the 'ca' utility with a new option -create_serial.
117 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
118 number file to 1, which is bound to cause problems. To avoid
119 the problems while respecting compatibility between different 0.9.7
120 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
121 CA.pl for serial number initialization. With the new release 0.9.8,
122 we can fix the problem directly in the 'ca' utility.)
125 *) Reduced header interdepencies by declaring more opaque objects in
126 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
127 give fewer recursive includes, which could break lazy source code - so
128 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
129 developers should define this symbol when building and using openssl to
130 ensure they track the recommended behaviour, interfaces, [etc], but
131 backwards-compatible behaviour prevails when this isn't defined.
134 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
137 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
138 This will generate a random key of the appropriate length based on the
139 cipher context. The EVP_CIPHER can provide its own random key generation
140 routine to support keys of a specific form. This is used in the des and
141 3des routines to generate a key of the correct parity. Update S/MIME
142 code to use new functions and hence generate correct parity DES keys.
143 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
144 valid (weak or incorrect parity).
147 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
148 as looking them up. This is useful when the verified structure may contain
149 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
150 present unless the new PKCS7_NO_CRL flag is asserted.
153 *) Extend ASN1 oid configuration module. It now additionally accepts the
156 shortName = some long name, 1.2.3.4
159 *) Reimplemented the BN_CTX implementation. There is now no more static
160 limitation on the number of variables it can handle nor the depth of the
161 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
162 information can now expand as required, and rather than having a single
163 static array of bignums, BN_CTX now uses a linked-list of such arrays
164 allowing it to expand on demand whilst maintaining the usefulness of
168 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
169 to allow all RSA operations to function using a single BN_CTX.
172 *) Preliminary support for certificate policy evaluation and checking. This
173 is initially intended to pass the tests outlined in "Conformance Testing
174 of Relying Party Client Certificate Path Processing Logic" v1.07.
177 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
178 remained unused and not that useful. A variety of other little bignum
179 tweaks and fixes have also been made continuing on from the audit (see
183 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
184 associated ASN1, EVP and SSL functions and old ASN1 macros.
187 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
188 and this should never fail. So the return value from the use of
189 BN_set_word() (which can fail due to needless expansion) is now deprecated;
190 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
193 *) BN_CTX_get() should return zero-valued bignums, providing the same
194 initialised value as BN_new().
195 [Geoff Thorpe, suggested by Ulf Möller]
197 *) Support for inhibitAnyPolicy certificate extension.
200 *) An audit of the BIGNUM code is underway, for which debugging code is
201 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
202 is considered valid when processing BIGNUMs, and causes execution to
203 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
204 further steps are taken to deliberately pollute unused data in BIGNUM
205 structures to try and expose faulty code further on. For now, openssl will
206 (in its default mode of operation) continue to tolerate the inconsistent
207 forms that it has tolerated in the past, but authors and packagers should
208 consider trying openssl and their own applications when compiled with
209 these debugging symbols defined. It will help highlight potential bugs in
210 their own code, and will improve the test coverage for OpenSSL itself. At
211 some point, these tighter rules will become openssl's default to improve
212 maintainability, though the assert()s and other overheads will remain only
213 in debugging configurations. See bn.h for more details.
214 [Geoff Thorpe, Nils Larsch, Ulf Möller]
216 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
217 that can only be obtained through BN_CTX_new() (which implicitly
218 initialises it). The presence of this function only made it possible
219 to overwrite an existing structure (and cause memory leaks).
222 *) Because of the callback-based approach for implementing LHASH as a
223 template type, lh_insert() adds opaque objects to hash-tables and
224 lh_doall() or lh_doall_arg() are typically used with a destructor callback
225 to clean up those corresponding objects before destroying the hash table
226 (and losing the object pointers). So some over-zealous constifications in
227 LHASH have been relaxed so that lh_insert() does not take (nor store) the
228 objects as "const" and the lh_doall[_arg] callback wrappers are not
229 prototyped to have "const" restrictions on the object pointers they are
230 given (and so aren't required to cast them away any more).
233 *) The tmdiff.h API was so ugly and minimal that our own timing utility
234 (speed) prefers to use its own implementation. The two implementations
235 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
236 its object type properly exposed (MS_TM) instead of casting to/from "char
237 *". This may still change yet if someone realises MS_TM and "ms_time_***"
238 aren't necessarily the greatest nomenclatures - but this is what was used
239 internally to the implementation so I've used that for now.
242 *) Ensure that deprecated functions do not get compiled when
243 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
244 the self-tests were still using deprecated key-generation functions so
245 these have been updated also.
248 *) Reorganise PKCS#7 code to separate the digest location functionality
249 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
250 New function PKCS7_set_digest() to set the digest type for PKCS#7
251 digestedData type. Add additional code to correctly generate the
252 digestedData type and add support for this type in PKCS7 initialization
256 *) New function PKCS7_set0_type_other() this initializes a PKCS7
257 structure of type "other".
260 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
261 sure the loop does correctly stop and breaking ("division by zero")
262 modulus operations are not performed. The (pre-generated) prime
263 table crypto/bn/bn_prime.h was already correct, but it could not be
264 re-generated on some platforms because of the "division by zero"
265 situation in the script.
266 [Ralf S. Engelschall]
268 *) Update support for ECC-based TLS ciphersuites according to
269 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
270 SHA-1 now is only used for "small" curves (where the
271 representation of a field element takes up to 24 bytes); for
272 larger curves, the field element resulting from ECDH is directly
273 used as premaster secret.
274 [Douglas Stebila (Sun Microsystems Laboratories)]
276 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
277 curve secp160r1 to the tests.
278 [Douglas Stebila (Sun Microsystems Laboratories)]
280 *) Add the possibility to load symbols globally with DSO.
281 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
283 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
284 control of the error stack.
287 *) Add support for STORE in ENGINE.
290 *) Add the STORE type. The intention is to provide a common interface
291 to certificate and key stores, be they simple file-based stores, or
292 HSM-type store, or LDAP stores, or...
293 NOTE: The code is currently UNTESTED and isn't really used anywhere.
296 *) Add a generic structure called OPENSSL_ITEM. This can be used to
297 pass a list of arguments to any function as well as provide a way
298 for a function to pass data back to the caller.
301 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
302 works like BUF_strdup() but can be used to duplicate a portion of
303 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
307 *) Add the function sk_find_ex() which works like sk_find(), but will
308 return an index to an element even if an exact match couldn't be
309 found. The index is guaranteed to point at the element where the
310 searched-for key would be inserted to preserve sorting order.
313 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
314 takes an extra flags argument for optional functionality. Currently,
315 the following flags are defined:
317 OBJ_BSEARCH_VALUE_ON_NOMATCH
318 This one gets OBJ_bsearch_ex() to return a pointer to the first
319 element where the comparing function returns a negative or zero
322 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
323 This one gets OBJ_bsearch_ex() to return a pointer to the first
324 element where the comparing function returns zero. This is useful
325 if there are more than one element where the comparing function
329 *) Make it possible to create self-signed certificates with 'openssl ca'
330 in such a way that the self-signed certificate becomes part of the
331 CA database and uses the same mechanisms for serial number generation
332 as all other certificate signing. The new flag '-selfsign' enables
333 this functionality. Adapt CA.sh and CA.pl.in.
336 *) Add functionality to check the public key of a certificate request
337 against a given private. This is useful to check that a certificate
338 request can be signed by that key (self-signing).
341 *) Make it possible to have multiple active certificates with the same
342 subject in the CA index file. This is done only if the keyword
343 'unique_subject' is set to 'no' in the main CA section (default
344 if 'CA_default') of the configuration file. The value is saved
345 with the database itself in a separate index attribute file,
346 named like the index file with '.attr' appended to the name.
349 *) Generate muti valued AVAs using '+' notation in config files for
353 *) Support for nameConstraints certificate extension.
356 *) Support for policyConstraints certificate extension.
359 *) Support for policyMappings certificate extension.
362 *) Make sure the default DSA_METHOD implementation only uses its
363 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
364 and change its own handlers to be NULL so as to remove unnecessary
365 indirection. This lets alternative implementations fallback to the
366 default implementation more easily.
369 *) Support for directoryName in GeneralName related extensions
373 *) Make it possible to link applications using Makefile.shared.
374 Make that possible even when linking against static libraries!
377 *) Support for single pass processing for S/MIME signing. This now
378 means that S/MIME signing can be done from a pipe, in addition
379 cleartext signing (multipart/signed type) is effectively streaming
380 and the signed data does not need to be all held in memory.
382 This is done with a new flag PKCS7_STREAM. When this flag is set
383 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
384 is done after the data is output (and digests calculated) in
388 *) Add full support for -rpath/-R, both in shared libraries and
389 applications, at least on the platforms where it's known how
393 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
394 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
395 will now compute a table of multiples of the generator that
396 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
397 faster (notably in the case of a single point multiplication,
399 [Nils Larsch, Bodo Moeller]
401 *) IPv6 support for certificate extensions. The various extensions
402 which use the IP:a.b.c.d can now take IPv6 addresses using the
403 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
407 *) Added an ENGINE that implements RSA by performing private key
408 exponentiations with the GMP library. The conversions to and from
409 GMP's mpz_t format aren't optimised nor are any montgomery forms
410 cached, and on x86 it appears OpenSSL's own performance has caught up.
411 However there are likely to be other architectures where GMP could
412 provide a boost. This ENGINE is not built in by default, but it can be
413 specified at Configure time and should be accompanied by the necessary
414 linker additions, eg;
415 ./config -DOPENSSL_USE_GMP -lgmp
418 *) "openssl engine" will not display ENGINE/DSO load failure errors when
419 testing availability of engines with "-t" - the old behaviour is
420 produced by increasing the feature's verbosity with "-tt".
423 *) ECDSA routines: under certain error conditions uninitialized BN objects
424 could be freed. Solution: make sure initialization is performed early
425 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
429 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
430 and DH_METHOD (eg. by ENGINE implementations) to override the normal
431 software implementations. For DSA and DH, parameter generation can
432 also be overriden by providing the appropriate method callbacks.
435 *) Change the "progress" mechanism used in key-generation and
436 primality testing to functions that take a new BN_GENCB pointer in
437 place of callback/argument pairs. The new API functions have "_ex"
438 postfixes and the older functions are reimplemented as wrappers for
439 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
440 declarations of the old functions to help (graceful) attempts to
441 migrate to the new functions. Also, the new key-generation API
442 functions operate on a caller-supplied key-structure and return
443 success/failure rather than returning a key or NULL - this is to
444 help make "keygen" another member function of RSA_METHOD etc.
446 Example for using the new callback interface:
448 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
452 BN_GENCB_set(&my_cb, my_callback, my_arg);
454 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
455 /* For the meaning of a, b in calls to my_callback(), see the
456 * documentation of the function that calls the callback.
457 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
458 * my_callback should return 1 if it wants BN_is_prime_ex()
459 * to continue, or 0 to stop.
464 *) Change the ZLIB compression method to be stateful, and make it
465 available to TLS with the number defined in
466 draft-ietf-tls-compression-04.txt.
469 *) Add the ASN.1 structures and functions for CertificatePair, which
470 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
472 CertificatePair ::= SEQUENCE {
473 forward [0] Certificate OPTIONAL,
474 reverse [1] Certificate OPTIONAL,
475 -- at least one of the pair shall be present -- }
477 Also implement the PEM functions to read and write certificate
478 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
480 This needed to be defined, mostly for the sake of the LDAP
481 attribute crossCertificatePair, but may prove useful elsewhere as
485 *) Make it possible to inhibit symlinking of shared libraries in
486 Makefile.shared, for Cygwin's sake.
489 *) Extend the BIGNUM API by creating a function
490 void BN_set_negative(BIGNUM *a, int neg);
491 and a macro that behave like
492 int BN_is_negative(const BIGNUM *a);
494 to avoid the need to access 'a->neg' directly in applications.
497 *) Implement fast modular reduction for pseudo-Mersenne primes
498 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
499 EC_GROUP_new_curve_GFp() will now automatically use this
501 [Nils Larsch <nla@trustcenter.de>]
503 *) Add new lock type (CRYPTO_LOCK_BN).
506 *) Change the ENGINE framework to automatically load engines
507 dynamically from specific directories unless they could be
508 found to already be built in or loaded. Move all the
509 current engines except for the cryptodev one to a new
511 The engines in engines/ are built as shared libraries if
512 the "shared" options was given to ./Configure or ./config.
513 Otherwise, they are inserted in libcrypto.a.
514 /usr/local/ssl/engines is the default directory for dynamic
515 engines, but that can be overriden at configure time through
516 the usual use of --prefix and/or --openssldir, and at run
517 time with the environment variable OPENSSL_ENGINES.
518 [Geoff Thorpe and Richard Levitte]
520 *) Add Makefile.shared, a helper makefile to build shared
521 libraries. Addapt Makefile.org.
524 *) Add version info to Win32 DLLs.
525 [Peter 'Luna' Runestig" <peter@runestig.com>]
527 *) Add new 'medium level' PKCS#12 API. Certificates and keys
528 can be added using this API to created arbitrary PKCS#12
529 files while avoiding the low level API.
531 New options to PKCS12_create(), key or cert can be NULL and
532 will then be omitted from the output file. The encryption
533 algorithm NIDs can be set to -1 for no encryption, the mac
534 iteration count can be set to 0 to omit the mac.
536 Enhance pkcs12 utility by making the -nokeys and -nocerts
537 options work when creating a PKCS#12 file. New option -nomac
538 to omit the mac, NONE can be set for an encryption algorithm.
539 New code is modified to use the enhanced PKCS12_create()
540 instead of the low level API.
543 *) Extend ASN1 encoder to support indefinite length constructed
544 encoding. This can output sequences tags and octet strings in
545 this form. Modify pk7_asn1.c to support indefinite length
546 encoding. This is experimental and needs additional code to
547 be useful, such as an ASN1 bio and some enhanced streaming
550 Extend template encode functionality so that tagging is passed
551 down to the template encoder.
554 *) Let 'openssl req' fail if an argument to '-newkey' is not
555 recognized instead of using RSA as a default.
558 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
559 As these are not official, they are not included in "ALL";
560 the "ECCdraft" ciphersuite group alias can be used to select them.
561 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
563 *) Add ECDH engine support.
564 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
566 *) Add ECDH in new directory crypto/ecdh/.
567 [Douglas Stebila (Sun Microsystems Laboratories)]
569 *) Let BN_rand_range() abort with an error after 100 iterations
570 without success (which indicates a broken PRNG).
573 *) Change BN_mod_sqrt() so that it verifies that the input value
574 is really the square of the return value. (Previously,
575 BN_mod_sqrt would show GIGO behaviour.)
578 *) Add named elliptic curves over binary fields from X9.62, SECG,
579 and WAP/WTLS; add OIDs that were still missing.
581 [Sheueling Chang Shantz and Douglas Stebila
582 (Sun Microsystems Laboratories)]
584 *) Extend the EC library for elliptic curves over binary fields
585 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
588 EC_GF2m_simple_method
592 EC_GROUP_new_curve_GF2m
593 EC_GROUP_set_curve_GF2m
594 EC_GROUP_get_curve_GF2m
595 EC_POINT_set_affine_coordinates_GF2m
596 EC_POINT_get_affine_coordinates_GF2m
597 EC_POINT_set_compressed_coordinates_GF2m
599 Point compression for binary fields is disabled by default for
600 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
603 As binary polynomials are represented as BIGNUMs, various members
604 of the EC_GROUP and EC_POINT data structures can be shared
605 between the implementations for prime fields and binary fields;
606 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
607 are essentially identical to their ..._GFp counterparts.
608 (For simplicity, the '..._GFp' prefix has been dropped from
609 various internal method names.)
611 An internal 'field_div' method (similar to 'field_mul' and
612 'field_sqr') has been added; this is used only for binary fields.
614 [Sheueling Chang Shantz and Douglas Stebila
615 (Sun Microsystems Laboratories)]
617 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
618 through methods ('mul', 'precompute_mult').
620 The generic implementations (now internally called 'ec_wNAF_mul'
621 and 'ec_wNAF_precomputed_mult') remain the default if these
622 methods are undefined.
624 [Sheueling Chang Shantz and Douglas Stebila
625 (Sun Microsystems Laboratories)]
627 *) New function EC_GROUP_get_degree, which is defined through
628 EC_METHOD. For curves over prime fields, this returns the bit
629 length of the modulus.
631 [Sheueling Chang Shantz and Douglas Stebila
632 (Sun Microsystems Laboratories)]
634 *) New functions EC_GROUP_dup, EC_POINT_dup.
635 (These simply call ..._new and ..._copy).
637 [Sheueling Chang Shantz and Douglas Stebila
638 (Sun Microsystems Laboratories)]
640 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
641 Polynomials are represented as BIGNUMs (where the sign bit is not
642 used) in the following functions [macros]:
645 BN_GF2m_sub [= BN_GF2m_add]
646 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
647 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
648 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
650 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
651 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
652 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
653 BN_GF2m_cmp [= BN_ucmp]
655 (Note that only the 'mod' functions are actually for fields GF(2^m).
656 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
658 For some functions, an the irreducible polynomial defining a
659 field can be given as an 'unsigned int[]' with strictly
660 decreasing elements giving the indices of those bits that are set;
661 i.e., p[] represents the polynomial
662 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
664 p[0] > p[1] > ... > p[k] = 0.
665 This applies to the following functions:
670 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
671 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
674 BN_GF2m_mod_solve_quad_arr
678 Conversion can be performed by the following functions:
683 bntest.c has additional tests for binary polynomial arithmetic.
685 Two implementations for BN_GF2m_mod_div() are available.
686 The default algorithm simply uses BN_GF2m_mod_inv() and
687 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
688 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
689 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
691 [Sheueling Chang Shantz and Douglas Stebila
692 (Sun Microsystems Laboratories)]
694 *) Add new error code 'ERR_R_DISABLED' that can be used when some
695 functionality is disabled at compile-time.
696 [Douglas Stebila <douglas.stebila@sun.com>]
698 *) Change default behaviour of 'openssl asn1parse' so that more
699 information is visible when viewing, e.g., a certificate:
701 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
702 mode the content of non-printable OCTET STRINGs is output in a
703 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
704 avoid the appearance of a printable string.
705 [Nils Larsch <nla@trustcenter.de>]
707 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
709 EC_GROUP_set_asn1_flag()
710 EC_GROUP_get_asn1_flag()
711 EC_GROUP_set_point_conversion_form()
712 EC_GROUP_get_point_conversion_form()
713 These control ASN1 encoding details:
714 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
715 has been set to OPENSSL_EC_NAMED_CURVE.
716 - Points are encoded in uncompressed form by default; options for
717 asn1_for are as for point2oct, namely
718 POINT_CONVERSION_COMPRESSED
719 POINT_CONVERSION_UNCOMPRESSED
720 POINT_CONVERSION_HYBRID
722 Also add 'seed' and 'seed_len' members to EC_GROUP with access
726 EC_GROUP_get_seed_len()
727 This is used only for ASN1 purposes (so far).
728 [Nils Larsch <nla@trustcenter.de>]
730 *) Add 'field_type' member to EC_METHOD, which holds the NID
731 of the appropriate field type OID. The new function
732 EC_METHOD_get_field_type() returns this value.
733 [Nils Larsch <nla@trustcenter.de>]
740 providing useful interfaces to EC_POINT_point2oct() and
741 EC_POINT_oct2point().
742 [Nils Larsch <nla@trustcenter.de>]
744 *) Change internals of the EC library so that the functions
745 EC_GROUP_set_generator()
746 EC_GROUP_get_generator()
748 EC_GROUP_get_cofactor()
749 are implemented directly in crypto/ec/ec_lib.c and not dispatched
750 to methods, which would lead to unnecessary code duplication when
751 adding different types of curves.
752 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
754 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
755 arithmetic, and such that modified wNAFs are generated
756 (which avoid length expansion in many cases).
759 *) Add a function EC_GROUP_check_discriminant() (defined via
760 EC_METHOD) that verifies that the curve discriminant is non-zero.
762 Add a function EC_GROUP_check() that makes some sanity tests
763 on a EC_GROUP, its generator and order. This includes
764 EC_GROUP_check_discriminant().
765 [Nils Larsch <nla@trustcenter.de>]
767 *) Add ECDSA in new directory crypto/ecdsa/.
769 Add applications 'openssl ecparam' and 'openssl ecdsa'
770 (these are based on 'openssl dsaparam' and 'openssl dsa').
772 ECDSA support is also included in various other files across the
773 library. Most notably,
774 - 'openssl req' now has a '-newkey ecdsa:file' option;
775 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
776 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
777 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
778 them suitable for ECDSA where domain parameters must be
779 extracted before the specific public key;
780 - ECDSA engine support has been added.
781 [Nils Larsch <nla@trustcenter.de>]
783 *) Include some named elliptic curves, and add OIDs from X9.62,
784 SECG, and WAP/WTLS. Each curve can be obtained from the new
786 EC_GROUP_new_by_curve_name(),
787 and the list of available named curves can be obtained with
788 EC_get_builtin_curves().
789 Also add a 'curve_name' member to EC_GROUP objects, which can be
791 EC_GROUP_set_curve_name()
792 EC_GROUP_get_curve_name()
793 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
795 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
796 was actually never needed) and in BN_mul(). The removal in BN_mul()
797 required a small change in bn_mul_part_recursive() and the addition
798 of the functions bn_cmp_part_words(), bn_sub_part_words() and
799 bn_add_part_words(), which do the same thing as bn_cmp_words(),
800 bn_sub_words() and bn_add_words() except they take arrays with
804 Changes between 0.9.7g and 0.9.7h [XX xxx XXXX]
806 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
807 the exponentiation using a fixed-length exponent. (Otherwise,
808 the information leaked through timing could expose the secret key
809 after many signatures; cf. Bleichenbacher's attack on DSA with
813 *) Make a new fixed-window mod_exp implementation the default for
814 RSA, DSA, and DH private-key operations so that the sequence of
815 squares and multiplies and the memory access pattern are
816 independent of the particular secret key. This will mitigate
817 cache-timing and potential related attacks.
819 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
820 and this is automatically used by BN_mod_exp_mont() if the new flag
821 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
822 will use this BN flag for private exponents unless the flag
823 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
824 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
826 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
828 *) Change the client implementation for SSLv23_method() and
829 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
830 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
831 (Previously, the SSL 2.0 backwards compatible Client Hello
832 message format would be used even with SSL_OP_NO_SSLv2.)
835 *) Add support for smime-type MIME parameter in S/MIME messages which some
839 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
840 a threadsafe manner. Modify rsa code to use new function and add calls
841 to dsa and dh code (which had race conditions before).
844 *) Include the fixed error library code in the C error file definitions
845 instead of fixing them up at runtime. This keeps the error code
849 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
851 *) Fixes for newer kerberos headers. NB: the casts are needed because
852 the 'length' field is signed on one version and unsigned on another
853 with no (?) obvious way to tell the difference, without these VC++
854 complains. Also the "definition" of FAR (blank) is no longer included
855 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
856 some needed definitions.
859 *) Undo Cygwin change.
862 *) Added support for proxy certificates according to RFC 3820.
863 Because they may be a security thread to unaware applications,
864 they must be explicitely allowed in run-time. See
865 docs/HOWTO/proxy_certificates.txt for further information.
868 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
870 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
871 server and client random values. Previously
872 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
873 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
875 This change has negligible security impact because:
877 1. Server and client random values still have 24 bytes of pseudo random
880 2. Server and client random values are sent in the clear in the initial
883 3. The master secret is derived using the premaster secret (48 bytes in
884 size for static RSA ciphersuites) as well as client server and random
887 The OpenSSL team would like to thank the UK NISCC for bringing this issue
890 [Stephen Henson, reported by UK NISCC]
892 *) Use Windows randomness collection on Cygwin.
895 *) Fix hang in EGD/PRNGD query when communication socket is closed
896 prematurely by EGD/PRNGD.
897 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
899 *) Prompt for pass phrases when appropriate for PKCS12 input format.
902 *) Back-port of selected performance improvements from development
903 branch, as well as improved support for PowerPC platforms.
906 *) Add lots of checks for memory allocation failure, error codes to indicate
907 failure and freeing up memory if a failure occurs.
908 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
910 *) Add new -passin argument to dgst.
913 *) Perform some character comparisons of different types in X509_NAME_cmp:
914 this is needed for some certificates that reencode DNs into UTF8Strings
915 (in violation of RFC3280) and can't or wont issue name rollover
919 *) Make an explicit check during certificate validation to see that
920 the CA setting in each certificate on the chain is correct. As a
921 side effect always do the following basic checks on extensions,
922 not just when there's an associated purpose to the check:
924 - if there is an unhandled critical extension (unless the user
925 has chosen to ignore this fault)
926 - if the path length has been exceeded (if one is set at all)
927 - that certain extensions fit the associated purpose (if one has
931 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
933 *) Avoid a race condition when CRLs are checked in a multi threaded
934 environment. This would happen due to the reordering of the revoked
935 entries during signature checking and serial number lookup. Now the
936 encoding is cached and the serial number sort performed under a lock.
937 Add new STACK function sk_is_sorted().
940 *) Add Delta CRL to the extension code.
943 *) Various fixes to s3_pkt.c so alerts are sent properly.
944 [David Holmes <d.holmes@f5.com>]
946 *) Reduce the chances of duplicate issuer name and serial numbers (in
947 violation of RFC3280) using the OpenSSL certificate creation utilities.
948 This is done by creating a random 64 bit value for the initial serial
949 number when a serial number file is created or when a self signed
950 certificate is created using 'openssl req -x509'. The initial serial
951 number file is created using 'openssl x509 -next_serial' in CA.pl
952 rather than being initialized to 1.
955 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
957 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
958 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
959 [Joe Orton, Steve Henson]
961 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
963 [Joe Orton, Steve Henson]
965 *) Make it possible to have multiple active certificates with the same
966 subject in the CA index file. This is done only if the keyword
967 'unique_subject' is set to 'no' in the main CA section (default
968 if 'CA_default') of the configuration file. The value is saved
969 with the database itself in a separate index attribute file,
970 named like the index file with '.attr' appended to the name.
973 *) X509 verify fixes. Disable broken certificate workarounds when
974 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
975 keyUsage extension present. Don't accept CRLs with unhandled critical
976 extensions: since verify currently doesn't process CRL extensions this
977 rejects a CRL with *any* critical extensions. Add new verify error codes
981 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
982 A clarification of RFC2560 will require the use of OCTET STRINGs and
983 some implementations cannot handle the current raw format. Since OpenSSL
984 copies and compares OCSP nonces as opaque blobs without any attempt at
985 parsing them this should not create any compatibility issues.
988 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
989 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
990 this HMAC (and other) operations are several times slower than OpenSSL
994 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
995 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
997 *) Use the correct content when signing type "other".
1000 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1002 *) Fix various bugs revealed by running the NISCC test suite:
1004 Stop out of bounds reads in the ASN1 code when presented with
1005 invalid tags (CAN-2003-0543 and CAN-2003-0544).
1007 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
1009 If verify callback ignores invalid public key errors don't try to check
1010 certificate signature with the NULL public key.
1014 *) New -ignore_err option in ocsp application to stop the server
1015 exiting on the first error in a request.
1018 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1019 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1023 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1024 extra data after the compression methods not only for TLS 1.0
1025 but also for SSL 3.0 (as required by the specification).
1026 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1028 *) Change X509_certificate_type() to mark the key as exported/exportable
1029 when it's 512 *bits* long, not 512 bytes.
1032 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1033 blocks during encryption.
1036 *) Various fixes to base64 BIO and non blocking I/O. On write
1037 flushes were not handled properly if the BIO retried. On read
1038 data was not being buffered properly and had various logic bugs.
1039 This also affects blocking I/O when the data being decoded is a
1043 *) Various S/MIME bugfixes and compatibility changes:
1044 output correct application/pkcs7 MIME type if
1045 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1046 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1047 of files as .eml work). Correctly handle very long lines in MIME
1051 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1053 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1054 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1055 a protocol version number mismatch like a decryption error
1056 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1059 *) Turn on RSA blinding by default in the default implementation
1060 to avoid a timing attack. Applications that don't want it can call
1061 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1062 They would be ill-advised to do so in most cases.
1063 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1065 *) Change RSA blinding code so that it works when the PRNG is not
1066 seeded (in this case, the secret RSA exponent is abused as
1067 an unpredictable seed -- if it is not unpredictable, there
1068 is no point in blinding anyway). Make RSA blinding thread-safe
1069 by remembering the creator's thread ID in rsa->blinding and
1070 having all other threads use local one-time blinding factors
1071 (this requires more computation than sharing rsa->blinding, but
1072 avoids excessive locking; and if an RSA object is not shared
1073 between threads, blinding will still be very fast).
1076 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1077 ENGINE as defaults for all supported algorithms irrespective of
1078 the 'flags' parameter. 'flags' is now honoured, so applications
1079 should make sure they are passing it correctly.
1082 *) Target "mingw" now allows native Windows code to be generated in
1083 the Cygwin environment as well as with the MinGW compiler.
1086 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1088 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1089 via timing by performing a MAC computation even if incorrrect
1090 block cipher padding has been found. This is a countermeasure
1091 against active attacks where the attacker has to distinguish
1092 between bad padding and a MAC verification error. (CAN-2003-0078)
1094 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1095 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1096 Martin Vuagnoux (EPFL, Ilion)]
1098 *) Make the no-err option work as intended. The intention with no-err
1099 is not to have the whole error stack handling routines removed from
1100 libcrypto, it's only intended to remove all the function name and
1101 reason texts, thereby removing some of the footprint that may not
1102 be interesting if those errors aren't displayed anyway.
1104 NOTE: it's still possible for any application or module to have it's
1105 own set of error texts inserted. The routines are there, just not
1106 used by default when no-err is given.
1109 *) Add support for FreeBSD on IA64.
1110 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1112 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1113 Kerberos function mit_des_cbc_cksum(). Before this change,
1114 the value returned by DES_cbc_cksum() was like the one from
1115 mit_des_cbc_cksum(), except the bytes were swapped.
1116 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1118 *) Allow an application to disable the automatic SSL chain building.
1119 Before this a rather primitive chain build was always performed in
1120 ssl3_output_cert_chain(): an application had no way to send the
1121 correct chain if the automatic operation produced an incorrect result.
1123 Now the chain builder is disabled if either:
1125 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1127 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1129 The reasoning behind this is that an application would not want the
1130 auto chain building to take place if extra chain certificates are
1131 present and it might also want a means of sending no additional
1132 certificates (for example the chain has two certificates and the
1136 *) Add the possibility to build without the ENGINE framework.
1137 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1139 *) Under Win32 gmtime() can return NULL: check return value in
1140 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1143 *) DSA routines: under certain error conditions uninitialized BN objects
1144 could be freed. Solution: make sure initialization is performed early
1145 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1146 Nils Larsch <nla@trustcenter.de> via PR#459)
1149 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1150 checked on reconnect on the client side, therefore session resumption
1151 could still fail with a "ssl session id is different" error. This
1152 behaviour is masked when SSL_OP_ALL is used due to
1153 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1154 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1155 followup to PR #377.
1158 *) IA-32 assembler support enhancements: unified ELF targets, support
1159 for SCO/Caldera platforms, fix for Cygwin shared build.
1162 *) Add support for FreeBSD on sparc64. As a consequence, support for
1163 FreeBSD on non-x86 processors is separate from x86 processors on
1164 the config script, much like the NetBSD support.
1165 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1167 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1169 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1172 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1173 code (06) was taken as the first octet of the session ID and the last
1174 octet was ignored consequently. As a result SSLv2 client side session
1175 caching could not have worked due to the session ID mismatch between
1177 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1181 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1182 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1186 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1187 seems that in spite of existing for more than a year, many application
1188 author have done nothing to provide the necessary callbacks, which
1189 means that this particular engine will not work properly anywhere.
1190 This is a very unfortunate situation which forces us, in the name
1191 of usability, to give the hw_ncipher.c a static lock, which is part
1193 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1194 appear in 0.9.8 or later. We EXPECT application authors to have
1195 dealt properly with this when 0.9.8 is released (unless we actually
1196 make such changes in the libcrypto locking code that changes will
1197 have to be made anyway).
1200 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1201 octets have been read, EOF or an error occurs. Without this change
1202 some truncated ASN1 structures will not produce an error.
1205 *) Disable Heimdal support, since it hasn't been fully implemented.
1206 Still give the possibility to force the use of Heimdal, but with
1207 warnings and a request that patches get sent to openssl-dev.
1210 *) Add the VC-CE target, introduce the WINCE sysname, and add
1211 INSTALL.WCE and appropriate conditionals to make it build.
1212 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1214 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1215 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1216 edit numbers of the version.
1217 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1219 *) Introduce safe string copy and catenation functions
1220 (BUF_strlcpy() and BUF_strlcat()).
1221 [Ben Laurie (CHATS) and Richard Levitte]
1223 *) Avoid using fixed-size buffers for one-line DNs.
1224 [Ben Laurie (CHATS)]
1226 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1227 resizing buffers containing secrets, and use where appropriate.
1228 [Ben Laurie (CHATS)]
1230 *) Avoid using fixed size buffers for configuration file location.
1231 [Ben Laurie (CHATS)]
1233 *) Avoid filename truncation for various CA files.
1234 [Ben Laurie (CHATS)]
1236 *) Use sizeof in preference to magic numbers.
1237 [Ben Laurie (CHATS)]
1239 *) Avoid filename truncation in cert requests.
1240 [Ben Laurie (CHATS)]
1242 *) Add assertions to check for (supposedly impossible) buffer
1244 [Ben Laurie (CHATS)]
1246 *) Don't cache truncated DNS entries in the local cache (this could
1247 potentially lead to a spoofing attack).
1248 [Ben Laurie (CHATS)]
1250 *) Fix various buffers to be large enough for hex/decimal
1251 representations in a platform independent manner.
1252 [Ben Laurie (CHATS)]
1254 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1255 resizing buffers containing secrets, and use where appropriate.
1256 [Ben Laurie (CHATS)]
1258 *) Add BIO_indent() to avoid much slightly worrying code to do
1260 [Ben Laurie (CHATS)]
1262 *) Convert sprintf()/BIO_puts() to BIO_printf().
1263 [Ben Laurie (CHATS)]
1265 *) buffer_gets() could terminate with the buffer only half
1267 [Ben Laurie (CHATS)]
1269 *) Add assertions to prevent user-supplied crypto functions from
1270 overflowing internal buffers by having large block sizes, etc.
1271 [Ben Laurie (CHATS)]
1273 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1275 [Ben Laurie (CHATS)]
1277 *) Eliminate unused copy of key in RC4.
1278 [Ben Laurie (CHATS)]
1280 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1281 [Ben Laurie (CHATS)]
1283 *) Fix off-by-one error in EGD path.
1284 [Ben Laurie (CHATS)]
1286 *) If RANDFILE path is too long, ignore instead of truncating.
1287 [Ben Laurie (CHATS)]
1289 *) Eliminate unused and incorrectly sized X.509 structure
1291 [Ben Laurie (CHATS)]
1293 *) Eliminate unused and dangerous function knumber().
1294 [Ben Laurie (CHATS)]
1296 *) Eliminate unused and dangerous structure, KSSL_ERR.
1297 [Ben Laurie (CHATS)]
1299 *) Protect against overlong session ID context length in an encoded
1300 session object. Since these are local, this does not appear to be
1302 [Ben Laurie (CHATS)]
1304 *) Change from security patch (see 0.9.6e below) that did not affect
1305 the 0.9.6 release series:
1307 Remote buffer overflow in SSL3 protocol - an attacker could
1308 supply an oversized master key in Kerberos-enabled versions.
1310 [Ben Laurie (CHATS)]
1312 *) Change the SSL kerb5 codes to match RFC 2712.
1315 *) Make -nameopt work fully for req and add -reqopt switch.
1316 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1318 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1319 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1321 *) Make sure tests can be performed even if the corresponding algorithms
1322 have been removed entirely. This was also the last step to make
1323 OpenSSL compilable with DJGPP under all reasonable conditions.
1324 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1326 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1327 to allow version independent disabling of normally unselected ciphers,
1328 which may be activated as a side-effect of selecting a single cipher.
1330 (E.g., cipher list string "RSA" enables ciphersuites that are left
1331 out of "ALL" because they do not provide symmetric encryption.
1332 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1333 [Lutz Jaenicke, Bodo Moeller]
1335 *) Add appropriate support for separate platform-dependent build
1336 directories. The recommended way to make a platform-dependent
1337 build directory is the following (tested on Linux), maybe with
1340 # Place yourself outside of the OpenSSL source tree. In
1341 # this example, the environment variable OPENSSL_SOURCE
1342 # is assumed to contain the absolute OpenSSL source directory.
1343 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1344 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1345 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1346 mkdir -p `dirname $F`
1347 ln -s $OPENSSL_SOURCE/$F $F
1350 To be absolutely sure not to disturb the source tree, a "make clean"
1351 is a good thing. If it isn't successfull, don't worry about it,
1352 it probably means the source directory is very clean.
1355 *) Make sure any ENGINE control commands make local copies of string
1356 pointers passed to them whenever necessary. Otherwise it is possible
1357 the caller may have overwritten (or deallocated) the original string
1358 data when a later ENGINE operation tries to use the stored values.
1359 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1361 *) Improve diagnostics in file reading and command-line digests.
1362 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1364 *) Add AES modes CFB and OFB to the object database. Correct an
1365 error in AES-CFB decryption.
1368 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1369 allows existing EVP_CIPHER_CTX structures to be reused after
1370 calling EVP_*Final(). This behaviour is used by encryption
1371 BIOs and some applications. This has the side effect that
1372 applications must explicitly clean up cipher contexts with
1373 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1376 *) Check the values of dna and dnb in bn_mul_recursive before calling
1377 bn_mul_comba (a non zero value means the a or b arrays do not contain
1378 n2 elements) and fallback to bn_mul_normal if either is not zero.
1381 *) Fix escaping of non-ASCII characters when using the -subj option
1382 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1385 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1386 form for "surname", serialNumber has no short form.
1387 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1388 therefore remove "mail" short name for "internet 7".
1389 The OID for unique identifiers in X509 certificates is
1390 x500UniqueIdentifier, not uniqueIdentifier.
1391 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1394 *) Add an "init" command to the ENGINE config module and auto initialize
1395 ENGINEs. Without any "init" command the ENGINE will be initialized
1396 after all ctrl commands have been executed on it. If init=1 the
1397 ENGINE is initailized at that point (ctrls before that point are run
1398 on the uninitialized ENGINE and after on the initialized one). If
1399 init=0 then the ENGINE will not be iniatialized at all.
1402 *) Fix the 'app_verify_callback' interface so that the user-defined
1403 argument is actually passed to the callback: In the
1404 SSL_CTX_set_cert_verify_callback() prototype, the callback
1405 declaration has been changed from
1408 int (*cb)(X509_STORE_CTX *,void *);
1409 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1410 i=s->ctx->app_verify_callback(&ctx)
1411 has been changed into
1412 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1414 To update applications using SSL_CTX_set_cert_verify_callback(),
1415 a dummy argument can be added to their callback functions.
1416 [D. K. Smetters <smetters@parc.xerox.com>]
1418 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1419 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1421 *) Add and OPENSSL_LOAD_CONF define which will cause
1422 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1423 This allows older applications to transparently support certain
1424 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1425 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1426 load the config file and OPENSSL_add_all_algorithms_conf() which will
1427 always load it have also been added.
1430 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1431 Adjust NIDs and EVP layer.
1432 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1434 *) Config modules support in openssl utility.
1436 Most commands now load modules from the config file,
1437 though in a few (such as version) this isn't done
1438 because it couldn't be used for anything.
1440 In the case of ca and req the config file used is
1441 the same as the utility itself: that is the -config
1442 command line option can be used to specify an
1446 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1447 use "openssl_conf" if filename is NULL use default openssl config file.
1450 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1451 config section name. Add a new flag to tolerate a missing config file
1452 and move code to CONF_modules_load_file().
1455 *) Support for crypto accelerator cards from Accelerated Encryption
1456 Processing, www.aep.ie. (Use engine 'aep')
1457 The support was copied from 0.9.6c [engine] and adapted/corrected
1458 to work with the new engine framework.
1459 [AEP Inc. and Richard Levitte]
1461 *) Support for SureWare crypto accelerator cards from Baltimore
1462 Technologies. (Use engine 'sureware')
1463 The support was copied from 0.9.6c [engine] and adapted
1464 to work with the new engine framework.
1467 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1468 make the newer ENGINE framework commands for the CHIL engine work.
1469 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1471 *) Make it possible to produce shared libraries on ReliantUNIX.
1472 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1474 *) Add the configuration target debug-linux-ppro.
1475 Make 'openssl rsa' use the general key loading routines
1476 implemented in apps.c, and make those routines able to
1477 handle the key format FORMAT_NETSCAPE and the variant
1479 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1481 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1482 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1484 *) Add -keyform to rsautl, and document -engine.
1485 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1487 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1488 BIO_R_NO_SUCH_FILE error code rather than the generic
1489 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1492 *) Add new functions
1494 ERR_peek_last_error_line
1495 ERR_peek_last_error_line_data.
1496 These are similar to
1499 ERR_peek_error_line_data,
1500 but report on the latest error recorded rather than the first one
1501 still in the error queue.
1502 [Ben Laurie, Bodo Moeller]
1504 *) default_algorithms option in ENGINE config module. This allows things
1506 default_algorithms = ALL
1507 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1510 *) Prelminary ENGINE config module.
1513 *) New experimental application configuration code.
1516 *) Change the AES code to follow the same name structure as all other
1517 symmetric ciphers, and behave the same way. Move everything to
1518 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1519 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1521 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1522 [Ben Laurie and Theo de Raadt]
1524 *) Add option to output public keys in req command.
1525 [Massimiliano Pala madwolf@openca.org]
1527 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1528 (up to about 10% better than before for P-192 and P-224).
1531 *) New functions/macros
1533 SSL_CTX_set_msg_callback(ctx, cb)
1534 SSL_CTX_set_msg_callback_arg(ctx, arg)
1535 SSL_set_msg_callback(ssl, cb)
1536 SSL_set_msg_callback_arg(ssl, arg)
1538 to request calling a callback function
1540 void cb(int write_p, int version, int content_type,
1541 const void *buf, size_t len, SSL *ssl, void *arg)
1543 whenever a protocol message has been completely received
1544 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1545 protocol version according to which the SSL library interprets
1546 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1547 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1548 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1549 specification (change_cipher_spec(20), alert(21), handshake(22)).
1550 'buf' and 'len' point to the actual message, 'ssl' to the
1551 SSL object, and 'arg' is the application-defined value set by
1552 SSL[_CTX]_set_msg_callback_arg().
1554 'openssl s_client' and 'openssl s_server' have new '-msg' options
1555 to enable a callback that displays all protocol messages.
1558 *) Change the shared library support so shared libraries are built as
1559 soon as the corresponding static library is finished, and thereby get
1560 openssl and the test programs linked against the shared library.
1561 This still only happens when the keyword "shard" has been given to
1562 the configuration scripts.
1564 NOTE: shared library support is still an experimental thing, and
1565 backward binary compatibility is still not guaranteed.
1566 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1568 *) Add support for Subject Information Access extension.
1569 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1571 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1572 additional bytes when new memory had to be allocated, not just
1573 when reusing an existing buffer.
1576 *) New command line and configuration option 'utf8' for the req command.
1577 This allows field values to be specified as UTF8 strings.
1580 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1581 runs for the former and machine-readable output for the latter.
1584 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1585 of the e-mail address in the DN (i.e., it will go into a certificate
1586 extension only). The new configuration file option 'email_in_dn = no'
1587 has the same effect.
1588 [Massimiliano Pala madwolf@openca.org]
1590 *) Change all functions with names starting with des_ to be starting
1591 with DES_ instead. Add wrappers that are compatible with libdes,
1592 but are named _ossl_old_des_*. Finally, add macros that map the
1593 des_* symbols to the corresponding _ossl_old_des_* if libdes
1594 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1595 desired, the des_* symbols will be mapped to DES_*, with one
1598 Since we provide two compatibility mappings, the user needs to
1599 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1600 compatibility is desired. The default (i.e., when that macro
1601 isn't defined) is OpenSSL 0.9.6c compatibility.
1603 There are also macros that enable and disable the support of old
1604 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1605 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1606 are defined, the default will apply: to support the old des routines.
1608 In either case, one must include openssl/des.h to get the correct
1609 definitions. Do not try to just include openssl/des_old.h, that
1612 NOTE: This is a major break of an old API into a new one. Software
1613 authors are encouraged to switch to the DES_ style functions. Some
1614 time in the future, des_old.h and the libdes compatibility functions
1615 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1616 default), and then completely removed.
1619 *) Test for certificates which contain unsupported critical extensions.
1620 If such a certificate is found during a verify operation it is
1621 rejected by default: this behaviour can be overridden by either
1622 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1623 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1624 X509_supported_extension() has also been added which returns 1 if a
1625 particular extension is supported.
1628 *) Modify the behaviour of EVP cipher functions in similar way to digests
1629 to retain compatibility with existing code.
1632 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1633 compatibility with existing code. In particular the 'ctx' parameter does
1634 not have to be to be initialized before the call to EVP_DigestInit() and
1635 it is tidied up after a call to EVP_DigestFinal(). New function
1636 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1637 EVP_MD_CTX_copy() changed to not require the destination to be
1638 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1639 requires the destination to be valid.
1641 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1642 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1645 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1646 so that complete 'Handshake' protocol structures are kept in memory
1647 instead of overwriting 'msg_type' and 'length' with 'body' data.
1650 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1651 [Massimo Santin via Richard Levitte]
1653 *) Major restructuring to the underlying ENGINE code. This includes
1654 reduction of linker bloat, separation of pure "ENGINE" manipulation
1655 (initialisation, etc) from functionality dealing with implementations
1656 of specific crypto iterfaces. This change also introduces integrated
1657 support for symmetric ciphers and digest implementations - so ENGINEs
1658 can now accelerate these by providing EVP_CIPHER and EVP_MD
1659 implementations of their own. This is detailed in crypto/engine/README
1660 as it couldn't be adequately described here. However, there are a few
1661 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1662 were changed in the original introduction of ENGINE code have now
1663 reverted back - the hooking from this code to ENGINE is now a good
1664 deal more passive and at run-time, operations deal directly with
1665 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1666 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1667 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1668 they were not being used by the framework as there is no concept of a
1669 BIGNUM_METHOD and they could not be generalised to the new
1670 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1671 ENGINE_cpy() has been removed as it cannot be consistently defined in
1675 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1678 *) Change mkdef.pl to sort symbols that get the same entry number,
1679 and make sure the automatically generated functions ERR_load_*
1680 become part of libeay.num as well.
1683 *) New function SSL_renegotiate_pending(). This returns true once
1684 renegotiation has been requested (either SSL_renegotiate() call
1685 or HelloRequest/ClientHello receveived from the peer) and becomes
1686 false once a handshake has been completed.
1687 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1688 sends a HelloRequest, but does not ensure that a handshake takes
1689 place. SSL_renegotiate_pending() is useful for checking if the
1690 client has followed the request.)
1693 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1694 By default, clients may request session resumption even during
1695 renegotiation (if session ID contexts permit); with this option,
1696 session resumption is possible only in the first handshake.
1698 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1699 more bits available for options that should not be part of
1700 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1703 *) Add some demos for certificate and certificate request creation.
1706 *) Make maximum certificate chain size accepted from the peer application
1707 settable (SSL*_get/set_max_cert_list()), as proposed by
1708 "Douglas E. Engert" <deengert@anl.gov>.
1711 *) Add support for shared libraries for Unixware-7
1712 (Boyd Lynn Gerber <gerberb@zenez.com>).
1715 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1716 be done prior to destruction. Use this to unload error strings from
1717 ENGINEs that load their own error strings. NB: This adds two new API
1718 functions to "get" and "set" this destroy handler in an ENGINE.
1721 *) Alter all existing ENGINE implementations (except "openssl" and
1722 "openbsd") to dynamically instantiate their own error strings. This
1723 makes them more flexible to be built both as statically-linked ENGINEs
1724 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1725 Also, add stub code to each that makes building them as self-contained
1726 shared-libraries easier (see README.ENGINE).
1729 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1730 implementations into applications that are completely implemented in
1731 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1732 commands that can be used to configure what shared-library to load and
1733 to control aspects of the way it is handled. Also, made an update to
1734 the README.ENGINE file that brings its information up-to-date and
1735 provides some information and instructions on the "dynamic" ENGINE
1736 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1739 *) Make it possible to unload ranges of ERR strings with a new
1740 "ERR_unload_strings" function.
1743 *) Add a copy() function to EVP_MD.
1746 *) Make EVP_MD routines take a context pointer instead of just the
1747 md_data void pointer.
1750 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1751 that the digest can only process a single chunk of data
1752 (typically because it is provided by a piece of
1753 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1754 is only going to provide a single chunk of data, and hence the
1755 framework needn't accumulate the data for oneshot drivers.
1758 *) As with "ERR", make it possible to replace the underlying "ex_data"
1759 functions. This change also alters the storage and management of global
1760 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1761 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1762 index counters. The API functions that use this state have been changed
1763 to take a "class_index" rather than pointers to the class's local STACK
1764 and counter, and there is now an API function to dynamically create new
1765 classes. This centralisation allows us to (a) plug a lot of the
1766 thread-safety problems that existed, and (b) makes it possible to clean
1767 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1768 such data would previously have always leaked in application code and
1769 workarounds were in place to make the memory debugging turn a blind eye
1770 to it. Application code that doesn't use this new function will still
1771 leak as before, but their memory debugging output will announce it now
1772 rather than letting it slide.
1774 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1775 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1776 has a return value to indicate success or failure.
1779 *) Make it possible to replace the underlying "ERR" functions such that the
1780 global state (2 LHASH tables and 2 locks) is only used by the "default"
1781 implementation. This change also adds two functions to "get" and "set"
1782 the implementation prior to it being automatically set the first time
1783 any other ERR function takes place. Ie. an application can call "get",
1784 pass the return value to a module it has just loaded, and that module
1785 can call its own "set" function using that value. This means the
1786 module's "ERR" operations will use (and modify) the error state in the
1787 application and not in its own statically linked copy of OpenSSL code.
1790 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1791 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1792 the operation, and provides a more encapsulated way for external code
1793 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1794 to use these functions rather than manually incrementing the counts.
1796 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1799 *) Add EVP test program.
1802 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1805 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1806 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1807 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1808 These allow a CRL to be built without having to access X509_CRL fields
1809 directly. Modify 'ca' application to use new functions.
1812 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1813 bug workarounds. Rollback attack detection is a security feature.
1814 The problem will only arise on OpenSSL servers when TLSv1 is not
1815 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1816 Software authors not wanting to support TLSv1 will have special reasons
1817 for their choice and can explicitly enable this option.
1818 [Bodo Moeller, Lutz Jaenicke]
1820 *) Rationalise EVP so it can be extended: don't include a union of
1821 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1822 (similar to those existing for EVP_CIPHER_CTX).
1827 EVP_MD_CTX_init(&md); /* new function call */
1828 EVP_DigestInit(&md, EVP_sha1());
1829 EVP_DigestUpdate(&md, in, len);
1830 EVP_DigestFinal(&md, out, NULL);
1831 EVP_MD_CTX_cleanup(&md); /* new function call */
1835 *) Make DES key schedule conform to the usual scheme, as well as
1836 correcting its structure. This means that calls to DES functions
1837 now have to pass a pointer to a des_key_schedule instead of a
1838 plain des_key_schedule (which was actually always a pointer
1841 des_key_schedule ks;
1843 des_set_key_checked(..., &ks);
1844 des_ncbc_encrypt(..., &ks, ...);
1846 (Note that a later change renames 'des_...' into 'DES_...'.)
1849 *) Initial reduction of linker bloat: the use of some functions, such as
1850 PEM causes large amounts of unused functions to be linked in due to
1851 poor organisation. For example pem_all.c contains every PEM function
1852 which has a knock on effect of linking in large amounts of (unused)
1853 ASN1 code. Grouping together similar functions and splitting unrelated
1854 functions prevents this.
1857 *) Cleanup of EVP macros.
1860 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1861 correct _ecb suffix.
1864 *) Add initial OCSP responder support to ocsp application. The
1865 revocation information is handled using the text based index
1866 use by the ca application. The responder can either handle
1867 requests generated internally, supplied in files (for example
1868 via a CGI script) or using an internal minimal server.
1871 *) Add configuration choices to get zlib compression for TLS.
1874 *) Changes to Kerberos SSL for RFC 2712 compliance:
1875 1. Implemented real KerberosWrapper, instead of just using
1876 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1877 2. Implemented optional authenticator field of KerberosWrapper.
1879 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1880 and authenticator structs; see crypto/krb5/.
1882 Generalized Kerberos calls to support multiple Kerberos libraries.
1883 [Vern Staats <staatsvr@asc.hpc.mil>,
1884 Jeffrey Altman <jaltman@columbia.edu>
1885 via Richard Levitte]
1887 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1888 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1889 values for each of the key sizes rather than having just
1890 parameters (and 'speed' generating keys each time).
1893 *) Speed up EVP routines.
1896 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1897 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1898 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1899 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1901 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1902 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1903 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1906 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1908 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1911 *) Added the OS2-EMX target.
1912 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1914 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1915 to support NCONF routines in extension code. New function CONF_set_nconf()
1916 to allow functions which take an NCONF to also handle the old LHASH
1917 structure: this means that the old CONF compatible routines can be
1918 retained (in particular wrt extensions) without having to duplicate the
1919 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1922 *) Enhance the general user interface with mechanisms for inner control
1923 and with possibilities to have yes/no kind of prompts.
1926 *) Change all calls to low level digest routines in the library and
1927 applications to use EVP. Add missing calls to HMAC_cleanup() and
1928 don't assume HMAC_CTX can be copied using memcpy().
1929 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1931 *) Add the possibility to control engines through control names but with
1932 arbitrary arguments instead of just a string.
1933 Change the key loaders to take a UI_METHOD instead of a callback
1934 function pointer. NOTE: this breaks binary compatibility with earlier
1935 versions of OpenSSL [engine].
1936 Adapt the nCipher code for these new conditions and add a card insertion
1940 *) Enhance the general user interface with mechanisms to better support
1941 dialog box interfaces, application-defined prompts, the possibility
1942 to use defaults (for example default passwords from somewhere else)
1943 and interrupts/cancellations.
1946 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1947 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1950 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1951 tidy up some unnecessarily weird code in 'sk_new()').
1952 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1954 *) Change the key loading routines for ENGINEs to use the same kind
1955 callback (pem_password_cb) as all other routines that need this
1959 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1960 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1961 than this minimum value is recommended.
1964 *) New random seeder for OpenVMS, using the system process statistics
1965 that are easily reachable.
1968 *) Windows apparently can't transparently handle global
1969 variables defined in DLLs. Initialisations such as:
1971 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1973 wont compile. This is used by the any applications that need to
1974 declare their own ASN1 modules. This was fixed by adding the option
1975 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1976 needed for static libraries under Win32.
1979 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1980 setting of purpose and trust fields. New X509_STORE trust and
1981 purpose functions and tidy up setting in other SSL functions.
1984 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1985 structure. These are inherited by X509_STORE_CTX when it is
1986 initialised. This allows various defaults to be set in the
1987 X509_STORE structure (such as flags for CRL checking and custom
1988 purpose or trust settings) for functions which only use X509_STORE_CTX
1989 internally such as S/MIME.
1991 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1992 trust settings if they are not set in X509_STORE. This allows X509_STORE
1993 purposes and trust (in S/MIME for example) to override any set by default.
1995 Add command line options for CRL checking to smime, s_client and s_server
1999 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2000 are set then the CRL is looked up in the X509_STORE structure and
2001 its validity and signature checked, then if the certificate is found
2002 in the CRL the verify fails with a revoked error.
2004 Various new CRL related callbacks added to X509_STORE_CTX structure.
2006 Command line options added to 'verify' application to support this.
2008 This needs some additional work, such as being able to handle multiple
2009 CRLs with different times, extension based lookup (rather than just
2010 by subject name) and ultimately more complete V2 CRL extension
2014 *) Add a general user interface API (crypto/ui/). This is designed
2015 to replace things like des_read_password and friends (backward
2016 compatibility functions using this new API are provided).
2017 The purpose is to remove prompting functions from the DES code
2018 section as well as provide for prompting through dialog boxes in
2019 a window system and the like.
2022 *) Add "ex_data" support to ENGINE so implementations can add state at a
2023 per-structure level rather than having to store it globally.
2026 *) Make it possible for ENGINE structures to be copied when retrieved by
2027 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2028 This causes the "original" ENGINE structure to act like a template,
2029 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2030 operational state can be localised to each ENGINE structure, despite the
2031 fact they all share the same "methods". New ENGINE structures returned in
2032 this case have no functional references and the return value is the single
2033 structural reference. This matches the single structural reference returned
2034 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2038 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2039 needs to match any other type at all we need to manually clear the
2043 *) Changes to the "openssl engine" utility to include;
2044 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2045 about an ENGINE's available control commands.
2046 - executing control commands from command line arguments using the
2047 '-pre' and '-post' switches. '-post' is only used if '-t' is
2048 specified and the ENGINE is successfully initialised. The syntax for
2049 the individual commands are colon-separated, for example;
2050 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2053 *) New dynamic control command support for ENGINEs. ENGINEs can now
2054 declare their own commands (numbers), names (strings), descriptions,
2055 and input types for run-time discovery by calling applications. A
2056 subset of these commands are implicitly classed as "executable"
2057 depending on their input type, and only these can be invoked through
2058 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2059 can be based on user input, config files, etc). The distinction is
2060 that "executable" commands cannot return anything other than a boolean
2061 result and can only support numeric or string input, whereas some
2062 discoverable commands may only be for direct use through
2063 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2064 pointers, or other custom uses. The "executable" commands are to
2065 support parameterisations of ENGINE behaviour that can be
2066 unambiguously defined by ENGINEs and used consistently across any
2067 OpenSSL-based application. Commands have been added to all the
2068 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2069 control over shared-library paths without source code alterations.
2072 *) Changed all ENGINE implementations to dynamically allocate their
2073 ENGINEs rather than declaring them statically. Apart from this being
2074 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2075 this also allows the implementations to compile without using the
2076 internal engine_int.h header.
2079 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2080 'const' value. Any code that should be able to modify a RAND_METHOD
2081 should already have non-const pointers to it (ie. they should only
2082 modify their own ones).
2085 *) Made a variety of little tweaks to the ENGINE code.
2086 - "atalla" and "ubsec" string definitions were moved from header files
2087 to C code. "nuron" string definitions were placed in variables
2088 rather than hard-coded - allowing parameterisation of these values
2089 later on via ctrl() commands.
2090 - Removed unused "#if 0"'d code.
2091 - Fixed engine list iteration code so it uses ENGINE_free() to release
2092 structural references.
2093 - Constified the RAND_METHOD element of ENGINE structures.
2094 - Constified various get/set functions as appropriate and added
2095 missing functions (including a catch-all ENGINE_cpy that duplicates
2096 all ENGINE values onto a new ENGINE except reference counts/state).
2097 - Removed NULL parameter checks in get/set functions. Setting a method
2098 or function to NULL is a way of cancelling out a previously set
2099 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2100 and doesn't justify the extra error symbols and code.
2101 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2102 flags from engine_int.h to engine.h.
2103 - Changed prototypes for ENGINE handler functions (init(), finish(),
2104 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2107 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2108 to the algorithm using long division. The binary algorithm can be
2109 used only if the modulus is odd. On 32-bit systems, it is faster
2110 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2111 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2112 up to 450 bits. In 64-bit environments, the binary algorithm
2113 appears to be advantageous for much longer moduli; here we use it
2114 for moduli up to 2048 bits.
2117 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2118 could not support the combine flag in choice fields.
2121 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2122 extensions from a certificate request to the certificate.
2125 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2126 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2127 file: this allows the display of the certificate about to be
2128 signed to be customised, to allow certain fields to be included
2129 or excluded and extension details. The old system didn't display
2130 multicharacter strings properly, omitted fields not in the policy
2131 and couldn't display additional details such as extensions.
2134 *) Function EC_POINTs_mul for multiple scalar multiplication
2135 of an arbitrary number of elliptic curve points
2136 \sum scalars[i]*points[i],
2137 optionally including the generator defined for the EC_GROUP:
2138 scalar*generator + \sum scalars[i]*points[i].
2140 EC_POINT_mul is a simple wrapper function for the typical case
2141 that the point list has just one item (besides the optional
2145 *) First EC_METHODs for curves over GF(p):
2147 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2148 operations and provides various method functions that can also
2149 operate with faster implementations of modular arithmetic.
2151 EC_GFp_mont_method() reuses most functions that are part of
2152 EC_GFp_simple_method, but uses Montgomery arithmetic.
2154 [Bodo Moeller; point addition and point doubling
2155 implementation directly derived from source code provided by
2156 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2158 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2159 crypto/ec/ec_lib.c):
2161 Curves are EC_GROUP objects (with an optional group generator)
2162 based on EC_METHODs that are built into the library.
2164 Points are EC_POINT objects based on EC_GROUP objects.
2166 Most of the framework would be able to handle curves over arbitrary
2167 finite fields, but as there are no obvious types for fields other
2168 than GF(p), some functions are limited to that for now.
2171 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2172 that the file contains a complete HTTP response.
2175 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2176 change the def and num file printf format specifier from "%-40sXXX"
2177 to "%-39s XXX". The latter will always guarantee a space after the
2178 field while the former will cause them to run together if the field
2179 is 40 of more characters long.
2182 *) Constify the cipher and digest 'method' functions and structures
2183 and modify related functions to take constant EVP_MD and EVP_CIPHER
2187 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2188 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2191 *) Modify EVP_Digest*() routines so they now return values. Although the
2192 internal software routines can never fail additional hardware versions
2196 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2198 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2199 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2202 ERR_R_NESTED_ASN1_ERROR
2204 ERR_R_MISSING_ASN1_EOS
2205 were 4 .. 9, conflicting with
2206 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2208 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2209 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2211 Add new error code 'ERR_R_INTERNAL_ERROR'.
2214 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2218 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2219 sets the subject name for a new request or supersedes the
2220 subject name in a given request. Formats that can be parsed are
2221 'CN=Some Name, OU=myOU, C=IT'
2223 'CN=Some Name/OU=myOU/C=IT'.
2225 Add options '-batch' and '-verbose' to 'openssl req'.
2226 [Massimiliano Pala <madwolf@hackmasters.net>]
2228 *) Introduce the possibility to access global variables through
2229 functions on platform were that's the best way to handle exporting
2230 global variables in shared libraries. To enable this functionality,
2231 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2232 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2233 is normally done by Configure or something similar).
2235 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2236 in the source file (foo.c) like this:
2238 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2239 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2241 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2242 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2244 OPENSSL_DECLARE_GLOBAL(int,foo);
2245 #define foo OPENSSL_GLOBAL_REF(foo)
2246 OPENSSL_DECLARE_GLOBAL(double,bar);
2247 #define bar OPENSSL_GLOBAL_REF(bar)
2249 The #defines are very important, and therefore so is including the
2250 header file everywhere where the defined globals are used.
2252 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2253 of ASN.1 items, but that structure is a bit different.
2255 The largest change is in util/mkdef.pl which has been enhanced with
2256 better and easier to understand logic to choose which symbols should
2257 go into the Windows .def files as well as a number of fixes and code
2258 cleanup (among others, algorithm keywords are now sorted
2259 lexicographically to avoid constant rewrites).
2262 *) In BN_div() keep a copy of the sign of 'num' before writing the
2263 result to 'rm' because if rm==num the value will be overwritten
2264 and produce the wrong result if 'num' is negative: this caused
2265 problems with BN_mod() and BN_nnmod().
2268 *) Function OCSP_request_verify(). This checks the signature on an
2269 OCSP request and verifies the signer certificate. The signer
2270 certificate is just checked for a generic purpose and OCSP request
2274 *) Add OCSP_check_validity() function to check the validity of OCSP
2275 responses. OCSP responses are prepared in real time and may only
2276 be a few seconds old. Simply checking that the current time lies
2277 between thisUpdate and nextUpdate max reject otherwise valid responses
2278 caused by either OCSP responder or client clock inaccuracy. Instead
2279 we allow thisUpdate and nextUpdate to fall within a certain period of
2280 the current time. The age of the response can also optionally be
2281 checked. Two new options -validity_period and -status_age added to
2285 *) If signature or public key algorithm is unrecognized print out its
2286 OID rather that just UNKNOWN.
2289 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2290 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2291 ID to be generated from the issuer certificate alone which can then be
2292 passed to OCSP_id_issuer_cmp().
2295 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2296 ASN1 modules to export functions returning ASN1_ITEM pointers
2297 instead of the ASN1_ITEM structures themselves. This adds several
2298 new macros which allow the underlying ASN1 function/structure to
2299 be accessed transparently. As a result code should not use ASN1_ITEM
2300 references directly (such as &X509_it) but instead use the relevant
2301 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2302 use of the new ASN1 code on platforms where exporting structures
2303 is problematical (for example in shared libraries) but exporting
2304 functions returning pointers to structures is not.
2307 *) Add support for overriding the generation of SSL/TLS session IDs.
2308 These callbacks can be registered either in an SSL_CTX or per SSL.
2309 The purpose of this is to allow applications to control, if they wish,
2310 the arbitrary values chosen for use as session IDs, particularly as it
2311 can be useful for session caching in multiple-server environments. A
2312 command-line switch for testing this (and any client code that wishes
2313 to use such a feature) has been added to "s_server".
2314 [Geoff Thorpe, Lutz Jaenicke]
2316 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2317 of the form '#if defined(...) || defined(...) || ...' and
2318 '#if !defined(...) && !defined(...) && ...'. This also avoids
2319 the growing number of special cases it was previously handling.
2322 *) Make all configuration macros available for application by making
2323 sure they are available in opensslconf.h, by giving them names starting
2324 with "OPENSSL_" to avoid conflicts with other packages and by making
2325 sure e_os2.h will cover all platform-specific cases together with
2327 Additionally, it is now possible to define configuration/platform-
2328 specific names (called "system identities"). In the C code, these
2329 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2330 macro with the name beginning with "OPENSSL_SYS_", which is determined
2331 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2335 *) New option -set_serial to 'req' and 'x509' this allows the serial
2336 number to use to be specified on the command line. Previously self
2337 signed certificates were hard coded with serial number 0 and the
2338 CA options of 'x509' had to use a serial number in a file which was
2342 *) New options to 'ca' utility to support V2 CRL entry extensions.
2343 Currently CRL reason, invalidity date and hold instruction are
2344 supported. Add new CRL extensions to V3 code and some new objects.
2347 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2348 disable standard block padding (aka PKCS#5 padding) in the EVP
2349 API, which was previously mandatory. This means that the data is
2350 not padded in any way and so the total length much be a multiple
2351 of the block size, otherwise an error occurs.
2354 *) Initial (incomplete) OCSP SSL support.
2357 *) New function OCSP_parse_url(). This splits up a URL into its host,
2358 port and path components: primarily to parse OCSP URLs. New -url
2359 option to ocsp utility.
2362 *) New nonce behavior. The return value of OCSP_check_nonce() now
2363 reflects the various checks performed. Applications can decide
2364 whether to tolerate certain situations such as an absent nonce
2365 in a response when one was present in a request: the ocsp application
2366 just prints out a warning. New function OCSP_add1_basic_nonce()
2367 this is to allow responders to include a nonce in a response even if
2368 the request is nonce-less.
2371 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2372 skipped when using openssl x509 multiple times on a single input file,
2373 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2376 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2377 set string type: to handle setting ASN1_TIME structures. Fix ca
2378 utility to correctly initialize revocation date of CRLs.
2381 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2382 the clients preferred ciphersuites and rather use its own preferences.
2383 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2384 Internet Explorer by ensuring unchanged hash method during stepup.
2385 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2388 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2389 to aes and add a new 'exist' option to print out symbols that don't
2393 *) Additional options to ocsp utility to allow flags to be set and
2394 additional certificates supplied.
2397 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2398 OCSP client a number of certificate to only verify the response
2402 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2403 handle the new API. Currently only ECB, CBC modes supported. Add new
2406 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2407 Encryption Standard (AES) Ciphersuites for Transport Layer
2408 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2409 not enabled by default and were not part of the "ALL" ciphersuite
2410 alias because they were not yet official; they could be
2411 explicitly requested by specifying the "AESdraft" ciphersuite
2412 group alias. In the final release of OpenSSL 0.9.7, the group
2413 alias is called "AES" and is part of "ALL".)
2414 [Ben Laurie, Steve Henson, Bodo Moeller]
2416 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2417 request to response.
2420 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2421 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2422 extract information from a certificate request. OCSP_response_create()
2423 creates a response and optionally adds a basic response structure.
2424 OCSP_basic_add1_status() adds a complete single response to a basic
2425 response and returns the OCSP_SINGLERESP structure just added (to allow
2426 extensions to be included for example). OCSP_basic_add1_cert() adds a
2427 certificate to a basic response and OCSP_basic_sign() signs a basic
2428 response with various flags. New helper functions ASN1_TIME_check()
2429 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2430 (converts ASN1_TIME to GeneralizedTime).
2433 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2434 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2435 structure from a certificate. X509_pubkey_digest() digests the public_key
2436 contents: this is used in various key identifiers.
2439 *) Make sk_sort() tolerate a NULL argument.
2440 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2442 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2443 passed by the function are trusted implicitly. If any of them signed the
2444 response then it is assumed to be valid and is not verified.
2447 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2448 to data. This was previously part of the PKCS7 ASN1 code. This
2449 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2450 [Steve Henson, reported by Kenneth R. Robinette
2451 <support@securenetterm.com>]
2453 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2454 routines: without these tracing memory leaks is very painful.
2455 Fix leaks in PKCS12 and PKCS7 routines.
2458 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2459 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2460 effectively meant GeneralizedTime would never be used. Now it
2461 is initialised to -1 but X509_time_adj() now has to check the value
2462 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2463 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2464 [Steve Henson, reported by Kenneth R. Robinette
2465 <support@securenetterm.com>]
2467 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2468 result in a zero length in the ASN1_INTEGER structure which was
2469 not consistent with the structure when d2i_ASN1_INTEGER() was used
2470 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2471 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2472 where it did not print out a minus for negative ASN1_INTEGER.
2475 *) Add summary printout to ocsp utility. The various functions which
2476 convert status values to strings have been renamed to:
2477 OCSP_response_status_str(), OCSP_cert_status_str() and
2478 OCSP_crl_reason_str() and are no longer static. New options
2479 to verify nonce values and to disable verification. OCSP response
2480 printout format cleaned up.
2483 *) Add additional OCSP certificate checks. These are those specified
2484 in RFC2560. This consists of two separate checks: the CA of the
2485 certificate being checked must either be the OCSP signer certificate
2486 or the issuer of the OCSP signer certificate. In the latter case the
2487 OCSP signer certificate must contain the OCSP signing extended key
2488 usage. This check is performed by attempting to match the OCSP
2489 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2490 in the OCSP_CERTID structures of the response.
2493 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2494 and related routines. This uses the standard OpenSSL certificate
2495 verify routines to perform initial checks (just CA validity) and
2496 to obtain the certificate chain. Then additional checks will be
2497 performed on the chain. Currently the root CA is checked to see
2498 if it is explicitly trusted for OCSP signing. This is used to set
2499 a root CA as a global signing root: that is any certificate that
2500 chains to that CA is an acceptable OCSP signing certificate.
2503 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2504 extensions from a separate configuration file.
2505 As when reading extensions from the main configuration file,
2506 the '-extensions ...' option may be used for specifying the
2508 [Massimiliano Pala <madwolf@comune.modena.it>]
2510 *) New OCSP utility. Allows OCSP requests to be generated or
2511 read. The request can be sent to a responder and the output
2512 parsed, outputed or printed in text form. Not complete yet:
2513 still needs to check the OCSP response validity.
2516 *) New subcommands for 'openssl ca':
2517 'openssl ca -status <serial>' prints the status of the cert with
2518 the given serial number (according to the index file).
2519 'openssl ca -updatedb' updates the expiry status of certificates
2521 [Massimiliano Pala <madwolf@comune.modena.it>]
2523 *) New '-newreq-nodes' command option to CA.pl. This is like
2524 '-newreq', but calls 'openssl req' with the '-nodes' option
2525 so that the resulting key is not encrypted.
2526 [Damien Miller <djm@mindrot.org>]
2528 *) New configuration for the GNU Hurd.
2529 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2531 *) Initial code to implement OCSP basic response verify. This
2532 is currently incomplete. Currently just finds the signer's
2533 certificate and verifies the signature on the response.
2536 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2537 value of OPENSSLDIR. This is available via the new '-d' option
2538 to 'openssl version', and is also included in 'openssl version -a'.
2541 *) Allowing defining memory allocation callbacks that will be given
2542 file name and line number information in additional arguments
2543 (a const char* and an int). The basic functionality remains, as
2544 well as the original possibility to just replace malloc(),
2545 realloc() and free() by functions that do not know about these
2546 additional arguments. To register and find out the current
2547 settings for extended allocation functions, the following
2548 functions are provided:
2550 CRYPTO_set_mem_ex_functions
2551 CRYPTO_set_locked_mem_ex_functions
2552 CRYPTO_get_mem_ex_functions
2553 CRYPTO_get_locked_mem_ex_functions
2555 These work the same way as CRYPTO_set_mem_functions and friends.
2556 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2557 extended allocation function is enabled.
2558 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2559 a conventional allocation function is enabled.
2560 [Richard Levitte, Bodo Moeller]
2562 *) Finish off removing the remaining LHASH function pointer casts.
2563 There should no longer be any prototype-casting required when using
2564 the LHASH abstraction, and any casts that remain are "bugs". See
2565 the callback types and macros at the head of lhash.h for details
2566 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2569 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2570 If /dev/[u]random devices are not available or do not return enough
2571 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2573 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2574 /etc/entropy will be queried once each in this sequence, quering stops
2575 when enough entropy was collected without querying more sockets.
2578 *) Change the Unix RAND_poll() variant to be able to poll several
2579 random devices, as specified by DEVRANDOM, until a sufficient amount
2580 of data has been collected. We spend at most 10 ms on each file
2581 (select timeout) and read in non-blocking mode. DEVRANDOM now
2582 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2583 (previously it was just the string "/dev/urandom"), so on typical
2584 platforms the 10 ms delay will never occur.
2585 Also separate out the Unix variant to its own file, rand_unix.c.
2586 For VMS, there's a currently-empty rand_vms.c.
2589 *) Move OCSP client related routines to ocsp_cl.c. These
2590 provide utility functions which an application needing
2591 to issue a request to an OCSP responder and analyse the
2592 response will typically need: as opposed to those which an
2593 OCSP responder itself would need which will be added later.
2595 OCSP_request_sign() signs an OCSP request with an API similar
2596 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2597 response. OCSP_response_get1_basic() extracts basic response
2598 from response. OCSP_resp_find_status(): finds and extracts status
2599 information from an OCSP_CERTID structure (which will be created
2600 when the request structure is built). These are built from lower
2601 level functions which work on OCSP_SINGLERESP structures but
2602 wont normally be used unless the application wishes to examine
2603 extensions in the OCSP response for example.
2605 Replace nonce routines with a pair of functions.
2606 OCSP_request_add1_nonce() adds a nonce value and optionally
2607 generates a random value. OCSP_check_nonce() checks the
2608 validity of the nonce in an OCSP response.
2611 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2612 This doesn't copy the supplied OCSP_CERTID and avoids the
2613 need to free up the newly created id. Change return type
2614 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2615 This can then be used to add extensions to the request.
2616 Deleted OCSP_request_new(), since most of its functionality
2617 is now in OCSP_REQUEST_new() (and the case insensitive name
2618 clash) apart from the ability to set the request name which
2619 will be added elsewhere.
2622 *) Update OCSP API. Remove obsolete extensions argument from
2623 various functions. Extensions are now handled using the new
2624 OCSP extension code. New simple OCSP HTTP function which
2625 can be used to send requests and parse the response.
2628 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2629 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2630 uses the special reorder version of SET OF to sort the attributes
2631 and reorder them to match the encoded order. This resolves a long
2632 standing problem: a verify on a PKCS7 structure just after signing
2633 it used to fail because the attribute order did not match the
2634 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2635 it uses the received order. This is necessary to tolerate some broken
2636 software that does not order SET OF. This is handled by encoding
2637 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2638 to produce the required SET OF.
2641 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2642 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2643 files to get correct declarations of the ASN.1 item variables.
2646 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2647 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2648 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2649 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2650 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2651 ASN1_ITEM and no wrapper functions.
2654 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2655 replace the old function pointer based I/O routines. Change most of
2656 the *_d2i_bio() and *_d2i_fp() functions to use these.
2659 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2660 lines, recognice more "algorithms" that can be deselected, and make
2661 it complain about algorithm deselection that isn't recognised.
2664 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2665 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2666 to use new functions. Add NO_ASN1_OLD which can be set to remove
2667 some old style ASN1 functions: this can be used to determine if old
2668 code will still work when these eventually go away.
2671 *) New extension functions for OCSP structures, these follow the
2672 same conventions as certificates and CRLs.
2675 *) New function X509V3_add1_i2d(). This automatically encodes and
2676 adds an extension. Its behaviour can be customised with various
2677 flags to append, replace or delete. Various wrappers added for
2678 certifcates and CRLs.
2681 *) Fix to avoid calling the underlying ASN1 print routine when
2682 an extension cannot be parsed. Correct a typo in the
2683 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2686 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2687 entries for variables.
2690 *) Add functionality to apps/openssl.c for detecting locking
2691 problems: As the program is single-threaded, all we have
2692 to do is register a locking callback using an array for
2693 storing which locks are currently held by the program.
2696 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2697 SSL_get_ex_data_X509_STORE_idx(), which is used in
2698 ssl_verify_cert_chain() and thus can be called at any time
2699 during TLS/SSL handshakes so that thread-safety is essential.
2700 Unfortunately, the ex_data design is not at all suited
2701 for multi-threaded use, so it probably should be abolished.
2704 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2705 [Broadcom, tweaked and integrated by Geoff Thorpe]
2707 *) Move common extension printing code to new function
2708 X509V3_print_extensions(). Reorganise OCSP print routines and
2709 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2712 *) New function X509_signature_print() to remove duplication in some
2716 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2717 set (this was treated exactly the same as SET OF previously). This
2718 is used to reorder the STACK representing the structure to match the
2719 encoding. This will be used to get round a problem where a PKCS7
2720 structure which was signed could not be verified because the STACK
2721 order did not reflect the encoded order.
2724 *) Reimplement the OCSP ASN1 module using the new code.
2727 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2728 for its ASN1 operations. The old style function pointers still exist
2729 for now but they will eventually go away.
2732 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2733 completely replaces the old ASN1 functionality with a table driven
2734 encoder and decoder which interprets an ASN1_ITEM structure describing
2735 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2736 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2737 has also been converted to the new form.
2740 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2741 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2742 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2743 for negative moduli.
2746 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2747 of not touching the result's sign bit.
2750 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2754 *) Changed the LHASH code to use prototypes for callbacks, and created
2755 macros to declare and implement thin (optionally static) functions
2756 that provide type-safety and avoid function pointer casting for the
2757 type-specific callbacks.
2760 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2762 [Veers Staats <staatsvr@asc.hpc.mil>,
2763 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2765 *) Reformat the FAQ so the different questions and answers can be divided
2766 in sections depending on the subject.
2769 *) Have the zlib compression code load ZLIB.DLL dynamically under
2773 *) New function BN_mod_sqrt for computing square roots modulo a prime
2774 (using the probabilistic Tonelli-Shanks algorithm unless
2775 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2776 be handled deterministically).
2777 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2779 *) Make BN_mod_inverse faster by explicitly handling small quotients
2780 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2781 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2784 *) New function BN_kronecker.
2787 *) Fix BN_gcd so that it works on negative inputs; the result is
2788 positive unless both parameters are zero.
2789 Previously something reasonably close to an infinite loop was
2790 possible because numbers could be growing instead of shrinking
2791 in the implementation of Euclid's algorithm.
2794 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2795 sign of the number in question.
2797 Fix BN_is_word(a,w) to work correctly for w == 0.
2799 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2800 because its test if the absolute value of 'a' equals 'w'.
2801 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2802 it exists mostly for use in the implementations of BN_is_zero(),
2803 BN_is_one(), and BN_is_word().
2806 *) New function BN_swap.
2809 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2810 the exponentiation functions are more likely to produce reasonable
2811 results on negative inputs.
2814 *) Change BN_mod_mul so that the result is always non-negative.
2815 Previously, it could be negative if one of the factors was negative;
2816 I don't think anyone really wanted that behaviour.
2819 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2820 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2821 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2822 and add new functions:
2831 BN_mod_lshift1_quick
2835 These functions always generate non-negative results.
2837 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2838 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2840 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2841 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2842 be reduced modulo m.
2843 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2846 The following entry accidentily appeared in the CHANGES file
2847 distributed with OpenSSL 0.9.7. The modifications described in
2848 it do *not* apply to OpenSSL 0.9.7.
2850 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2851 was actually never needed) and in BN_mul(). The removal in BN_mul()
2852 required a small change in bn_mul_part_recursive() and the addition
2853 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2854 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2855 bn_sub_words() and bn_add_words() except they take arrays with
2860 *) In 'openssl passwd', verify passwords read from the terminal
2861 unless the '-salt' option is used (which usually means that
2862 verification would just waste user's time since the resulting
2863 hash is going to be compared with some given password hash)
2864 or the new '-noverify' option is used.
2866 This is an incompatible change, but it does not affect
2867 non-interactive use of 'openssl passwd' (passwords on the command
2868 line, '-stdin' option, '-in ...' option) and thus should not
2872 *) Remove all references to RSAref, since there's no more need for it.
2875 *) Make DSO load along a path given through an environment variable
2876 (SHLIB_PATH) with shl_load().
2879 *) Constify the ENGINE code as a result of BIGNUM constification.
2880 Also constify the RSA code and most things related to it. In a
2881 few places, most notable in the depth of the ASN.1 code, ugly
2882 casts back to non-const were required (to be solved at a later
2886 *) Make it so the openssl application has all engines loaded by default.
2889 *) Constify the BIGNUM routines a little more.
2892 *) Add the following functions:
2894 ENGINE_load_cswift()
2896 ENGINE_load_atalla()
2898 ENGINE_load_builtin_engines()
2900 That way, an application can itself choose if external engines that
2901 are built-in in OpenSSL shall ever be used or not. The benefit is
2902 that applications won't have to be linked with libdl or other dso
2903 libraries unless it's really needed.
2905 Changed 'openssl engine' to load all engines on demand.
2906 Changed the engine header files to avoid the duplication of some
2907 declarations (they differed!).
2910 *) 'openssl engine' can now list capabilities.
2913 *) Better error reporting in 'openssl engine'.
2916 *) Never call load_dh_param(NULL) in s_server.
2919 *) Add engine application. It can currently list engines by name and
2920 identity, and test if they are actually available.
2923 *) Improve RPM specification file by forcing symbolic linking and making
2924 sure the installed documentation is also owned by root.root.
2925 [Damien Miller <djm@mindrot.org>]
2927 *) Give the OpenSSL applications more possibilities to make use of
2928 keys (public as well as private) handled by engines.
2931 *) Add OCSP code that comes from CertCo.
2934 *) Add VMS support for the Rijndael code.
2937 *) Added untested support for Nuron crypto accelerator.
2940 *) Add support for external cryptographic devices. This code was
2941 previously distributed separately as the "engine" branch.
2942 [Geoff Thorpe, Richard Levitte]
2944 *) Rework the filename-translation in the DSO code. It is now possible to
2945 have far greater control over how a "name" is turned into a filename
2946 depending on the operating environment and any oddities about the
2947 different shared library filenames on each system.
2950 *) Support threads on FreeBSD-elf in Configure.
2953 *) Fix for SHA1 assembly problem with MASM: it produces
2954 warnings about corrupt line number information when assembling
2955 with debugging information. This is caused by the overlapping
2957 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2960 NCONF_get_number() has no error checking at all. As a replacement,
2961 NCONF_get_number_e() is defined (_e for "error checking") and is
2962 promoted strongly. The old NCONF_get_number is kept around for
2963 binary backward compatibility.
2964 Make it possible for methods to load from something other than a BIO,
2965 by providing a function pointer that is given a name instead of a BIO.
2966 For example, this could be used to load configuration data from an
2970 *) Fix for non blocking accept BIOs. Added new I/O special reason
2971 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2972 with non blocking I/O was not possible because no retry code was
2973 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2977 *) Added the beginnings of Rijndael support.
2980 *) Fix for bug in DirectoryString mask setting. Add support for
2981 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2982 to allow certificate printing to more controllable, additional
2983 'certopt' option to 'x509' to allow new printing options to be
2987 *) Clean old EAY MD5 hack from e_os.h.
2990 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2992 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2993 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2994 [Joe Orton, Steve Henson]
2996 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2998 *) Fix additional bug revealed by the NISCC test suite:
3000 Stop bug triggering large recursion when presented with
3001 certain ASN.1 tags (CAN-2003-0851)
3004 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3006 *) Fix various bugs revealed by running the NISCC test suite:
3008 Stop out of bounds reads in the ASN1 code when presented with
3009 invalid tags (CAN-2003-0543 and CAN-2003-0544).
3011 If verify callback ignores invalid public key errors don't try to check
3012 certificate signature with the NULL public key.
3016 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3017 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3021 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3022 extra data after the compression methods not only for TLS 1.0
3023 but also for SSL 3.0 (as required by the specification).
3024 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3026 *) Change X509_certificate_type() to mark the key as exported/exportable
3027 when it's 512 *bits* long, not 512 bytes.
3030 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3032 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3033 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3034 a protocol version number mismatch like a decryption error
3035 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3038 *) Turn on RSA blinding by default in the default implementation
3039 to avoid a timing attack. Applications that don't want it can call
3040 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3041 They would be ill-advised to do so in most cases.
3042 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3044 *) Change RSA blinding code so that it works when the PRNG is not
3045 seeded (in this case, the secret RSA exponent is abused as
3046 an unpredictable seed -- if it is not unpredictable, there
3047 is no point in blinding anyway). Make RSA blinding thread-safe
3048 by remembering the creator's thread ID in rsa->blinding and
3049 having all other threads use local one-time blinding factors
3050 (this requires more computation than sharing rsa->blinding, but
3051 avoids excessive locking; and if an RSA object is not shared
3052 between threads, blinding will still be very fast).
3055 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3057 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3058 via timing by performing a MAC computation even if incorrrect
3059 block cipher padding has been found. This is a countermeasure
3060 against active attacks where the attacker has to distinguish
3061 between bad padding and a MAC verification error. (CAN-2003-0078)
3063 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3064 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3065 Martin Vuagnoux (EPFL, Ilion)]
3067 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3069 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3070 memory from it's contents. This is done with a counter that will
3071 place alternating values in each byte. This can be used to solve
3072 two issues: 1) the removal of calls to memset() by highly optimizing
3073 compilers, and 2) cleansing with other values than 0, since those can
3074 be read through on certain media, for example a swap space on disk.
3077 *) Bugfix: client side session caching did not work with external caching,
3078 because the session->cipher setting was not restored when reloading
3079 from the external cache. This problem was masked, when
3080 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3081 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3084 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3085 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3086 [Zeev Lieber <zeev-l@yahoo.com>]
3088 *) Undo an undocumented change introduced in 0.9.6e which caused
3089 repeated calls to OpenSSL_add_all_ciphers() and
3090 OpenSSL_add_all_digests() to be ignored, even after calling
3094 *) Change the default configuration reader to deal with last line not
3095 being properly terminated.
3098 *) Change X509_NAME_cmp() so it applies the special rules on handling
3099 DN values that are of type PrintableString, as well as RDNs of type
3100 emailAddress where the value has the type ia5String.
3101 [stefank@valicert.com via Richard Levitte]
3103 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3104 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3105 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3106 the bitwise-OR of the two for use by the majority of applications
3107 wanting this behaviour, and update the docs. The documented
3108 behaviour and actual behaviour were inconsistent and had been
3109 changing anyway, so this is more a bug-fix than a behavioural
3111 [Geoff Thorpe, diagnosed by Nadav Har'El]
3113 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3114 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3117 *) Fix initialization code race conditions in
3118 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3119 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3120 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3121 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3122 ssl2_get_cipher_by_char(),
3123 ssl3_get_cipher_by_char().
3124 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3126 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3127 the cached sessions are flushed, as the remove_cb() might use ex_data
3128 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3129 (see [openssl.org #212]).
3130 [Geoff Thorpe, Lutz Jaenicke]
3132 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3133 length, instead of the encoding length to d2i_ASN1_OBJECT.
3136 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3138 *) [In 0.9.6g-engine release:]
3139 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3140 [Lynn Gazis <lgazis@rainbow.com>]
3142 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3144 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3145 and get fix the header length calculation.
3146 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3147 Alon Kantor <alonk@checkpoint.com> (and others),
3150 *) Use proper error handling instead of 'assertions' in buffer
3151 overflow checks added in 0.9.6e. This prevents DoS (the
3152 assertions could call abort()).
3153 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3155 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3157 *) Add various sanity checks to asn1_get_length() to reject
3158 the ASN1 length bytes if they exceed sizeof(long), will appear
3159 negative or the content length exceeds the length of the
3161 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3163 *) Fix cipher selection routines: ciphers without encryption had no flags
3164 for the cipher strength set and where therefore not handled correctly
3165 by the selection routines (PR #130).
3168 *) Fix EVP_dsa_sha macro.
3172 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3173 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3174 that was added in OpenSSL 0.9.6d.
3176 As the countermeasure turned out to be incompatible with some
3177 broken SSL implementations, the new option is part of SSL_OP_ALL.
3178 SSL_OP_ALL is usually employed when compatibility with weird SSL
3179 implementations is desired (e.g. '-bugs' option to 's_client' and
3180 's_server'), so the new option is automatically set in many
3184 *) Changes in security patch:
3186 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3187 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3188 Air Force Materiel Command, USAF, under agreement number
3191 *) Add various sanity checks to asn1_get_length() to reject
3192 the ASN1 length bytes if they exceed sizeof(long), will appear
3193 negative or the content length exceeds the length of the
3194 supplied buffer. (CAN-2002-0659)
3195 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3197 *) Assertions for various potential buffer overflows, not known to
3199 [Ben Laurie (CHATS)]
3201 *) Various temporary buffers to hold ASCII versions of integers were
3202 too small for 64 bit platforms. (CAN-2002-0655)
3203 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3205 *) Remote buffer overflow in SSL3 protocol - an attacker could
3206 supply an oversized session ID to a client. (CAN-2002-0656)
3207 [Ben Laurie (CHATS)]
3209 *) Remote buffer overflow in SSL2 protocol - an attacker could
3210 supply an oversized client master key. (CAN-2002-0656)
3211 [Ben Laurie (CHATS)]
3213 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3215 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3216 encoded as NULL) with id-dsa-with-sha1.
3217 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3219 *) Check various X509_...() return values in apps/req.c.
3220 [Nils Larsch <nla@trustcenter.de>]
3222 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3223 an end-of-file condition would erronously be flagged, when the CRLF
3224 was just at the end of a processed block. The bug was discovered when
3225 processing data through a buffering memory BIO handing the data to a
3226 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3227 <ptsekov@syntrex.com> and Nedelcho Stanev.
3230 *) Implement a countermeasure against a vulnerability recently found
3231 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3232 before application data chunks to avoid the use of known IVs
3233 with data potentially chosen by the attacker.
3236 *) Fix length checks in ssl3_get_client_hello().
3239 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3240 to prevent ssl3_read_internal() from incorrectly assuming that
3241 ssl3_read_bytes() found application data while handshake
3242 processing was enabled when in fact s->s3->in_read_app_data was
3243 merely automatically cleared during the initial handshake.
3244 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3246 *) Fix object definitions for Private and Enterprise: they were not
3247 recognized in their shortname (=lowercase) representation. Extend
3248 obj_dat.pl to issue an error when using undefined keywords instead
3249 of silently ignoring the problem (Svenning Sorensen
3250 <sss@sss.dnsalias.net>).
3253 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3254 generators, i.e. generators other than 2 and 5. (Previously, the
3255 code did not properly initialise the 'add' and 'rem' values to
3256 BN_generate_prime().)
3258 In the new general case, we do not insist that 'generator' is
3259 actually a primitive root: This requirement is rather pointless;
3260 a generator of the order-q subgroup is just as good, if not
3264 *) Map new X509 verification errors to alerts. Discovered and submitted by
3265 Tom Wu <tom@arcot.com>.
3268 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3269 returning non-zero before the data has been completely received
3270 when using non-blocking I/O.
3271 [Bodo Moeller; problem pointed out by John Hughes]
3273 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3274 [Ben Laurie, Lutz Jaenicke]
3276 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3277 Yoram Zahavi <YoramZ@gilian.com>).
3280 *) Add information about CygWin 1.3 and on, and preserve proper
3281 configuration for the versions before that.
3282 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3284 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3285 check whether we deal with a copy of a session and do not delete from
3286 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3287 <izhar@checkpoint.com>.
3290 *) Do not store session data into the internal session cache, if it
3291 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3292 flag is set). Proposed by Aslam <aslam@funk.com>.
3295 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3299 *) [In 0.9.6d-engine release:]
3300 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3301 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3303 *) Add the configuration target linux-s390x.
3304 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3306 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3307 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3308 variable as an indication that a ClientHello message has been
3309 received. As the flag value will be lost between multiple
3310 invocations of ssl3_accept when using non-blocking I/O, the
3311 function may not be aware that a handshake has actually taken
3312 place, thus preventing a new session from being added to the
3315 To avoid this problem, we now set s->new_session to 2 instead of
3316 using a local variable.
3317 [Lutz Jaenicke, Bodo Moeller]
3319 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3320 if the SSL_R_LENGTH_MISMATCH error is detected.
3321 [Geoff Thorpe, Bodo Moeller]
3323 *) New 'shared_ldflag' column in Configure platform table.
3326 *) Fix EVP_CIPHER_mode macro.
3327 ["Dan S. Camper" <dan@bti.net>]
3329 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3330 type, we must throw them away by setting rr->length to 0.
3331 [D P Chang <dpc@qualys.com>]
3333 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3335 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3336 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3337 worked incorrectly for those cases where range = 10..._2 and
3338 3*range is two bits longer than range.)
3341 *) Only add signing time to PKCS7 structures if it is not already
3345 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3346 OBJ_ld_ce should be OBJ_id_ce.
3347 Also some ip-pda OIDs in crypto/objects/objects.txt were
3348 incorrect (cf. RFC 3039).
3349 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3351 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3352 returns early because it has nothing to do.
3353 [Andy Schneider <andy.schneider@bjss.co.uk>]
3355 *) [In 0.9.6c-engine release:]
3356 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3357 [Andy Schneider <andy.schneider@bjss.co.uk>]
3359 *) [In 0.9.6c-engine release:]
3360 Add support for Cryptographic Appliance's keyserver technology.
3361 (Use engine 'keyclient')
3362 [Cryptographic Appliances and Geoff Thorpe]
3364 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3365 is called via tools/c89.sh because arguments have to be
3366 rearranged (all '-L' options must appear before the first object
3368 [Richard Shapiro <rshapiro@abinitio.com>]
3370 *) [In 0.9.6c-engine release:]
3371 Add support for Broadcom crypto accelerator cards, backported
3373 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3375 *) [In 0.9.6c-engine release:]
3376 Add support for SureWare crypto accelerator cards from
3377 Baltimore Technologies. (Use engine 'sureware')
3378 [Baltimore Technologies and Mark Cox]
3380 *) [In 0.9.6c-engine release:]
3381 Add support for crypto accelerator cards from Accelerated
3382 Encryption Processing, www.aep.ie. (Use engine 'aep')
3383 [AEP Inc. and Mark Cox]
3385 *) Add a configuration entry for gcc on UnixWare.
3386 [Gary Benson <gbenson@redhat.com>]
3388 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3389 messages are stored in a single piece (fixed-length part and
3390 variable-length part combined) and fix various bugs found on the way.
3393 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3394 instead. BIO_gethostbyname() does not know what timeouts are
3395 appropriate, so entries would stay in cache even when they have
3397 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3399 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3400 faced with a pathologically small ClientHello fragment that does
3401 not contain client_version: Instead of aborting with an error,
3402 simply choose the highest available protocol version (i.e.,
3403 TLS 1.0 unless it is disabled). In practice, ClientHello
3404 messages are never sent like this, but this change gives us
3405 strictly correct behaviour at least for TLS.
3408 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3409 never resets s->method to s->ctx->method when called from within
3410 one of the SSL handshake functions.
3411 [Bodo Moeller; problem pointed out by Niko Baric]
3413 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3414 (sent using the client's version number) if client_version is
3415 smaller than the protocol version in use. Also change
3416 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3417 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3418 the client will at least see that alert.
3421 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3425 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3426 client receives HelloRequest while in a handshake.
3427 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3429 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3430 should end in 'break', not 'goto end' which circuments various
3431 cleanups done in state SSL_ST_OK. But session related stuff
3432 must be disabled for SSL_ST_OK in the case that we just sent a
3435 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3436 before just sending a HelloRequest.
3437 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3439 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3440 reveal whether illegal block cipher padding was found or a MAC
3441 verification error occured. (Neither SSLerr() codes nor alerts
3442 are directly visible to potential attackers, but the information
3443 may leak via logfiles.)
3445 Similar changes are not required for the SSL 2.0 implementation
3446 because the number of padding bytes is sent in clear for SSL 2.0,
3447 and the extra bytes are just ignored. However ssl/s2_pkt.c
3448 failed to verify that the purported number of padding bytes is in
3452 *) Add OpenUNIX-8 support including shared libraries
3453 (Boyd Lynn Gerber <gerberb@zenez.com>).
3456 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3457 'wristwatch attack' using huge encoding parameters (cf.
3458 James H. Manger's CRYPTO 2001 paper). Note that the
3459 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3460 encoding parameters and hence was not vulnerable.
3463 *) BN_sqr() bug fix.
3464 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3466 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3467 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3468 followed by modular reduction.
3469 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3471 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3472 equivalent based on BN_pseudo_rand() instead of BN_rand().
3475 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3476 This function was broken, as the check for a new client hello message
3477 to handle SGC did not allow these large messages.
3478 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3481 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3484 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3485 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3488 *) Rework the configuration and shared library support for Tru64 Unix.
3489 The configuration part makes use of modern compiler features and
3490 still retains old compiler behavior for those that run older versions
3491 of the OS. The shared library support part includes a variant that
3492 uses the RPATH feature, and is available through the special
3493 configuration target "alpha-cc-rpath", which will never be selected
3495 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3497 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3498 with the same message size as in ssl3_get_certificate_request().
3499 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3500 messages might inadvertently be reject as too long.
3501 [Petr Lampa <lampa@fee.vutbr.cz>]
3503 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3506 *) Modified SSL library such that the verify_callback that has been set
3507 specificly for an SSL object with SSL_set_verify() is actually being
3508 used. Before the change, a verify_callback set with this function was
3509 ignored and the verify_callback() set in the SSL_CTX at the time of
3510 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3511 to allow the necessary settings.
3514 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3515 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3516 done automatically (in contradiction to the requirements of the C
3517 standard). This made problems when used from OpenSSH.
3520 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3521 dh->length and always used
3523 BN_rand_range(priv_key, dh->p).
3525 BN_rand_range() is not necessary for Diffie-Hellman, and this
3526 specific range makes Diffie-Hellman unnecessarily inefficient if
3527 dh->length (recommended exponent length) is much smaller than the
3528 length of dh->p. We could use BN_rand_range() if the order of
3529 the subgroup was stored in the DH structure, but we only have
3534 BN_rand(priv_key, l, ...)
3536 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3542 RSA_eay_public_encrypt
3543 RSA_eay_private_decrypt
3544 RSA_eay_private_encrypt (signing)
3545 RSA_eay_public_decrypt (signature verification)
3547 (default implementations for RSA_public_encrypt,
3548 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3549 always reject numbers >= n.
3552 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3553 to synchronize access to 'locking_thread'. This is necessary on
3554 systems where access to 'locking_thread' (an 'unsigned long'
3555 variable) is not atomic.
3558 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3559 *before* setting the 'crypto_lock_rand' flag. The previous code had
3560 a race condition if 0 is a valid thread ID.
3561 [Travis Vitek <vitek@roguewave.com>]
3563 *) Add support for shared libraries under Irix.
3564 [Albert Chin-A-Young <china@thewrittenword.com>]
3566 *) Add configuration option to build on Linux on both big-endian and
3568 [Ralf Baechle <ralf@uni-koblenz.de>]
3570 *) Add the possibility to create shared libraries on HP-UX.
3573 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3575 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3576 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3577 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3578 PRNG state recovery was possible based on the output of
3579 one PRNG request appropriately sized to gain knowledge on
3580 'md' followed by enough consecutive 1-byte PRNG requests
3581 to traverse all of 'state'.
3583 1. When updating 'md_local' (the current thread's copy of 'md')
3584 during PRNG output generation, hash all of the previous
3585 'md_local' value, not just the half used for PRNG output.
3587 2. Make the number of bytes from 'state' included into the hash
3588 independent from the number of PRNG bytes requested.
3590 The first measure alone would be sufficient to avoid
3591 Markku-Juhani's attack. (Actually it had never occurred
3592 to me that the half of 'md_local' used for chaining was the
3593 half from which PRNG output bytes were taken -- I had always
3594 assumed that the secret half would be used.) The second
3595 measure makes sure that additional data from 'state' is never
3596 mixed into 'md_local' in small portions; this heuristically
3597 further strengthens the PRNG.
3600 *) Fix crypto/bn/asm/mips3.s.
3603 *) When only the key is given to "enc", the IV is undefined. Print out
3604 an error message in this case.
3607 *) Handle special case when X509_NAME is empty in X509 printing routines.
3610 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3611 positive and less than q.
3614 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3615 used: it isn't thread safe and the add_lock_callback should handle
3617 [Paul Rose <Paul.Rose@bridge.com>]
3619 *) Verify that incoming data obeys the block size in
3620 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3624 [Ulf Möller, Bodo Möller]
3626 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3627 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3628 when fixing the server behaviour for backwards-compatible 'client
3629 hello' messages. (Note that the attack is impractical against
3630 SSL 3.0 and TLS 1.0 anyway because length and version checking
3631 means that the probability of guessing a valid ciphertext is
3632 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3635 Before 0.9.5, the countermeasure (hide the error by generating a
3636 random 'decryption result') did not work properly because
3637 ERR_clear_error() was missing, meaning that SSL_get_error() would
3638 detect the supposedly ignored error.
3640 Both problems are now fixed.
3643 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3644 (previously it was 1024).
3647 *) Fix for compatibility mode trust settings: ignore trust settings
3648 unless some valid trust or reject settings are present.
3651 *) Fix for blowfish EVP: its a variable length cipher.
3654 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3655 parameters in DSA public key structures and return an error in the
3656 DSA routines if parameters are absent.
3659 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3660 in the current directory if neither $RANDFILE nor $HOME was set.
3661 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3662 caused some confusion to Windows users who haven't defined $HOME.
3663 Thus RAND_file_name() is changed again: e_os.h can define a
3664 DEFAULT_HOME, which will be used if $HOME is not set.
3665 For Windows, we use "C:"; on other platforms, we still require
3666 environment variables.
3668 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3669 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3670 having multiple threads call RAND_poll() concurrently.
3673 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3674 combination of a flag and a thread ID variable.
3675 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3676 flag), *other* threads can enter ssleay_add_bytes without obeying
3677 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3678 that they do not hold after the first thread unsets add_do_not_lock).
3681 *) Change bctest again: '-x' expressions are not available in all
3685 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3687 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3688 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3690 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3691 the default extension for executables, if any. Also, make the perl
3692 scripts that use symlink() to test if it really exists and use "cp"
3693 if it doesn't. All this made OpenSSL compilable and installable in
3697 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3698 If SEQUENCE is length is indefinite just set c->slen to the total
3699 amount of data available.
3700 [Steve Henson, reported by shige@FreeBSD.org]
3701 [This change does not apply to 0.9.7.]
3703 *) Change bctest to avoid here-documents inside command substitution
3704 (workaround for FreeBSD /bin/sh bug).
3705 For compatibility with Ultrix, avoid shell functions (introduced
3706 in the bctest version that searches along $PATH).
3709 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3710 with des_encrypt() defined on some operating systems, like Solaris
3714 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3715 On the Importance of Eliminating Errors in Cryptographic
3716 Computations, J. Cryptology 14 (2001) 2, 101-119,
3717 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3720 *) MIPS assembler BIGNUM division bug fix.
3723 *) Disabled incorrect Alpha assembler code.
3726 *) Fix PKCS#7 decode routines so they correctly update the length
3727 after reading an EOC for the EXPLICIT tag.
3729 [This change does not apply to 0.9.7.]
3731 *) Fix bug in PKCS#12 key generation routines. This was triggered
3732 if a 3DES key was generated with a 0 initial byte. Include
3733 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3734 (but broken) behaviour.
3737 *) Enhance bctest to search for a working bc along $PATH and print
3739 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3741 *) Fix memory leaks in err.c: free err_data string if necessary;
3742 don't write to the wrong index in ERR_set_error_data.
3745 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3749 *) Replace rdtsc with _emit statements for VC++ version 5.
3750 [Jeremy Cooper <jeremy@baymoo.org>]
3752 *) Make it possible to reuse SSLv2 sessions.
3755 *) In copy_email() check for >= 0 as a return value for
3756 X509_NAME_get_index_by_NID() since 0 is a valid index.
3757 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3759 *) Avoid coredump with unsupported or invalid public keys by checking if
3760 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3761 PKCS7_verify() fails with non detached data.
3764 *) Don't use getenv in library functions when run as setuid/setgid.
3765 New function OPENSSL_issetugid().
3768 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3769 due to incorrect handling of multi-threading:
3771 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3773 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3775 3. Count how many times MemCheck_off() has been called so that
3776 nested use can be treated correctly. This also avoids
3777 inband-signalling in the previous code (which relied on the
3778 assumption that thread ID 0 is impossible).
3781 *) Add "-rand" option also to s_client and s_server.
3784 *) Fix CPU detection on Irix 6.x.
3785 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3786 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3788 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3791 [This change does not apply to 0.9.7.]
3793 *) Use the cached encoding of an X509_NAME structure rather than
3794 copying it. This is apparently the reason for the libsafe "errors"
3795 but the code is actually correct.
3798 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3799 Bleichenbacher's DSA attack.
3800 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3801 to be set and top=0 forces the highest bit to be set; top=-1 is new
3802 and leaves the highest bit random.
3803 [Ulf Moeller, Bodo Moeller]
3805 *) In the NCONF_...-based implementations for CONF_... queries
3806 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3807 a temporary CONF structure with the data component set to NULL
3808 (which gives segmentation faults in lh_retrieve).
3809 Instead, use NULL for the CONF pointer in CONF_get_string and
3810 CONF_get_number (which may use environment variables) and directly
3811 return NULL from CONF_get_section.
3814 *) Fix potential buffer overrun for EBCDIC.
3817 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3818 keyUsage if basicConstraints absent for a CA.
3821 *) Make SMIME_write_PKCS7() write mail header values with a format that
3822 is more generally accepted (no spaces before the semicolon), since
3823 some programs can't parse those values properly otherwise. Also make
3824 sure BIO's that break lines after each write do not create invalid
3828 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3829 macros previously used would not encode an empty SEQUENCE OF
3830 and break the signature.
3832 [This change does not apply to 0.9.7.]
3834 *) Zero the premaster secret after deriving the master secret in
3838 *) Add some EVP_add_digest_alias registrations (as found in
3839 OpenSSL_add_all_digests()) to SSL_library_init()
3840 aka OpenSSL_add_ssl_algorithms(). This provides improved
3841 compatibility with peers using X.509 certificates
3842 with unconventional AlgorithmIdentifier OIDs.
3845 *) Fix for Irix with NO_ASM.
3846 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3848 *) ./config script fixes.
3849 [Ulf Moeller, Richard Levitte]
3851 *) Fix 'openssl passwd -1'.
3854 *) Change PKCS12_key_gen_asc() so it can cope with non null
3855 terminated strings whose length is passed in the passlen
3856 parameter, for example from PEM callbacks. This was done
3857 by adding an extra length parameter to asc2uni().
3858 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3860 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3861 call failed, free the DSA structure.
3864 *) Fix to uni2asc() to cope with zero length Unicode strings.
3865 These are present in some PKCS#12 files.
3868 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3869 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3870 when writing a 32767 byte record.
3871 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3873 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3874 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3876 (RSA objects have a reference count access to which is protected
3877 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3878 so they are meant to be shared between threads.)
3879 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3880 "Reddie, Steven" <Steven.Reddie@ca.com>]
3882 *) Fix a deadlock in CRYPTO_mem_leaks().
3885 *) Use better test patterns in bntest.
3888 *) rand_win.c fix for Borland C.
3891 *) BN_rshift bugfix for n == 0.
3894 *) Add a 'bctest' script that checks for some known 'bc' bugs
3895 so that 'make test' does not abort just because 'bc' is broken.
3898 *) Store verify_result within SSL_SESSION also for client side to
3899 avoid potential security hole. (Re-used sessions on the client side
3900 always resulted in verify_result==X509_V_OK, not using the original
3901 result of the server certificate verification.)
3904 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3905 SSL3_RT_APPLICATION_DATA, return 0.
3906 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3910 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3911 releases, have been re-implemented by renaming the previous
3912 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3913 and ssl3_read_internal, respectively, and adding 'peek' parameters
3914 to them. The new ssl[23]_{read,peek} functions are calls to
3915 ssl[23]_read_internal with the 'peek' flag set appropriately.
3916 A 'peek' parameter has also been added to ssl3_read_bytes, which
3917 does the actual work for ssl3_read_internal.
3920 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3921 the method-specific "init()" handler. Also clean up ex_data after
3922 calling the method-specific "finish()" handler. Previously, this was
3923 happening the other way round.
3926 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3927 The previous value, 12, was not always sufficient for BN_mod_exp().
3930 *) Make sure that shared libraries get the internal name engine with
3931 the full version number and not just 0. This should mark the
3932 shared libraries as not backward compatible. Of course, this should
3933 be changed again when we can guarantee backward binary compatibility.
3936 *) Fix typo in get_cert_by_subject() in by_dir.c
3937 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3939 *) Rework the system to generate shared libraries:
3941 - Make note of the expected extension for the shared libraries and
3942 if there is a need for symbolic links from for example libcrypto.so.0
3943 to libcrypto.so.0.9.7. There is extended info in Configure for
3946 - Make as few rebuilds of the shared libraries as possible.
3948 - Still avoid linking the OpenSSL programs with the shared libraries.
3950 - When installing, install the shared libraries separately from the
3954 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3956 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3957 and not in SSL_clear because the latter is also used by the
3958 accept/connect functions; previously, the settings made by
3959 SSL_set_read_ahead would be lost during the handshake.
3960 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3962 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3963 Previously, it would create entries for disableed algorithms no
3967 *) Added several new manual pages for SSL_* function.
3970 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3972 *) In ssl23_get_client_hello, generate an error message when faced
3973 with an initial SSL 3.0/TLS record that is too small to contain the
3974 first two bytes of the ClientHello message, i.e. client_version.
3975 (Note that this is a pathologic case that probably has never happened
3976 in real life.) The previous approach was to use the version number
3977 from the record header as a substitute; but our protocol choice
3978 should not depend on that one because it is not authenticated
3979 by the Finished messages.
3982 *) More robust randomness gathering functions for Windows.
3983 [Jeffrey Altman <jaltman@columbia.edu>]
3985 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3986 not set then we don't setup the error code for issuer check errors
3987 to avoid possibly overwriting other errors which the callback does
3988 handle. If an application does set the flag then we assume it knows
3989 what it is doing and can handle the new informational codes
3993 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3994 a general "ANY" type, as such it should be able to decode anything
3995 including tagged types. However it didn't check the class so it would
3996 wrongly interpret tagged types in the same way as their universal
3997 counterpart and unknown types were just rejected. Changed so that the
3998 tagged and unknown types are handled in the same way as a SEQUENCE:
3999 that is the encoding is stored intact. There is also a new type
4000 "V_ASN1_OTHER" which is used when the class is not universal, in this
4001 case we have no idea what the actual type is so we just lump them all
4005 *) On VMS, stdout may very well lead to a file that is written to
4006 in a record-oriented fashion. That means that every write() will
4007 write a separate record, which will be read separately by the
4008 programs trying to read from it. This can be very confusing.
4010 The solution is to put a BIO filter in the way that will buffer
4011 text until a linefeed is reached, and then write everything a
4012 line at a time, so every record written will be an actual line,
4013 not chunks of lines and not (usually doesn't happen, but I've
4014 seen it once) several lines in one record. BIO_f_linebuffer() is
4017 Currently, it's a VMS-only method, because that's where it has
4018 been tested well enough.
4021 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4022 it can return incorrect results.
4023 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4024 but it was in 0.9.6-beta[12].)
4027 *) Disable the check for content being present when verifying detached
4028 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4029 include zero length content when signing messages.
4032 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4033 BIO_ctrl (for BIO pairs).
4036 *) Add DSO method for VMS.
4039 *) Bug fix: Montgomery multiplication could produce results with the
4043 *) Add RPM specification openssl.spec and modify it to build three
4044 packages. The default package contains applications, application
4045 documentation and run-time libraries. The devel package contains
4046 include files, static libraries and function documentation. The
4047 doc package contains the contents of the doc directory. The original
4048 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4051 *) Add a large number of documentation files for many SSL routines.
4052 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4054 *) Add a configuration entry for Sony News 4.
4055 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4057 *) Don't set the two most significant bits to one when generating a
4058 random number < q in the DSA library.
4061 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4062 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4063 the underlying transport is blocking) if a handshake took place.
4064 (The default behaviour is needed by applications such as s_client
4065 and s_server that use select() to determine when to use SSL_read;
4066 but for applications that know in advance when to expect data, it
4067 just makes things more complicated.)
4070 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4074 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4075 work better on such systems.
4076 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4078 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4079 Update PKCS12_parse() so it copies the friendlyName and the
4080 keyid to the certificates aux info.
4083 *) Fix bug in PKCS7_verify() which caused an infinite loop
4084 if there was more than one signature.
4085 [Sven Uszpelkat <su@celocom.de>]
4087 *) Major change in util/mkdef.pl to include extra information
4088 about each symbol, as well as presentig variables as well
4089 as functions. This change means that there's n more need
4090 to rebuild the .num files when some algorithms are excluded.
4093 *) Allow the verify time to be set by an application,
4094 rather than always using the current time.
4097 *) Phase 2 verify code reorganisation. The certificate
4098 verify code now looks up an issuer certificate by a
4099 number of criteria: subject name, authority key id
4100 and key usage. It also verifies self signed certificates
4101 by the same criteria. The main comparison function is
4102 X509_check_issued() which performs these checks.
4104 Lot of changes were necessary in order to support this
4105 without completely rewriting the lookup code.
4107 Authority and subject key identifier are now cached.
4109 The LHASH 'certs' is X509_STORE has now been replaced
4110 by a STACK_OF(X509_OBJECT). This is mainly because an
4111 LHASH can't store or retrieve multiple objects with
4112 the same hash value.
4114 As a result various functions (which were all internal
4115 use only) have changed to handle the new X509_STORE
4116 structure. This will break anything that messed round
4117 with X509_STORE internally.
4119 The functions X509_STORE_add_cert() now checks for an
4120 exact match, rather than just subject name.
4122 The X509_STORE API doesn't directly support the retrieval
4123 of multiple certificates matching a given criteria, however
4124 this can be worked round by performing a lookup first
4125 (which will fill the cache with candidate certificates)
4126 and then examining the cache for matches. This is probably
4127 the best we can do without throwing out X509_LOOKUP
4128 entirely (maybe later...).
4130 The X509_VERIFY_CTX structure has been enhanced considerably.
4132 All certificate lookup operations now go via a get_issuer()
4133 callback. Although this currently uses an X509_STORE it
4134 can be replaced by custom lookups. This is a simple way
4135 to bypass the X509_STORE hackery necessary to make this
4136 work and makes it possible to use more efficient techniques
4137 in future. A very simple version which uses a simple
4138 STACK for its trusted certificate store is also provided
4139 using X509_STORE_CTX_trusted_stack().
4141 The verify_cb() and verify() callbacks now have equivalents
4142 in the X509_STORE_CTX structure.
4144 X509_STORE_CTX also has a 'flags' field which can be used
4145 to customise the verify behaviour.
4148 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4149 excludes S/MIME capabilities.
4152 *) When a certificate request is read in keep a copy of the
4153 original encoding of the signed data and use it when outputing
4154 again. Signatures then use the original encoding rather than
4155 a decoded, encoded version which may cause problems if the
4156 request is improperly encoded.
4159 *) For consistency with other BIO_puts implementations, call
4160 buffer_write(b, ...) directly in buffer_puts instead of calling
4163 In BIO_puts, increment b->num_write as in BIO_write.
4164 [Peter.Sylvester@EdelWeb.fr]
4166 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4167 BN_zero, we may not return a BIGNUM with an array consisting of
4171 *) Avoid calling abort() from within the library when problems are
4172 detected, except if preprocessor symbols have been defined
4173 (such as REF_CHECK, BN_DEBUG etc.).
4176 *) New openssl application 'rsautl'. This utility can be
4177 used for low level RSA operations. DER public key
4178 BIO/fp routines also added.
4181 *) New Configure entry and patches for compiling on QNX 4.
4182 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4184 *) A demo state-machine implementation was sponsored by
4185 Nuron (http://www.nuron.com/) and is now available in
4186 demos/state_machine.
4189 *) New options added to the 'dgst' utility for signature
4190 generation and verification.
4193 *) Unrecognized PKCS#7 content types are now handled via a
4194 catch all ASN1_TYPE structure. This allows unsupported
4195 types to be stored as a "blob" and an application can
4196 encode and decode it manually.
4199 *) Fix various signed/unsigned issues to make a_strex.c
4201 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4203 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4204 length if passed a buffer. ASN1_INTEGER_to_BN failed
4205 if passed a NULL BN and its argument was negative.
4206 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4208 *) Modification to PKCS#7 encoding routines to output definite
4209 length encoding. Since currently the whole structures are in
4210 memory there's not real point in using indefinite length
4211 constructed encoding. However if OpenSSL is compiled with
4212 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4215 *) Added BIO_vprintf() and BIO_vsnprintf().
4218 *) Added more prefixes to parse for in the the strings written
4219 through a logging bio, to cover all the levels that are available
4220 through syslog. The prefixes are now:
4222 PANIC, EMERG, EMR => LOG_EMERG
4223 ALERT, ALR => LOG_ALERT
4224 CRIT, CRI => LOG_CRIT
4225 ERROR, ERR => LOG_ERR
4226 WARNING, WARN, WAR => LOG_WARNING
4227 NOTICE, NOTE, NOT => LOG_NOTICE
4228 INFO, INF => LOG_INFO
4229 DEBUG, DBG => LOG_DEBUG
4231 and as before, if none of those prefixes are present at the
4232 beginning of the string, LOG_ERR is chosen.
4234 On Win32, the LOG_* levels are mapped according to this:
4236 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4237 LOG_WARNING => EVENTLOG_WARNING_TYPE
4238 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4242 *) Made it possible to reconfigure with just the configuration
4243 argument "reconf" or "reconfigure". The command line arguments
4244 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4245 and are retrieved from there when reconfiguring.
4249 [Assar Westerlund <assar@sics.se>, Richard Levitte]
projects / openssl.git / blob