5 Changes between 0.9.6 and 0.9.7 [xx XXX 2001]
7 Both OpenSSL 0.9.6a (bugfix release, 5 Apr 2001) and OpenSSL 0.9.7
8 are based on OpenSSL 0.9.6.
9 Change log entries are tagged as follows:
10 -) applies to 0.9.6a (/0.9.6b) only
11 *) applies to 0.9.6a (/0.9.6b) and 0.9.7
12 +) applies to 0.9.7 only
14 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15 positive and less than q.
18 +) Enhance the general user interface with mechanisms for inner control
19 and with pssibilities to have yes/no kind of prompts.
22 *) Change all calls to low level digest routines in the library and
23 applications to use EVP. Add missing calls to HMAC_cleanup() and
24 don't assume HMAC_CTX can be copied using memcpy().
25 [Verdon Walker <VWalker@novell.com>, Steve Henson]
27 +) Add the possibility to control engines through control names but with
28 arbitrary arguments instead of just a string.
29 Change the key loaders to take a UI_METHOD instead of a callback
30 function pointer. NOTE: this breaks binary compatibility with earlier
31 versions of OpenSSL [engine].
32 Addapt the nCipher code for these new conditions and add a card insertion
36 +) Enhance the general user interface with mechanisms to better support
37 dialog box interfaces, application-defined prompts, the possibility
38 to use defaults (for example default passwords from somewhere else)
39 and interrupts/cancelations.
42 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
43 used: it isn't thread safe and the add_lock_callback should handle
45 [Paul Rose <Paul.Rose@bridge.com>]
47 *) Verify that incoming data obeys the block size in
48 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
51 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
52 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
56 [Ulf Möller, Bodo Möller]
58 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
59 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5
60 when fixing the server behaviour for backwards-compatible 'client
61 hello' messages. (Note that the attack is impractical against
62 SSL 3.0 and TLS 1.0 anyway because length and version checking
63 means that the probability of guessing a valid ciphertext is
64 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
67 Before 0.9.5, the countermeasure (hide the error by generating a
68 random 'decryption result') did not work properly because
69 ERR_clear_error() was missing, meaning that SSL_get_error() would
70 detect the supposedly ignored error.
72 Both problems are now fixed.
75 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
76 (previously it was 1024).
79 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
80 tidy up some unecessarily weird code in 'sk_new()').
81 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
83 +) Change the key loading routines for ENGINEs to use the same kind
84 callback (pem_password_cb) as all other routines that need this
88 *) Fix for compatibility mode trust settings: ignore trust settings
89 unless some valid trust or reject settings are present.
92 *) Fix for blowfish EVP: its a variable length cipher.
95 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
96 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
97 than this minimum value is recommended.
100 +) New random seeder for OpenVMS, using the system process statistics
101 that are easily reachable.
104 +) Windows apparently can't transparently handle global
105 variables defined in DLLs. Initialisations such as:
107 const ASN1_ITEM *it = &ASN1_INTEGER_it;
109 wont compile. This is used by the any applications that need to
110 delcare their own ASN1 modules. This was fixed by adding the option
111 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
112 needed for static libraries under Win32.
115 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
116 setting of purpose and trust fields. New X509_STORE trust and
117 purpose functions and tidy up setting in other SSL functions.
120 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
121 structure. These are inherited by X509_STORE_CTX when it is
122 initialised. This allows various defaults to be set in the
123 X509_STORE structure (such as flags for CRL checking and custom
124 purpose or trust settings) for functions which only use X509_STORE_CTX
125 internally such as S/MIME.
127 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
128 trust settings if they are not set in X509_STORE. This allows X509_STORE
129 purposes and trust (in S/MIME for example) to override any set by default.
131 Add command line options for CRL checking to smime, s_client and s_server
135 +) Initial CRL based revocation checking. If the CRL checking flag(s)
136 are set then the CRL is looked up in the X509_STORE structure and
137 its validity and signature checked, then if the certificate is found
138 in the CRL the verify fails with a revoked error.
140 Various new CRL related callbacks added to X509_STORE_CTX structure.
142 Command line options added to 'verify' application to support this.
144 This needs some additional work, such as being able to handle multiple
145 CRLs with different times, extension based lookup (rather than just
146 by subject name) and ultimately more complete V2 CRL extension
150 +) Add a general user interface API (crypto/ui/). This is designed
151 to replace things like des_read_password and friends (backward
152 compatibility functions using this new API are provided).
153 The purpose is to remove prompting functions from the DES code
154 section as well as provide for prompting through dialog boxes in
155 a window system and the like.
158 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
159 in the current directory if neither $RANDFILE nor $HOME was set.
160 RAND_file_name() in 0.9.6a returned NULL in this case. This has
161 caused some confusion to Windows users who haven't defined $HOME.
162 Thus RAND_file_name() is changed again: e_os.h can define a
163 DEFAULT_HOME, which will be used if $HOME is not set.
164 For Windows, we use "C:"; on other platforms, we still require
165 environment variables.
167 +) Add "ex_data" support to ENGINE so implementations can add state at a
168 per-structure level rather than having to store it globally.
171 +) Make it possible for ENGINE structures to be copied when retrieved by
172 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
173 This causes the "original" ENGINE structure to act like a template,
174 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
175 operational state can be localised to each ENGINE structure, despite the
176 fact they all share the same "methods". New ENGINE structures returned in
177 this case have no functional references and the return value is the single
178 structural reference. This matches the single structural reference returned
179 by ENGINE_by_id() normally, when it is incremented on the pre-existing
183 *) Fix various bugs related to DSA S/MIME verification. Handle missing
184 parameters in DSA public key structures and return an error in the
185 DSA routines if parameters are absent.
188 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
189 needs to match any other type at all we need to manually clear the
193 +) Changes to the "openssl engine" utility to include;
194 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
195 about an ENGINE's available control commands.
196 - executing control commands from command line arguments using the
197 '-pre' and '-post' switches. '-post' is only used if '-t' is
198 specified and the ENGINE is successfully initialised. The syntax for
199 the individual commands are colon-separated, for example;
200 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
203 +) New dynamic control command support for ENGINEs. ENGINEs can now
204 declare their own commands (numbers), names (strings), descriptions,
205 and input types for run-time discovery by calling applications. A
206 subset of these commands are implicitly classed as "executable"
207 depending on their input type, and only these can be invoked through
208 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
209 can be based on user input, config files, etc). The distinction is
210 that "executable" commands cannot return anything other than a boolean
211 result and can only support numeric or string input, whereas some
212 discoverable commands may only be for direct use through
213 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
214 pointers, or other custom uses. The "executable" commands are to
215 support parameterisations of ENGINE behaviour that can be
216 unambiguously defined by ENGINEs and used consistently across any
217 OpenSSL-based application. Commands have been added to all the
218 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
219 control over shared-library paths without source code alterations.
222 +) Changed all ENGINE implementations to dynamically allocate their
223 ENGINEs rather than declaring them statically. Apart from this being
224 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
225 this also allows the implementations to compile without using the
226 internal engine_int.h header.
229 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
230 'const' value. Any code that should be able to modify a RAND_METHOD
231 should already have non-const pointers to it (ie. they should only
232 modify their own ones).
235 +) Made a variety of little tweaks to the ENGINE code.
236 - "atalla" and "ubsec" string definitions were moved from header files
237 to C code. "nuron" string definitions were placed in variables
238 rather than hard-coded - allowing parameterisation of these values
239 later on via ctrl() commands.
240 - Removed unused "#if 0"'d code.
241 - Fixed engine list iteration code so it uses ENGINE_free() to release
242 structural references.
243 - Constified the RAND_METHOD element of ENGINE structures.
244 - Constified various get/set functions as appropriate and added
245 missing functions (including a catch-all ENGINE_cpy that duplicates
246 all ENGINE values onto a new ENGINE except reference counts/state).
247 - Removed NULL parameter checks in get/set functions. Setting a method
248 or function to NULL is a way of cancelling out a previously set
249 value. Passing a NULL ENGINE parameter is just plain stupid anyway
250 and doesn't justify the extra error symbols and code.
251 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
252 flags from engine_int.h to engine.h.
253 - Changed prototypes for ENGINE handler functions (init(), finish(),
254 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
257 *) Move 'if (!initialized) RAND_poll()' into regions protected by
258 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
259 having multiple threads call RAND_poll() concurrently.
262 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
263 combination of a flag and a thread ID variable.
264 Otherwise while one thread is in ssleay_rand_bytes (which sets the
265 flag), *other* threads can enter ssleay_add_bytes without obeying
266 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock
267 that they do not hold after the first thread unsets add_do_not_lock).
270 +) Implement binary inversion algorithm for BN_mod_inverse in addition
271 to the algorithm using long divison. The binary algorithm can be
272 used only if the modulus is odd. On 32-bit systems, it is faster
273 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
274 roughly 5-15% for 256-bit moduli), so we use it only for moduli
275 up to 450 bits. In 64-bit environments, the binary algorithm
276 appears to be advantageous for much longer moduli; here we use it
277 for moduli up to 2048 bits.
280 *) Change bctest again: '-x' expressions are not available in all
284 -) OpenSSL 0.9.6a released [5 Apr 2001]
286 *) Fix a couple of memory leaks in PKCS7_dataDecode()
287 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
289 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
290 the default extension for executables, if any. Also, make the perl
291 scripts that use symlink() to test if it really exists and use "cp"
292 if it doesn't. All this made OpenSSL compilable and installable in
296 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
297 could not support the combine flag in choice fields.
300 -) Fix for asn1_GetSequence() for indefinite length constructed data.
301 If SEQUENCE is length is indefinite just set c->slen to the total
302 amount of data available.
303 [Steve Henson, reported by shige@FreeBSD.org]
304 [This change does not apply to 0.9.7.]
306 *) Change bctest to avoid here-documents inside command substitution
307 (workaround for FreeBSD /bin/sh bug).
308 For compatibility with Ultrix, avoid shell functions (introduced
309 in the bctest version that searches along $PATH).
312 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
313 with des_encrypt() defined on some operating systems, like Solaris
317 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
318 On the Importance of Eliminating Errors in Cryptographic
319 Computations, J. Cryptology 14 (2001) 2, 101-119,
320 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
323 *) MIPS assembler BIGNUM division bug fix.
326 *) Disabled incorrect Alpha assembler code.
329 -) Fix PKCS#7 decode routines so they correctly update the length
330 after reading an EOC for the EXPLICIT tag.
332 [This change does not apply to 0.9.7.]
334 *) Fix bug in PKCS#12 key generation routines. This was triggered
335 if a 3DES key was generated with a 0 initial byte. Include
336 PKCS12_BROKEN_KEYGEN compilation option to retain the old
337 (but broken) behaviour.
340 *) Enhance bctest to search for a working bc along $PATH and print
342 [Tim Rice <tim@multitalents.net> via Richard Levitte]
344 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
345 extensions from a certificate request to the certificate.
348 +) Allow multiple 'certopt' and 'nameopt' options to be separated
349 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
350 file: this allows the display of the certificate about to be
351 signed to be customised, to allow certain fields to be included
352 or excluded and extension details. The old system didn't display
353 multicharacter strings properly, omitted fields not in the policy
354 and couldn't display additional details such as extensions.
357 *) Fix memory leaks in err.c: free err_data string if necessary;
358 don't write to the wrong index in ERR_set_error_data.
361 +) Function EC_POINTs_mul for simultaneous scalar multiplication
362 of an arbitrary number of elliptic curve points, optionally
363 including the generator defined for the EC_GROUP.
364 EC_POINT_mul is a simple wrapper function for the typical case
365 that the point list has just one item (besides the optional
369 +) First EC_METHODs for curves over GF(p):
371 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
372 operations and provides various method functions that can also
373 operate with faster implementations of modular arithmetic.
375 EC_GFp_mont_method() reuses most functions that are part of
376 EC_GFp_simple_method, but uses Montgomery arithmetic.
378 [Bodo Moeller; point addition and point doubling
379 implementation directly derived from source code provided by
380 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
382 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
385 Curves are EC_GROUP objects (with an optional group generator)
386 based on EC_METHODs that are built into the library.
388 Points are EC_POINT objects based on EC_GROUP objects.
390 Most of the framework would be able to handle curves over arbitrary
391 finite fields, but as there are no obvious types for fields other
392 than GF(p), some functions are limited to that for now.
395 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
396 that the file contains a complete HTTP response.
399 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
400 change the def and num file printf format specifier from "%-40sXXX"
401 to "%-39s XXX". The latter will always guarantee a space after the
402 field while the former will cause them to run together if the field
403 is 40 of more characters long.
406 +) Constify the cipher and digest 'method' functions and structures
407 and modify related functions to take constant EVP_MD and EVP_CIPHER
411 *) Implement ssl23_peek (analogous to ssl23_read), which previously
415 *) Replace rdtsc with _emit statements for VC++ version 5.
416 [Jeremy Cooper <jeremy@baymoo.org>]
418 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
419 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
422 +) Modify EVP_Digest*() routines so they now return values. Although the
423 internal software routines can never fail additional hardware versions
427 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
429 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
430 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
433 ERR_R_NESTED_ASN1_ERROR
435 ERR_R_MISSING_ASN1_EOS
436 were 4 .. 9, conflicting with
437 ERR_LIB_RSA (= ERR_R_RSA_LIB)
439 ERR_LIB_PEM (= ERR_R_PEM_LIB).
440 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
442 Add new error code 'ERR_R_INTERNAL_ERROR'.
445 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
449 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
450 sets the subject name for a new request or supersedes the
451 subject name in a given request. Formats that can be parsed are
452 'CN=Some Name, OU=myOU, C=IT'
454 'CN=Some Name/OU=myOU/C=IT'.
456 Add options '-batch' and '-verbose' to 'openssl req'.
457 [Massimiliano Pala <madwolf@hackmasters.net>]
459 *) Make it possible to reuse SSLv2 sessions.
462 +) Introduce the possibility to access global variables through
463 functions on platform were that's the best way to handle exporting
464 global variables in shared libraries. To enable this functionality,
465 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
466 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
467 is normally done by Configure or something similar).
469 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
470 in the source file (foo.c) like this:
472 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
473 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
475 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
476 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
478 OPENSSL_DECLARE_GLOBAL(int,foo);
479 #define foo OPENSSL_GLOBAL_REF(foo)
480 OPENSSL_DECLARE_GLOBAL(double,bar);
481 #define bar OPENSSL_GLOBAL_REF(bar)
483 The #defines are very important, and therefore so is including the
484 header file everywere where the defined globals are used.
486 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
487 of ASN.1 items, but that structure is a bt different.
489 The largest change is in util/mkdef.pl which has been enhanced with
490 better and easier to understand logic to choose which symbols should
491 go into the Windows .def files as well as a number of fixes and code
492 cleanup (among others, algorithm keywords are now sorted
493 lexicographically to avoid constant rewrites).
496 *) In copy_email() check for >= 0 as a return value for
497 X509_NAME_get_index_by_NID() since 0 is a valid index.
498 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
500 +) In BN_div() keep a copy of the sign of 'num' before writing the
501 result to 'rm' because if rm==num the value will be overwritten
502 and produce the wrong result if 'num' is negative: this caused
503 problems with BN_mod() and BN_nnmod().
506 +) Function OCSP_request_verify(). This checks the signature on an
507 OCSP request and verifies the signer certificate. The signer
508 certificate is just checked for a generic purpose and OCSP request
512 +) Add OCSP_check_validity() function to check the validity of OCSP
513 responses. OCSP responses are prepared in real time and may only
514 be a few seconds old. Simply checking that the current time lies
515 between thisUpdate and nextUpdate max reject otherwise valid responses
516 caused by either OCSP responder or client clock innacuracy. Instead
517 we allow thisUpdate and nextUpdate to fall within a certain period of
518 the current time. The age of the response can also optionally be
519 checked. Two new options -validity_period and -status_age added to
523 +) If signature or public key algorithm is unrecognized print out its
524 OID rather that just UNKOWN.
527 *) Avoid coredump with unsupported or invalid public keys by checking if
528 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
529 PKCS7_verify() fails with non detached data.
532 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
533 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
534 ID to be generated from the issuer certificate alone which can then be
535 passed to OCSP_id_issuer_cmp().
538 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
539 ASN1 modules to export functions returning ASN1_ITEM pointers
540 instead of the ASN1_ITEM structures themselves. This adds several
541 new macros which allow the underlying ASN1 function/structure to
542 be accessed transparently. As a result code should not use ASN1_ITEM
543 references directly (such as &X509_it) but instead use the relevant
544 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
545 use of the new ASN1 code on platforms where exporting structures
546 is problematical (for example in shared libraries) but exporting
547 functions returning pointers to structures is not.
550 +) Add support for overriding the generation of SSL/TLS session IDs.
551 These callbacks can be registered either in an SSL_CTX or per SSL.
552 The purpose of this is to allow applications to control, if they wish,
553 the arbitrary values chosen for use as session IDs, particularly as it
554 can be useful for session caching in multiple-server environments. A
555 command-line switch for testing this (and any client code that wishes
556 to use such a feature) has been added to "s_server".
557 [Geoff Thorpe, Lutz Jaenicke]
559 +) Modify mkdef.pl to recognise and parse prprocessor conditionals
560 of the form '#if defined(...) || defined(...) || ...' and
561 '#if !defined(...) && !defined(...) && ...'. This also avoids
562 the growing number of special cases it was previously handling.
565 *) Don't use getenv in library functions when run as setuid/setgid.
566 New function OPENSSL_issetugid().
569 +) Make all configuration macros available for application by making
570 sure they are available in opensslconf.h, by giving them names starting
571 with "OPENSSL_" to avoid conflicts with other packages and by making
572 sure e_os2.h will cover all platform-specific cases together with
574 Additionally, it is now possible to define configuration/platform-
575 specific names (called "system identities"). In the C code, these
576 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
577 macro with the name beginning with "OPENSSL_SYS_", which is determined
578 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
582 +) New option -set_serial to 'req' and 'x509' this allows the serial
583 number to use to be specified on the command line. Previously self
584 signed certificates were hard coded with serial number 0 and the
585 CA options of 'x509' had to use a serial number in a file which was
589 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
590 due to incorrect handling of multi-threading:
592 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
594 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
596 3. Count how many times MemCheck_off() has been called so that
597 nested use can be treated correctly. This also avoids
598 inband-signalling in the previous code (which relied on the
599 assumption that thread ID 0 is impossible).
602 +) New options to 'ca' utility to support V2 CRL entry extensions.
603 Currently CRL reason, invalidity date and hold instruction are
604 supported. Add new CRL extensions to V3 code and some new objects.
607 *) Add "-rand" option also to s_client and s_server.
610 +) New function EVP_CIPHER_CTX_set_padding() this is used to
611 disable standard block padding (aka PKCS#5 padding) in the EVP
612 API, which was previously mandatory. This means that the data is
613 not padded in any way and so the total length much be a multiple
614 of the block size, otherwise an error occurs.
617 +) Initial (incomplete) OCSP SSL support.
620 *) Fix CPU detection on Irix 6.x.
621 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
622 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
624 +) New function OCSP_parse_url(). This splits up a URL into its host,
625 port and path components: primarily to parse OCSP URLs. New -url
626 option to ocsp utility.
629 +) New nonce behavior. The return value of OCSP_check_nonce() now
630 reflects the various checks performed. Applications can decide
631 whether to tolerate certain situations such as an absent nonce
632 in a response when one was present in a request: the ocsp application
633 just prints out a warning. New function OCSP_add1_basic_nonce()
634 this is to allow responders to include a nonce in a response even if
635 the request is nonce-less.
638 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
640 [This change does not apply to 0.9.7.]
642 *) Use the cached encoding of an X509_NAME structure rather than
643 copying it. This is apparently the reason for the libsafe "errors"
644 but the code is actually correct.
647 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
648 skipped when using openssl x509 multiple times on a single input file,
649 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
652 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
653 set string type: to handle setting ASN1_TIME structures. Fix ca
654 utility to correctly initialize revocation date of CRLs.
657 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
658 the clients preferred ciphersuites and rather use its own preferences.
659 Should help to work around M$ SGC (Server Gated Cryptography) bug in
660 Internet Explorer by ensuring unchanged hash method during stepup.
663 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
664 to aes and add a new 'exist' option to print out symbols that don't
668 +) Additional options to ocsp utility to allow flags to be set and
669 additional certificates supplied.
672 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
673 OCSP client a number of certificate to only verify the response
677 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
678 Bleichenbacher's DSA attack.
679 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
680 to be set and top=0 forces the highest bit to be set; top=-1 is new
681 and leaves the highest bit random.
682 [Ulf Moeller, Bodo Moeller]
684 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
685 handle the new API. Currently only ECB, CBC modes supported. Add new
686 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
687 for TLS" draft-ietf-tls-ciphersuite-03.txt.
688 [Ben Laurie, Steve Henson]
690 *) In the NCONF_...-based implementations for CONF_... queries
691 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
692 a temporary CONF structure with the data component set to NULL
693 (which gives segmentation faults in lh_retrieve).
694 Instead, use NULL for the CONF pointer in CONF_get_string and
695 CONF_get_number (which may use environment variables) and directly
696 return NULL from CONF_get_section.
699 *) Fix potential buffer overrun for EBCDIC.
702 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
706 +) Functions for OCSP responders. OCSP_request_onereq_count(),
707 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
708 extract information from a certificate request. OCSP_response_create()
709 creates a response and optionally adds a basic response structure.
710 OCSP_basic_add1_status() adds a complete single response to a basic
711 reponse and returns the OCSP_SINGLERESP structure just added (to allow
712 extensions to be included for example). OCSP_basic_add1_cert() adds a
713 certificate to a basic response and OCSP_basic_sign() signs a basic
714 response with various flags. New helper functions ASN1_TIME_check()
715 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
716 (converts ASN1_TIME to GeneralizedTime).
719 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
720 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
721 structure from a certificate. X509_pubkey_digest() digests tha public_key
722 contents: this is used in various key identifiers.
725 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
726 keyUsage if basicConstraints absent for a CA.
729 *) Make SMIME_write_PKCS7() write mail header values with a format that
730 is more generally accepted (no spaces before the semicolon), since
731 some programs can't parse those values properly otherwise. Also make
732 sure BIO's that break lines after each write do not create invalid
736 +) Make sk_sort() tolerate a NULL argument.
737 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
739 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
740 passed by the function are trusted implicitly. If any of them signed the
741 reponse then it is assumed to be valid and is not verified.
744 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
745 macros previously used would not encode an empty SEQUENCE OF
746 and break the signature.
748 [This change does not apply to 0.9.7.]
750 *) Zero the premaster secret after deriving the master secret in
754 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
755 to data. This was previously part of the PKCS7 ASN1 code. This
756 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
757 [Steve Henson, reported by Kenneth R. Robinette
758 <support@securenetterm.com>]
760 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
761 routines: without these tracing memory leaks is very painful.
762 Fix leaks in PKCS12 and PKCS7 routines.
765 *) Add some EVP_add_digest_alias registrations (as found in
766 OpenSSL_add_all_digests()) to SSL_library_init()
767 aka OpenSSL_add_ssl_algorithms(). This provides improved
768 compatibility with peers using X.509 certificates
769 with unconventional AlgorithmIdentifier OIDs.
772 *) Fix for Irix with NO_ASM.
773 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
775 *) ./config script fixes.
776 [Ulf Moeller, Richard Levitte]
778 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
779 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
780 effectively meant GeneralizedTime would never be used. Now it
781 is initialised to -1 but X509_time_adj() now has to check the value
782 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
783 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
784 [Steve Henson, reported by Kenneth R. Robinette
785 <support@securenetterm.com>]
787 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
788 result in a zero length in the ASN1_INTEGER structure which was
789 not consistent with the structure when d2i_ASN1_INTEGER() was used
790 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
791 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
792 where it did not print out a minus for negative ASN1_INTEGER.
795 *) Fix 'openssl passwd -1'.
798 +) Add summary printout to ocsp utility. The various functions which
799 convert status values to strings have been renamed to:
800 OCSP_response_status_str(), OCSP_cert_status_str() and
801 OCSP_crl_reason_str() and are no longer static. New options
802 to verify nonce values and to disable verification. OCSP response
803 printout format cleaned up.
806 +) Add additional OCSP certificate checks. These are those specified
807 in RFC2560. This consists of two separate checks: the CA of the
808 certificate being checked must either be the OCSP signer certificate
809 or the issuer of the OCSP signer certificate. In the latter case the
810 OCSP signer certificate must contain the OCSP signing extended key
811 usage. This check is performed by attempting to match the OCSP
812 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
813 in the OCSP_CERTID structures of the response.
816 +) Initial OCSP certificate verification added to OCSP_basic_verify()
817 and related routines. This uses the standard OpenSSL certificate
818 verify routines to perform initial checks (just CA validity) and
819 to obtain the certificate chain. Then additional checks will be
820 performed on the chain. Currently the root CA is checked to see
821 if it is explicitly trusted for OCSP signing. This is used to set
822 a root CA as a global signing root: that is any certificate that
823 chains to that CA is an acceptable OCSP signing certificate.
826 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
827 extensions from a separate configuration file.
828 As when reading extensions from the main configuration file,
829 the '-extensions ...' option may be used for specifying the
831 [Massimiliano Pala <madwolf@comune.modena.it>]
833 *) Change PKCS12_key_gen_asc() so it can cope with non null
834 terminated strings whose length is passed in the passlen
835 parameter, for example from PEM callbacks. This was done
836 by adding an extra length parameter to asc2uni().
837 [Steve Henson, reported by <oddissey@samsung.co.kr>]
839 +) New OCSP utility. Allows OCSP requests to be generated or
840 read. The request can be sent to a responder and the output
841 parsed, outputed or printed in text form. Not complete yet:
842 still needs to check the OCSP response validity.
845 +) New subcommands for 'openssl ca':
846 'openssl ca -status <serial>' prints the status of the cert with
847 the given serial number (according to the index file).
848 'openssl ca -updatedb' updates the expiry status of certificates
850 [Massimiliano Pala <madwolf@comune.modena.it>]
852 +) New '-newreq-nodes' command option to CA.pl. This is like
853 '-newreq', but calls 'openssl req' with the '-nodes' option
854 so that the resulting key is not encrypted.
855 [Damien Miller <djm@mindrot.org>]
857 +) New configuration for the GNU Hurd.
858 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
860 +) Initial code to implement OCSP basic response verify. This
861 is currently incomplete. Currently just finds the signer's
862 certificate and verifies the signature on the response.
865 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
866 value of OPENSSLDIR. This is available via the new '-d' option
867 to 'openssl version', and is also included in 'openssl version -a'.
870 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
871 call failed, free the DSA structure.
874 +) Allowing defining memory allocation callbacks that will be given
875 file name and line number information in additional arguments
876 (a const char* and an int). The basic functionality remains, as
877 well as the original possibility to just replace malloc(),
878 realloc() and free() by functions that do not know about these
879 additional arguments. To register and find out the current
880 settings for extended allocation functions, the following
881 functions are provided:
883 CRYPTO_set_mem_ex_functions
884 CRYPTO_set_locked_mem_ex_functions
885 CRYPTO_get_mem_ex_functions
886 CRYPTO_get_locked_mem_ex_functions
888 These work the same way as CRYPTO_set_mem_functions and friends.
889 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
890 extended allocation function is enabled.
891 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
892 a conventional allocation function is enabled.
893 [Richard Levitte, Bodo Moeller]
895 *) Fix to uni2asc() to cope with zero length Unicode strings.
896 These are present in some PKCS#12 files.
899 +) Finish off removing the remaining LHASH function pointer casts.
900 There should no longer be any prototype-casting required when using
901 the LHASH abstraction, and any casts that remain are "bugs". See
902 the callback types and macros at the head of lhash.h for details
903 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
906 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
907 If /dev/[u]random devices are not available or do not return enough
908 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
910 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
911 /etc/entropy will be queried once each in this sequence, quering stops
912 when enough entropy was collected without querying more sockets.
915 +) Change the Unix RAND_poll() variant to be able to poll several
916 random devices, as specified by DEVRANDOM, until a sufficient amount
917 of data has been collected. We spend at most 10 ms on each file
918 (select timeout) and read in non-blocking mode. DEVRANDOM now
919 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
920 (previously it was just the string "/dev/urandom"), so on typical
921 platforms the 10 ms delay will never occur.
922 Also separate out the Unix variant to its own file, rand_unix.c.
923 For VMS, there's a currently-empty rand_vms.c.
926 +) Move OCSP client related routines to ocsp_cl.c. These
927 provide utility functions which an application needing
928 to issue a request to an OCSP responder and analyse the
929 response will typically need: as opposed to those which an
930 OCSP responder itself would need which will be added later.
932 OCSP_request_sign() signs an OCSP request with an API similar
933 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
934 response. OCSP_response_get1_basic() extracts basic response
935 from response. OCSP_resp_find_status(): finds and extracts status
936 information from an OCSP_CERTID structure (which will be created
937 when the request structure is built). These are built from lower
938 level functions which work on OCSP_SINGLERESP structures but
939 wont normally be used unless the application wishes to examine
940 extensions in the OCSP response for example.
942 Replace nonce routines with a pair of functions.
943 OCSP_request_add1_nonce() adds a nonce value and optionally
944 generates a random value. OCSP_check_nonce() checks the
945 validity of the nonce in an OCSP response.
948 +) Change function OCSP_request_add() to OCSP_request_add0_id().
949 This doesn't copy the supplied OCSP_CERTID and avoids the
950 need to free up the newly created id. Change return type
951 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
952 This can then be used to add extensions to the request.
953 Deleted OCSP_request_new(), since most of its functionality
954 is now in OCSP_REQUEST_new() (and the case insensitive name
955 clash) apart from the ability to set the request name which
956 will be added elsewhere.
959 +) Update OCSP API. Remove obsolete extensions argument from
960 various functions. Extensions are now handled using the new
961 OCSP extension code. New simple OCSP HTTP function which
962 can be used to send requests and parse the response.
965 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
966 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
967 uses the special reorder version of SET OF to sort the attributes
968 and reorder them to match the encoded order. This resolves a long
969 standing problem: a verify on a PKCS7 structure just after signing
970 it used to fail because the attribute order did not match the
971 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
972 it uses the received order. This is necessary to tolerate some broken
973 software that does not order SET OF. This is handled by encoding
974 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
975 to produce the required SET OF.
978 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
979 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
980 files to get correct declarations of the ASN.1 item variables.
983 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
984 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
985 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
986 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
987 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
988 ASN1_ITEM and no wrapper functions.
991 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
992 replace the old function pointer based I/O routines. Change most of
993 the *_d2i_bio() and *_d2i_fp() functions to use these.
996 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
997 lines, recognice more "algorithms" that can be deselected, and make
998 it complain about algorithm deselection that isn't recognised.
1001 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1002 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1003 to use new functions. Add NO_ASN1_OLD which can be set to remove
1004 some old style ASN1 functions: this can be used to determine if old
1005 code will still work when these eventually go away.
1008 +) New extension functions for OCSP structures, these follow the
1009 same conventions as certificates and CRLs.
1012 +) New function X509V3_add1_i2d(). This automatically encodes and
1013 adds an extension. Its behaviour can be customised with various
1014 flags to append, replace or delete. Various wrappers added for
1015 certifcates and CRLs.
1018 +) Fix to avoid calling the underlying ASN1 print routine when
1019 an extension cannot be parsed. Correct a typo in the
1020 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1023 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1024 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1025 when writing a 32767 byte record.
1026 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1028 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1029 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1031 (RSA objects have a reference count access to which is protected
1032 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1033 so they are meant to be shared between threads.)
1034 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1035 "Reddie, Steven" <Steven.Reddie@ca.com>]
1037 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1038 entries for variables.
1041 *) Fix a deadlock in CRYPTO_mem_leaks().
1044 +) Add functionality to apps/openssl.c for detecting locking
1045 problems: As the program is single-threaded, all we have
1046 to do is register a locking callback using an array for
1047 storing which locks are currently held by the program.
1050 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1051 SSL_get_ex_data_X509_STORE_idx(), which is used in
1052 ssl_verify_cert_chain() and thus can be called at any time
1053 during TLS/SSL handshakes so that thread-safety is essential.
1054 Unfortunately, the ex_data design is not at all suited
1055 for multi-threaded use, so it probably should be abolished.
1058 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1059 [Broadcom, tweaked and integrated by Geoff Thorpe]
1061 +) Move common extension printing code to new function
1062 X509V3_print_extensions(). Reorganise OCSP print routines and
1063 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1066 +) New function X509_signature_print() to remove duplication in some
1070 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1071 set (this was treated exactly the same as SET OF previously). This
1072 is used to reorder the STACK representing the structure to match the
1073 encoding. This will be used to get round a problem where a PKCS7
1074 structure which was signed could not be verified because the STACK
1075 order did not reflect the encoded order.
1078 +) Reimplement the OCSP ASN1 module using the new code.
1081 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1082 for its ASN1 operations. The old style function pointers still exist
1083 for now but they will eventually go away.
1086 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1087 completely replaces the old ASN1 functionality with a table driven
1088 encoder and decoder which interprets an ASN1_ITEM structure describing
1089 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1090 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1091 has also been converted to the new form.
1094 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1095 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1096 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1097 for negative moduli.
1100 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1101 of not touching the result's sign bit.
1104 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1108 +) Changed the LHASH code to use prototypes for callbacks, and created
1109 macros to declare and implement thin (optionally static) functions
1110 that provide type-safety and avoid function pointer casting for the
1111 type-specific callbacks.
1114 *) Use better test patterns in bntest.
1117 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1119 [Veers Staats <staatsvr@asc.hpc.mil>,
1120 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1122 *) rand_win.c fix for Borland C.
1125 *) BN_rshift bugfix for n == 0.
1128 +) Reformat the FAQ so the different questions and answers can be divided
1129 in sections depending on the subject.
1132 +) Have the zlib compression code load ZLIB.DLL dynamically under
1136 +) New function BN_mod_sqrt for computing square roots modulo a prime
1137 (using the probabilistic Tonelli-Shanks algorithm unless
1138 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1139 be handled deterministically).
1140 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1142 *) Add a 'bctest' script that checks for some known 'bc' bugs
1143 so that 'make test' does not abort just because 'bc' is broken.
1146 *) Store verify_result within SSL_SESSION also for client side to
1147 avoid potential security hole. (Re-used sessions on the client side
1148 always resulted in verify_result==X509_V_OK, not using the original
1149 result of the server certificate verification.)
1152 +) Make BN_mod_inverse faster by explicitly handling small quotients
1153 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1154 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1157 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1158 SSL3_RT_APPLICATION_DATA, return 0.
1159 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1163 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1164 releases, have been re-implemented by renaming the previous
1165 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1166 and ssl3_read_internal, respectively, and adding 'peek' parameters
1167 to them. The new ssl[23]_{read,peek} functions are calls to
1168 ssl[23]_read_internal with the 'peek' flag set appropriately.
1169 A 'peek' parameter has also been added to ssl3_read_bytes, which
1170 does the actual work for ssl3_read_internal.
1173 +) New function BN_kronecker.
1176 +) Fix BN_gcd so that it works on negative inputs; the result is
1177 positive unless both parameters are zero.
1178 Previously something reasonably close to an infinite loop was
1179 possible because numbers could be growing instead of shrinking
1180 in the implementation of Euclid's algorithm.
1183 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1184 sign of the number in question.
1186 Fix BN_is_word(a,w) to work correctly for w == 0.
1188 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1189 because its test if the absolute value of 'a' equals 'w'.
1190 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1191 it exists mostly for use in the implementations of BN_is_zero(),
1192 BN_is_one(), and BN_is_word().
1195 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1196 the method-specific "init()" handler. Also clean up ex_data after
1197 calling the method-specific "finish()" handler. Previously, this was
1198 happening the other way round.
1201 +) New function BN_swap.
1204 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1205 the exponentiation functions are more likely to produce reasonable
1206 results on negative inputs.
1209 +) Change BN_mod_mul so that the result is always non-negative.
1210 Previously, it could be negative if one of the factors was negative;
1211 I don't think anyone really wanted that behaviour.
1214 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1215 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1216 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1217 and add new functions:
1226 BN_mod_lshift1_quick
1230 These functions always generate non-negative results.
1232 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1233 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1235 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1236 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1237 be reduced modulo m.
1238 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1240 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1241 was actually never needed) and in BN_mul(). The removal in BN_mul()
1242 required a small change in bn_mul_part_recursive() and the addition
1243 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1244 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1245 bn_sub_words() and bn_add_words() except they take arrays with
1249 +) In 'openssl passwd', verify passwords read from the terminal
1250 unless the '-salt' option is used (which usually means that
1251 verification would just waste user's time since the resulting
1252 hash is going to be compared with some given password hash)
1253 or the new '-noverify' option is used.
1255 This is an incompatible change, but it does not affect
1256 non-interactive use of 'openssl passwd' (passwords on the command
1257 line, '-stdin' option, '-in ...' option) and thus should not
1261 +) Remove all references to RSAref, since there's no more need for it.
1264 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1265 The previous value, 12, was not always sufficient for BN_mod_exp().
1268 +) Make DSO load along a path given through an environment variable
1269 (SHLIB_PATH) with shl_load().
1272 +) Constify the ENGINE code as a result of BIGNUM constification.
1273 Also constify the RSA code and most things related to it. In a
1274 few places, most notable in the depth of the ASN.1 code, ugly
1275 casts back to non-const were required (to be solved at a later
1279 +) Make it so the openssl application has all engines loaded by default.
1282 +) Constify the BIGNUM routines a little more.
1285 *) Make sure that shared libraries get the internal name engine with
1286 the full version number and not just 0. This should mark the
1287 shared libraries as not backward compatible. Of course, this should
1288 be changed again when we can guarantee backward binary compatibility.
1291 +) Add the following functions:
1293 ENGINE_load_cswift()
1295 ENGINE_load_atalla()
1297 ENGINE_load_builtin_engines()
1299 That way, an application can itself choose if external engines that
1300 are built-in in OpenSSL shall ever be used or not. The benefit is
1301 that applications won't have to be linked with libdl or other dso
1302 libraries unless it's really needed.
1304 Changed 'openssl engine' to load all engines on demand.
1305 Changed the engine header files to avoid the duplication of some
1306 declarations (they differed!).
1309 +) 'openssl engine' can now list capabilities.
1312 +) Better error reporting in 'openssl engine'.
1315 +) Never call load_dh_param(NULL) in s_server.
1318 +) Add engine application. It can currently list engines by name and
1319 identity, and test if they are actually available.
1322 +) Add support for shared libraries under Irix.
1323 [Albert Chin-A-Young <china@thewrittenword.com>]
1325 +) Improve RPM specification file by forcing symbolic linking and making
1326 sure the installed documentation is also owned by root.root.
1327 [Damien Miller <djm@mindrot.org>]
1329 +) Add configuration option to build on Linux on both big-endian and
1331 [Ralf Baechle <ralf@uni-koblenz.de>]
1333 +) Give the OpenSSL applications more possibilities to make use of
1334 keys (public as well as private) handled by engines.
1337 +) Add OCSP code that comes from CertCo.
1340 +) Add VMS support for the Rijndael code.
1343 +) Added untested support for Nuron crypto accelerator.
1346 +) Add support for external cryptographic devices. This code was
1347 previously distributed separately as the "engine" branch.
1348 [Geoff Thorpe, Richard Levitte]
1350 +) Rework the filename-translation in the DSO code. It is now possible to
1351 have far greater control over how a "name" is turned into a filename
1352 depending on the operating environment and any oddities about the
1353 different shared library filenames on each system.
1356 +) Support threads on FreeBSD-elf in Configure.
1359 +) Add the possibility to create shared libraries on HP-UX
1362 +) Fix for SHA1 assembly problem with MASM: it produces
1363 warnings about corrupt line number information when assembling
1364 with debugging information. This is caused by the overlapping
1366 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1369 NCONF_get_number() has no error checking at all. As a replacement,
1370 NCONF_get_number_e() is defined (_e for "error checking") and is
1371 promoted strongly. The old NCONF_get_number is kept around for
1372 binary backward compatibility.
1373 Make it possible for methods to load from something other than a BIO,
1374 by providing a function pointer that is given a name instead of a BIO.
1375 For example, this could be used to load configuration data from an
1379 *) Fix typo in get_cert_by_subject() in by_dir.c
1380 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1382 *) Rework the system to generate shared libraries:
1384 - Make note of the expected extension for the shared libraries and
1385 if there is a need for symbolic links from for example libcrypto.so.0
1386 to libcrypto.so.0.9.7. There is extended info in Configure for
1389 - Make as few rebuilds of the shared libraries as possible.
1391 - Still avoid linking the OpenSSL programs with the shared libraries.
1393 - When installing, install the shared libraries separately from the
1397 +) Fix for non blocking accept BIOs. Added new I/O special reason
1398 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1399 with non blocking I/O was not possible because no retry code was
1400 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1404 +) Added the beginnings of Rijndael support.
1407 +) Fix for bug in DirectoryString mask setting. Add support for
1408 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1409 to allow certificate printing to more controllable, additional
1410 'certopt' option to 'x509' to allow new printing options to be
1414 +) Clean old EAY MD5 hack from e_os.h.
1417 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1419 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1420 and not in SSL_clear because the latter is also used by the
1421 accept/connect functions; previously, the settings made by
1422 SSL_set_read_ahead would be lost during the handshake.
1423 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1425 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1426 Previously, it would create entries for disableed algorithms no
1430 *) Added several new manual pages for SSL_* function.
1433 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1435 *) In ssl23_get_client_hello, generate an error message when faced
1436 with an initial SSL 3.0/TLS record that is too small to contain the
1437 first two bytes of the ClientHello message, i.e. client_version.
1438 (Note that this is a pathologic case that probably has never happened
1439 in real life.) The previous approach was to use the version number
1440 from the record header as a substitute; but our protocol choice
1441 should not depend on that one because it is not authenticated
1442 by the Finished messages.
1445 *) More robust randomness gathering functions for Windows.
1446 [Jeffrey Altman <jaltman@columbia.edu>]
1448 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1449 not set then we don't setup the error code for issuer check errors
1450 to avoid possibly overwriting other errors which the callback does
1451 handle. If an application does set the flag then we assume it knows
1452 what it is doing and can handle the new informational codes
1456 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1457 a general "ANY" type, as such it should be able to decode anything
1458 including tagged types. However it didn't check the class so it would
1459 wrongly interpret tagged types in the same way as their universal
1460 counterpart and unknown types were just rejected. Changed so that the
1461 tagged and unknown types are handled in the same way as a SEQUENCE:
1462 that is the encoding is stored intact. There is also a new type
1463 "V_ASN1_OTHER" which is used when the class is not universal, in this
1464 case we have no idea what the actual type is so we just lump them all
1468 *) On VMS, stdout may very well lead to a file that is written to
1469 in a record-oriented fashion. That means that every write() will
1470 write a separate record, which will be read separately by the
1471 programs trying to read from it. This can be very confusing.
1473 The solution is to put a BIO filter in the way that will buffer
1474 text until a linefeed is reached, and then write everything a
1475 line at a time, so every record written will be an actual line,
1476 not chunks of lines and not (usually doesn't happen, but I've
1477 seen it once) several lines in one record. BIO_f_linebuffer() is
1480 Currently, it's a VMS-only method, because that's where it has
1481 been tested well enough.
1484 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1485 it can return incorrect results.
1486 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1487 but it was in 0.9.6-beta[12].)
1490 *) Disable the check for content being present when verifying detached
1491 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1492 include zero length content when signing messages.
1495 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1496 BIO_ctrl (for BIO pairs).
1499 *) Add DSO method for VMS.
1502 *) Bug fix: Montgomery multiplication could produce results with the
1506 *) Add RPM specification openssl.spec and modify it to build three
1507 packages. The default package contains applications, application
1508 documentation and run-time libraries. The devel package contains
1509 include files, static libraries and function documentation. The
1510 doc package contains the contents of the doc directory. The original
1511 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1514 *) Add a large number of documentation files for many SSL routines.
1515 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1517 *) Add a configuration entry for Sony News 4.
1518 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1520 *) Don't set the two most significant bits to one when generating a
1521 random number < q in the DSA library.
1524 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1525 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1526 the underlying transport is blocking) if a handshake took place.
1527 (The default behaviour is needed by applications such as s_client
1528 and s_server that use select() to determine when to use SSL_read;
1529 but for applications that know in advance when to expect data, it
1530 just makes things more complicated.)
1533 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1537 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1538 work better on such systems.
1539 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1541 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1542 Update PKCS12_parse() so it copies the friendlyName and the
1543 keyid to the certificates aux info.
1546 *) Fix bug in PKCS7_verify() which caused an infinite loop
1547 if there was more than one signature.
1548 [Sven Uszpelkat <su@celocom.de>]
1550 *) Major change in util/mkdef.pl to include extra information
1551 about each symbol, as well as presentig variables as well
1552 as functions. This change means that there's n more need
1553 to rebuild the .num files when some algorithms are excluded.
1556 *) Allow the verify time to be set by an application,
1557 rather than always using the current time.
1560 *) Phase 2 verify code reorganisation. The certificate
1561 verify code now looks up an issuer certificate by a
1562 number of criteria: subject name, authority key id
1563 and key usage. It also verifies self signed certificates
1564 by the same criteria. The main comparison function is
1565 X509_check_issued() which performs these checks.
1567 Lot of changes were necessary in order to support this
1568 without completely rewriting the lookup code.
1570 Authority and subject key identifier are now cached.
1572 The LHASH 'certs' is X509_STORE has now been replaced
1573 by a STACK_OF(X509_OBJECT). This is mainly because an
1574 LHASH can't store or retrieve multiple objects with
1575 the same hash value.
1577 As a result various functions (which were all internal
1578 use only) have changed to handle the new X509_STORE
1579 structure. This will break anything that messed round
1580 with X509_STORE internally.
1582 The functions X509_STORE_add_cert() now checks for an
1583 exact match, rather than just subject name.
1585 The X509_STORE API doesn't directly support the retrieval
1586 of multiple certificates matching a given criteria, however
1587 this can be worked round by performing a lookup first
1588 (which will fill the cache with candidate certificates)
1589 and then examining the cache for matches. This is probably
1590 the best we can do without throwing out X509_LOOKUP
1591 entirely (maybe later...).
1593 The X509_VERIFY_CTX structure has been enhanced considerably.
1595 All certificate lookup operations now go via a get_issuer()
1596 callback. Although this currently uses an X509_STORE it
1597 can be replaced by custom lookups. This is a simple way
1598 to bypass the X509_STORE hackery necessary to make this
1599 work and makes it possible to use more efficient techniques
1600 in future. A very simple version which uses a simple
1601 STACK for its trusted certificate store is also provided
1602 using X509_STORE_CTX_trusted_stack().
1604 The verify_cb() and verify() callbacks now have equivalents
1605 in the X509_STORE_CTX structure.
1607 X509_STORE_CTX also has a 'flags' field which can be used
1608 to customise the verify behaviour.
1611 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1612 excludes S/MIME capabilities.
1615 *) When a certificate request is read in keep a copy of the
1616 original encoding of the signed data and use it when outputing
1617 again. Signatures then use the original encoding rather than
1618 a decoded, encoded version which may cause problems if the
1619 request is improperly encoded.
1622 *) For consistency with other BIO_puts implementations, call
1623 buffer_write(b, ...) directly in buffer_puts instead of calling
1626 In BIO_puts, increment b->num_write as in BIO_write.
1627 [Peter.Sylvester@EdelWeb.fr]
1629 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1630 BN_zero, we may not return a BIGNUM with an array consisting of
1634 *) Avoid calling abort() from within the library when problems are
1635 detected, except if preprocessor symbols have been defined
1636 (such as REF_CHECK, BN_DEBUG etc.).
1639 *) New openssl application 'rsautl'. This utility can be
1640 used for low level RSA operations. DER public key
1641 BIO/fp routines also added.
1644 *) New Configure entry and patches for compiling on QNX 4.
1645 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1647 *) A demo state-machine implementation was sponsored by
1648 Nuron (http://www.nuron.com/) and is now available in
1649 demos/state_machine.
1652 *) New options added to the 'dgst' utility for signature
1653 generation and verification.
1656 *) Unrecognized PKCS#7 content types are now handled via a
1657 catch all ASN1_TYPE structure. This allows unsupported
1658 types to be stored as a "blob" and an application can
1659 encode and decode it manually.
1662 *) Fix various signed/unsigned issues to make a_strex.c
1664 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1666 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1667 length if passed a buffer. ASN1_INTEGER_to_BN failed
1668 if passed a NULL BN and its argument was negative.
1669 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1671 *) Modification to PKCS#7 encoding routines to output definite
1672 length encoding. Since currently the whole structures are in
1673 memory there's not real point in using indefinite length
1674 constructed encoding. However if OpenSSL is compiled with
1675 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1678 *) Added BIO_vprintf() and BIO_vsnprintf().
1681 *) Added more prefixes to parse for in the the strings written
1682 through a logging bio, to cover all the levels that are available
1683 through syslog. The prefixes are now:
1685 PANIC, EMERG, EMR => LOG_EMERG
1686 ALERT, ALR => LOG_ALERT
1687 CRIT, CRI => LOG_CRIT
1688 ERROR, ERR => LOG_ERR
1689 WARNING, WARN, WAR => LOG_WARNING
1690 NOTICE, NOTE, NOT => LOG_NOTICE
1691 INFO, INF => LOG_INFO
1692 DEBUG, DBG => LOG_DEBUG
1694 and as before, if none of those prefixes are present at the
1695 beginning of the string, LOG_ERR is chosen.
1697 On Win32, the LOG_* levels are mapped according to this:
1699 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1700 LOG_WARNING => EVENTLOG_WARNING_TYPE
1701 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1705 *) Made it possible to reconfigure with just the configuration
1706 argument "reconf" or "reconfigure". The command line arguments
1707 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1708 and are retrieved from there when reconfiguring.
1712 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1714 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1717 *) The obj_dat.pl script was messing up the sorting of object
1718 names. The reason was that it compared the quoted version
1719 of strings as a result "OCSP" > "OCSP Signing" because
1720 " > SPACE. Changed script to store unquoted versions of
1721 names and add quotes on output. It was also omitting some
1722 names from the lookup table if they were given a default
1723 value (that is if SN is missing it is given the same
1724 value as LN and vice versa), these are now added on the
1725 grounds that if an object has a name we should be able to
1726 look it up. Finally added warning output when duplicate
1727 short or long names are found.
1730 *) Changes needed for Tandem NSK.
1731 [Scott Uroff <scott@xypro.com>]
1733 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1734 RSA_padding_check_SSLv23(), special padding was never detected
1735 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1736 version rollback attacks was not effective.
1738 In s23_clnt.c, don't use special rollback-attack detection padding
1739 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1740 client; similarly, in s23_srvr.c, don't do the rollback check if
1741 SSL 2.0 is the only protocol enabled in the server.
1744 *) Make it possible to get hexdumps of unprintable data with 'openssl
1745 asn1parse'. By implication, the functions ASN1_parse_dump() and
1746 BIO_dump_indent() are added.
1749 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1750 these print out strings and name structures based on various
1751 flags including RFC2253 support and proper handling of
1752 multibyte characters. Added options to the 'x509' utility
1753 to allow the various flags to be set.
1756 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1757 Also change the functions X509_cmp_current_time() and
1758 X509_gmtime_adj() work with an ASN1_TIME structure,
1759 this will enable certificates using GeneralizedTime in validity
1760 dates to be checked.
1763 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1764 negative public key encodings) on by default,
1765 NO_NEG_PUBKEY_BUG can be set to disable it.
1768 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1769 content octets. An i2c_ASN1_OBJECT is unnecessary because
1770 the encoding can be trivially obtained from the structure.
1773 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1774 not read locks (CRYPTO_r_[un]lock).
1777 *) A first attempt at creating official support for shared
1778 libraries through configuration. I've kept it so the
1779 default is static libraries only, and the OpenSSL programs
1780 are always statically linked for now, but there are
1781 preparations for dynamic linking in place.
1782 This has been tested on Linux and True64.
1785 *) Randomness polling function for Win9x, as described in:
1786 Peter Gutmann, Software Generation of Practically Strong
1790 *) Fix so PRNG is seeded in req if using an already existing
1794 *) New options to smime application. -inform and -outform
1795 allow alternative formats for the S/MIME message including
1796 PEM and DER. The -content option allows the content to be
1797 specified separately. This should allow things like Netscape
1798 form signing output easier to verify.
1801 *) Fix the ASN1 encoding of tags using the 'long form'.
1804 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1805 STRING types. These convert content octets to and from the
1806 underlying type. The actual tag and length octets are
1807 already assumed to have been read in and checked. These
1808 are needed because all other string types have virtually
1809 identical handling apart from the tag. By having versions
1810 of the ASN1 functions that just operate on content octets
1811 IMPLICIT tagging can be handled properly. It also allows
1812 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1813 and ASN1_INTEGER are identical apart from the tag.
1816 *) Change the handling of OID objects as follows:
1818 - New object identifiers are inserted in objects.txt, following
1819 the syntax given in objects.README.
1820 - objects.pl is used to process obj_mac.num and create a new
1822 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1825 This is currently kind of a hack, and the perl code in objects.pl
1826 isn't very elegant, but it works as I intended. The simplest way
1827 to check that it worked correctly is to look in obj_dat.h and
1828 check the array nid_objs and make sure the objects haven't moved
1829 around (this is important!). Additions are OK, as well as
1830 consistent name changes.
1833 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1836 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1837 The given file adds to whatever has already been seeded into the
1838 random pool through the RANDFILE configuration file option or
1839 environment variable, or the default random state file.
1842 *) mkstack.pl now sorts each macro group into lexical order.
1843 Previously the output order depended on the order the files
1844 appeared in the directory, resulting in needless rewriting
1848 *) Patches to make OpenSSL compile under Win32 again. Mostly
1849 work arounds for the VC++ problem that it treats func() as
1850 func(void). Also stripped out the parts of mkdef.pl that
1851 added extra typesafe functions: these no longer exist.
1854 *) Reorganisation of the stack code. The macros are now all
1855 collected in safestack.h . Each macro is defined in terms of
1856 a "stack macro" of the form SKM_<name>(type, a, b). The
1857 DEBUG_SAFESTACK is now handled in terms of function casts,
1858 this has the advantage of retaining type safety without the
1859 use of additional functions. If DEBUG_SAFESTACK is not defined
1860 then the non typesafe macros are used instead. Also modified the
1861 mkstack.pl script to handle the new form. Needs testing to see
1862 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1863 the default if no major problems. Similar behaviour for ASN1_SET_OF
1864 and PKCS12_STACK_OF.
1867 *) When some versions of IIS use the 'NET' form of private key the
1868 key derivation algorithm is different. Normally MD5(password) is
1869 used as a 128 bit RC4 key. In the modified case
1870 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1871 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1872 as the old Netscape_RSA functions except they have an additional
1873 'sgckey' parameter which uses the modified algorithm. Also added
1874 an -sgckey command line option to the rsa utility. Thanks to
1875 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1876 algorithm to openssl-dev.
1879 *) The evp_local.h macros were using 'c.##kname' which resulted in
1880 invalid expansion on some systems (SCO 5.0.5 for example).
1881 Corrected to 'c.kname'.
1882 [Phillip Porch <root@theporch.com>]
1884 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1885 a STACK of email addresses from a certificate or request, these look
1886 in the subject name and the subject alternative name extensions and
1887 omit any duplicate addresses.
1890 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1891 This makes DSA verification about 2 % faster.
1894 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1895 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1896 plus overhead for 1024 bit moduli).
1897 This makes exponentiations about 0.5 % faster for 1024 bit
1898 exponents (as measured by "openssl speed rsa2048").
1901 *) Rename memory handling macros to avoid conflicts with other
1903 Malloc => OPENSSL_malloc
1904 Malloc_locked => OPENSSL_malloc_locked
1905 Realloc => OPENSSL_realloc
1906 Free => OPENSSL_free
1909 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1910 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1913 *) CygWin32 support.
1914 [John Jarvie <jjarvie@newsguy.com>]
1916 *) The type-safe stack code has been rejigged. It is now only compiled
1917 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1918 by default all type-specific stack functions are "#define"d back to
1919 standard stack functions. This results in more streamlined output
1920 but retains the type-safety checking possibilities of the original
1924 *) The STACK code has been cleaned up, and certain type declarations
1925 that didn't make a lot of sense have been brought in line. This has
1926 also involved a cleanup of sorts in safestack.h to more correctly
1927 map type-safe stack functions onto their plain stack counterparts.
1928 This work has also resulted in a variety of "const"ifications of
1929 lots of the code, especially "_cmp" operations which should normally
1930 be prototyped with "const" parameters anyway.
1933 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1934 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1935 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1936 where all of 'md' is used each time the PRNG is used, but 'state'
1937 is used only indexed by a cyclic counter. As entropy may not be
1938 well distributed from the beginning, 'md' is important as a
1939 chaining variable. However, the output function chains only half
1940 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1941 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1942 in all of 'state' being rewritten, with the new values depending
1943 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1946 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1947 the handshake is continued after ssl_verify_cert_chain();
1948 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1949 can lead to 'unexplainable' connection aborts later.
1950 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1952 *) Major EVP API cipher revision.
1953 Add hooks for extra EVP features. This allows various cipher
1954 parameters to be set in the EVP interface. Support added for variable
1955 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1956 setting of RC2 and RC5 parameters.
1958 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1961 Remove lots of duplicated code from the EVP library. For example *every*
1962 cipher init() function handles the 'iv' in the same way according to the
1963 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1964 for CFB and OFB modes they zero ctx->num.
1966 New functionality allows removal of S/MIME code RC2 hack.
1968 Most of the routines have the same form and so can be declared in terms
1971 By shifting this to the top level EVP_CipherInit() it can be removed from
1972 all individual ciphers. If the cipher wants to handle IVs or keys
1973 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1976 Change lots of functions like EVP_EncryptUpdate() to now return a
1977 value: although software versions of the algorithms cannot fail
1978 any installed hardware versions can.
1981 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1982 this option is set, tolerate broken clients that send the negotiated
1983 protocol version number instead of the requested protocol version
1987 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1988 i.e. non-zero for export ciphersuites, zero otherwise.
1989 Previous versions had this flag inverted, inconsistent with
1990 rsa_tmp_cb (..._TMP_RSA_CB).
1991 [Bodo Moeller; problem reported by Amit Chopra]
1993 *) Add missing DSA library text string. Work around for some IIS
1994 key files with invalid SEQUENCE encoding.
1997 *) Add a document (doc/standards.txt) that list all kinds of standards
1998 and so on that are implemented in OpenSSL.
2001 *) Enhance c_rehash script. Old version would mishandle certificates
2002 with the same subject name hash and wouldn't handle CRLs at all.
2003 Added -fingerprint option to crl utility, to support new c_rehash
2007 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2010 *) Fix for SSL server purpose checking. Server checking was
2011 rejecting certificates which had extended key usage present
2012 but no ssl client purpose.
2013 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2015 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2016 is a little unclear about how a blank password is handled.
2017 Since the password in encoded as a BMPString with terminating
2018 double NULL a zero length password would end up as just the
2019 double NULL. However no password at all is different and is
2020 handled differently in the PKCS#12 key generation code. NS
2021 treats a blank password as zero length. MSIE treats it as no
2022 password on export: but it will try both on import. We now do
2023 the same: PKCS12_parse() tries zero length and no password if
2024 the password is set to "" or NULL (NULL is now a valid password:
2025 it wasn't before) as does the pkcs12 application.
2028 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2029 perror when PEM_read_bio_X509_REQ fails, the error message must
2030 be obtained from the error queue.
2033 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2034 it in ERR_remove_state if appropriate, and change ERR_get_state
2035 accordingly to avoid race conditions (this is necessary because
2036 thread_hash is no longer constant once set).
2039 *) Bugfix for linux-elf makefile.one.
2042 *) RSA_get_default_method() will now cause a default
2043 RSA_METHOD to be chosen if one doesn't exist already.
2044 Previously this was only set during a call to RSA_new()
2045 or RSA_new_method(NULL) meaning it was possible for
2046 RSA_get_default_method() to return NULL.
2049 *) Added native name translation to the existing DSO code
2050 that will convert (if the flag to do so is set) filenames
2051 that are sufficiently small and have no path information
2052 into a canonical native form. Eg. "blah" converted to
2053 "libblah.so" or "blah.dll" etc.
2056 *) New function ERR_error_string_n(e, buf, len) which is like
2057 ERR_error_string(e, buf), but writes at most 'len' bytes
2058 including the 0 terminator. For ERR_error_string_n, 'buf'
2060 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2062 *) CONF library reworked to become more general. A new CONF
2063 configuration file reader "class" is implemented as well as a
2064 new functions (NCONF_*, for "New CONF") to handle it. The now
2065 old CONF_* functions are still there, but are reimplemented to
2066 work in terms of the new functions. Also, a set of functions
2067 to handle the internal storage of the configuration data is
2068 provided to make it easier to write new configuration file
2069 reader "classes" (I can definitely see something reading a
2070 configuration file in XML format, for example), called _CONF_*,
2071 or "the configuration storage API"...
2073 The new configuration file reading functions are:
2075 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2076 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2078 NCONF_default, NCONF_WIN32
2080 NCONF_dump_fp, NCONF_dump_bio
2082 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2083 NCONF_new creates a new CONF object. This works in the same way
2084 as other interfaces in OpenSSL, like the BIO interface.
2085 NCONF_dump_* dump the internal storage of the configuration file,
2086 which is useful for debugging. All other functions take the same
2087 arguments as the old CONF_* functions wth the exception of the
2088 first that must be a `CONF *' instead of a `LHASH *'.
2090 To make it easer to use the new classes with the old CONF_* functions,
2091 the function CONF_set_default_method is provided.
2094 *) Add '-tls1' option to 'openssl ciphers', which was already
2095 mentioned in the documentation but had not been implemented.
2096 (This option is not yet really useful because even the additional
2097 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2100 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2101 OpenSSL-based applications) load shared libraries and bind to
2102 them in a portable way.
2103 [Geoff Thorpe, with contributions from Richard Levitte]
2105 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2107 *) Make sure _lrotl and _lrotr are only used with MSVC.
2109 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2110 (the default implementation of RAND_status).
2112 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2113 to '-clrext' (= clear extensions), as intended and documented.
2114 [Bodo Moeller; inconsistency pointed out by Michael Attili
2117 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2118 was larger than the MD block size.
2119 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2121 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2122 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2123 using the passed key: if the passed key was a private key the result
2124 of X509_print(), for example, would be to print out all the private key
2128 *) des_quad_cksum() byte order bug fix.
2129 [Ulf Möller, using the problem description in krb4-0.9.7, where
2130 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2132 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2134 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2136 *) For easily testing in shell scripts whether some command
2137 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2138 returns with exit code 0 iff no command of the given name is available.
2139 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2140 the output goes to stdout and nothing is printed to stderr.
2141 Additional arguments are always ignored.
2143 Since for each cipher there is a command of the same name,
2144 the 'no-cipher' compilation switches can be tested this way.
2146 ('openssl no-XXX' is not able to detect pseudo-commands such
2147 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2150 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2153 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2154 is set; it will be thrown away anyway because each handshake creates
2156 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2157 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2158 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2159 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2162 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2163 'Q' and 'R' lose their special meanings (quit/renegotiate).
2164 This is part of what -quiet does; unlike -quiet, -ign_eof
2165 does not suppress any output.
2168 *) Add compatibility options to the purpose and trust code. The
2169 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2170 accepts a certificate or CA, this was the previous behaviour,
2171 with all the associated security issues.
2173 X509_TRUST_COMPAT is the old trust behaviour: only and
2174 automatically trust self signed roots in certificate store. A
2175 new trust setting X509_TRUST_DEFAULT is used to specify that
2176 a purpose has no associated trust setting and it should instead
2177 use the value in the default purpose.
2180 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2181 and fix a memory leak.
2184 *) In util/mkerr.pl (which implements 'make errors'), preserve
2185 reason strings from the previous version of the .c file, as
2186 the default to have only downcase letters (and digits) in
2187 automatically generated reasons codes is not always appropriate.
2190 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2191 using strerror. Previously, ERR_reason_error_string() returned
2192 library names as reason strings for SYSerr; but SYSerr is a special
2193 case where small numbers are errno values, not library numbers.
2196 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2197 converts DSA parameters into DH parameters. (When creating parameters,
2198 DSA_generate_parameters is used.)
2201 *) Include 'length' (recommended exponent length) in C code generated
2202 by 'openssl dhparam -C'.
2205 *) The second argument to set_label in perlasm was already being used
2206 so couldn't be used as a "file scope" flag. Moved to third argument
2210 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2211 instead of RAND_bytes for encryption IVs and salts.
2214 *) Include RAND_status() into RAND_METHOD instead of implementing
2215 it only for md_rand.c Otherwise replacing the PRNG by calling
2216 RAND_set_rand_method would be impossible.
2219 *) Don't let DSA_generate_key() enter an infinite loop if the random
2220 number generation fails.
2223 *) New 'rand' application for creating pseudo-random output.
2226 *) Added configuration support for Linux/IA64
2227 [Rolf Haberrecker <rolf@suse.de>]
2229 *) Assembler module support for Mingw32.
2232 *) Shared library support for HPUX (in shlib/).
2233 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2235 *) Shared library support for Solaris gcc.
2236 [Lutz Behnke <behnke@trustcenter.de>]
2238 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2240 *) PKCS7_encrypt() was adding text MIME headers twice because they
2241 were added manually and by SMIME_crlf_copy().
2244 *) In bntest.c don't call BN_rand with zero bits argument.
2245 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2247 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2248 case was implemented. This caused BN_div_recp() to fail occasionally.
2251 *) Add an optional second argument to the set_label() in the perl
2252 assembly language builder. If this argument exists and is set
2253 to 1 it signals that the assembler should use a symbol whose
2254 scope is the entire file, not just the current function. This
2255 is needed with MASM which uses the format label:: for this scope.
2256 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2258 *) Change the ASN1 types so they are typedefs by default. Before
2259 almost all types were #define'd to ASN1_STRING which was causing
2260 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2264 *) Change names of new functions to the new get1/get0 naming
2265 convention: After 'get1', the caller owns a reference count
2266 and has to call ..._free; 'get0' returns a pointer to some
2267 data structure without incrementing reference counters.
2268 (Some of the existing 'get' functions increment a reference
2269 counter, some don't.)
2270 Similarly, 'set1' and 'add1' functions increase reference
2271 counters or duplicate objects.
2274 *) Allow for the possibility of temp RSA key generation failure:
2275 the code used to assume it always worked and crashed on failure.
2278 *) Fix potential buffer overrun problem in BIO_printf().
2279 [Ulf Möller, using public domain code by Patrick Powell; problem
2280 pointed out by David Sacerdote <das33@cornell.edu>]
2282 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2283 RAND_egd() and RAND_status(). In the command line application,
2284 the EGD socket can be specified like a seed file using RANDFILE
2288 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2289 Some CAs (e.g. Verisign) distribute certificates in this form.
2292 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2293 list to exclude them. This means that no special compilation option
2294 is needed to use anonymous DH: it just needs to be included in the
2298 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2299 EVP_MD_type. The old functionality is available in a new macro called
2300 EVP_MD_md(). Change code that uses it and update docs.
2303 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2304 where the 'void *' argument is replaced by a function pointer argument.
2305 Previously 'void *' was abused to point to functions, which works on
2306 many platforms, but is not correct. As these functions are usually
2307 called by macros defined in OpenSSL header files, most source code
2308 should work without changes.
2311 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2312 sections with information on -D... compiler switches used for
2313 compiling the library so that applications can see them. To enable
2314 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2315 must be defined. E.g.,
2316 #define OPENSSL_ALGORITHM_DEFINES
2317 #include <openssl/opensslconf.h>
2318 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2319 [Richard Levitte, Ulf and Bodo Möller]
2321 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2325 *) Change the 'other' type in certificate aux info to a STACK_OF
2326 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2327 the required ASN1 format: arbitrary types determined by an OID.
2330 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2331 argument to 'req'. This is not because the function is newer or
2332 better than others it just uses the work 'NEW' in the certificate
2333 request header lines. Some software needs this.
2336 *) Reorganise password command line arguments: now passwords can be
2337 obtained from various sources. Delete the PEM_cb function and make
2338 it the default behaviour: i.e. if the callback is NULL and the
2339 usrdata argument is not NULL interpret it as a null terminated pass
2340 phrase. If usrdata and the callback are NULL then the pass phrase
2341 is prompted for as usual.
2344 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2345 the support is automatically enabled. The resulting binaries will
2346 autodetect the card and use it if present.
2347 [Ben Laurie and Compaq Inc.]
2349 *) Work around for Netscape hang bug. This sends certificate request
2350 and server done in one record. Since this is perfectly legal in the
2351 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2352 the bugs/SSLv3 entry for more info.
2355 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2358 *) Add -rand argument to smime and pkcs12 applications and read/write
2362 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2365 *) Add command line password options to the remaining applications.
2368 *) Bug fix for BN_div_recp() for numerators with an even number of
2372 *) More tests in bntest.c, and changed test_bn output.
2375 *) ./config recognizes MacOS X now.
2378 *) Bug fix for BN_div() when the first words of num and divsor are
2379 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2382 *) Add support for various broken PKCS#8 formats, and command line
2383 options to produce them.
2386 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2387 get temporary BIGNUMs from a BN_CTX.
2390 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2394 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2395 include a #define from the old name to the new. The original intent
2396 was that statically linked binaries could for example just call
2397 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2398 link with digests. This never worked becayse SSLeay_add_all_digests()
2399 and SSLeay_add_all_ciphers() were in the same source file so calling
2400 one would link with the other. They are now in separate source files.
2403 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2406 *) Use a less unusual form of the Miller-Rabin primality test (it used
2407 a binary algorithm for exponentiation integrated into the Miller-Rabin
2408 loop, our standard modexp algorithms are faster).
2411 *) Support for the EBCDIC character set completed.
2412 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2414 *) Source code cleanups: use const where appropriate, eliminate casts,
2415 use void * instead of char * in lhash.
2418 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2419 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2420 this the server could overwrite ephemeral keys that the client
2424 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2425 using 50 iterations of the Rabin-Miller test.
2427 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2428 iterations of the Rabin-Miller test as required by the appendix
2429 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2430 As BN_is_prime_fasttest includes trial division, DSA parameter
2431 generation becomes much faster.
2433 This implies a change for the callback functions in DSA_is_prime
2434 and DSA_generate_parameters: The callback function is called once
2435 for each positive witness in the Rabin-Miller test, not just
2436 occasionally in the inner loop; and the parameters to the
2437 callback function now provide an iteration count for the outer
2438 loop rather than for the current invocation of the inner loop.
2439 DSA_generate_parameters additionally can call the callback
2440 function with an 'iteration count' of -1, meaning that a
2441 candidate has passed the trial division test (when q is generated
2442 from an application-provided seed, trial division is skipped).
2445 *) New function BN_is_prime_fasttest that optionally does trial
2446 division before starting the Rabin-Miller test and has
2447 an additional BN_CTX * argument (whereas BN_is_prime always
2448 has to allocate at least one BN_CTX).
2449 'callback(1, -1, cb_arg)' is called when a number has passed the
2450 trial division stage.
2453 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2457 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2460 *) New function BN_pseudo_rand().
2463 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2464 bignum version of BN_from_montgomery() with the working code from
2465 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2469 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2470 made it impossible to use the same SSL_SESSION data structure in
2471 SSL2 clients in multiple threads.
2474 *) The return value of RAND_load_file() no longer counts bytes obtained
2475 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2476 to seed the PRNG (previously an explicit byte count was required).
2477 [Ulf Möller, Bodo Möller]
2479 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2480 used (char *) instead of (void *) and had casts all over the place.
2483 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2486 *) Retain source code compatibility for BN_prime_checks macro:
2487 BN_is_prime(..., BN_prime_checks, ...) now uses
2488 BN_prime_checks_for_size to determine the appropriate number of
2489 Rabin-Miller iterations.
2492 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2493 DH_CHECK_P_NOT_SAFE_PRIME.
2494 (Check if this is true? OpenPGP calls them "strong".)
2497 *) Merge the functionality of "dh" and "gendh" programs into a new program
2498 "dhparam". The old programs are retained for now but will handle DH keys
2499 (instead of parameters) in future.
2502 *) Make the ciphers, s_server and s_client programs check the return values
2503 when a new cipher list is set.
2506 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2507 ciphers. Before when the 56bit ciphers were enabled the sorting was
2510 The syntax for the cipher sorting has been extended to support sorting by
2511 cipher-strength (using the strength_bits hard coded in the tables).
2512 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2514 Fix a bug in the cipher-command parser: when supplying a cipher command
2515 string with an "undefined" symbol (neither command nor alphanumeric
2516 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2517 an error is flagged.
2519 Due to the strength-sorting extension, the code of the
2520 ssl_create_cipher_list() function was completely rearranged. I hope that
2521 the readability was also increased :-)
2522 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2524 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2525 for the first serial number and places 2 in the serial number file. This
2526 avoids problems when the root CA is created with serial number zero and
2527 the first user certificate has the same issuer name and serial number
2531 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2532 the new code. Add documentation for this stuff.
2535 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2536 X509_*() to X509at_*() on the grounds that they don't handle X509
2537 structures and behave in an analagous way to the X509v3 functions:
2538 they shouldn't be called directly but wrapper functions should be used
2541 So we also now have some wrapper functions that call the X509at functions
2542 when passed certificate requests. (TO DO: similar things can be done with
2543 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2544 things. Some of these need some d2i or i2d and print functionality
2545 because they handle more complex structures.)
2548 *) Add missing #ifndefs that caused missing symbols when building libssl
2549 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2550 NO_RSA in ssl/s2*.c.
2551 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2553 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2554 has a return value which indicates the quality of the random data
2555 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2556 error queue. New function RAND_pseudo_bytes() generates output that is
2557 guaranteed to be unique but not unpredictable. RAND_add is like
2558 RAND_seed, but takes an extra argument for an entropy estimate
2559 (RAND_seed always assumes full entropy).
2562 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2563 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2564 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2565 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2566 false-positive rate of at most 2^-80 for random input.
2569 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2572 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2573 in the 0.9.5 release), this returns the chain
2574 from an X509_CTX structure with a dup of the stack and all
2575 the X509 reference counts upped: so the stack will exist
2576 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2579 Also make SSL_SESSION_print() print out the verify return
2583 *) Add manpage for the pkcs12 command. Also change the default
2584 behaviour so MAC iteration counts are used unless the new
2585 -nomaciter option is used. This improves file security and
2586 only older versions of MSIE (4.0 for example) need it.
2589 *) Honor the no-xxx Configure options when creating .DEF files.
2592 *) Add PKCS#10 attributes to field table: challengePassword,
2593 unstructuredName and unstructuredAddress. These are taken from
2594 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2595 international characters are used.
2597 More changes to X509_ATTRIBUTE code: allow the setting of types
2598 based on strings. Remove the 'loc' parameter when adding
2599 attributes because these will be a SET OF encoding which is sorted
2603 *) Initial changes to the 'req' utility to allow request generation
2604 automation. This will allow an application to just generate a template
2605 file containing all the field values and have req construct the
2608 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2609 used all over the place including certificate requests and PKCS#7
2610 structures. They are currently handled manually where necessary with
2611 some primitive wrappers for PKCS#7. The new functions behave in a
2612 manner analogous to the X509 extension functions: they allow
2613 attributes to be looked up by NID and added.
2615 Later something similar to the X509V3 code would be desirable to
2616 automatically handle the encoding, decoding and printing of the
2617 more complex types. The string types like challengePassword can
2618 be handled by the string table functions.
2620 Also modified the multi byte string table handling. Now there is
2621 a 'global mask' which masks out certain types. The table itself
2622 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2623 is useful when for example there is only one permissible type
2624 (as in countryName) and using the mask might result in no valid
2628 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2629 SSL_get_peer_finished to allow applications to obtain the latest
2630 Finished messages sent to the peer or expected from the peer,
2631 respectively. (SSL_get_peer_finished is usually the Finished message
2632 actually received from the peer, otherwise the protocol will be aborted.)
2634 As the Finished message are message digests of the complete handshake
2635 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2636 be used for external authentication procedures when the authentication
2637 provided by SSL/TLS is not desired or is not enough.
2640 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2641 the host supports BWX extension and if Compaq C is present on the
2642 $PATH. Just exploiting of the BWX extension results in 20-30%
2643 performance kick for some algorithms, e.g. DES and RC4 to mention
2644 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2648 *) Add support for MS "fast SGC". This is arguably a violation of the
2649 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2650 weak crypto and after checking the certificate is SGC a second one
2651 with strong crypto. MS SGC stops the first handshake after receiving
2652 the server certificate message and sends a second client hello. Since
2653 a server will typically do all the time consuming operations before
2654 expecting any further messages from the client (server key exchange
2655 is the most expensive) there is little difference between the two.
2657 To get OpenSSL to support MS SGC we have to permit a second client
2658 hello message after we have sent server done. In addition we have to
2659 reset the MAC if we do get this second client hello.
2662 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2663 if a DER encoded private key is RSA or DSA traditional format. Changed
2664 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2665 format DER encoded private key. Newer code should use PKCS#8 format which
2666 has the key type encoded in the ASN1 structure. Added DER private key
2667 support to pkcs8 application.
2670 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2671 ciphersuites has been selected (as required by the SSL 3/TLS 1
2672 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2673 is set, we interpret this as a request to violate the specification
2674 (the worst that can happen is a handshake failure, and 'correct'
2675 behaviour would result in a handshake failure anyway).
2678 *) In SSL_CTX_add_session, take into account that there might be multiple
2679 SSL_SESSION structures with the same session ID (e.g. when two threads
2680 concurrently obtain them from an external cache).
2681 The internal cache can handle only one SSL_SESSION with a given ID,
2682 so if there's a conflict, we now throw out the old one to achieve
2686 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2687 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2688 some routines that use cipher OIDs: some ciphers do not have OIDs
2689 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2693 *) Simplify the trust setting structure and code. Now we just have
2694 two sequences of OIDs for trusted and rejected settings. These will
2695 typically have values the same as the extended key usage extension
2696 and any application specific purposes.
2698 The trust checking code now has a default behaviour: it will just
2699 check for an object with the same NID as the passed id. Functions can
2700 be provided to override either the default behaviour or the behaviour
2701 for a given id. SSL client, server and email already have functions
2702 in place for compatibility: they check the NID and also return "trusted"
2703 if the certificate is self signed.
2706 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2707 traditional format into an EVP_PKEY structure.
2710 *) Add a password callback function PEM_cb() which either prompts for
2711 a password if usr_data is NULL or otherwise assumes it is a null
2712 terminated password. Allow passwords to be passed on command line
2713 environment or config files in a few more utilities.
2716 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2717 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2718 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2719 Update documentation.
2722 *) Support for ASN1 "NULL" type. This could be handled before by using
2723 ASN1_TYPE but there wasn't any function that would try to read a NULL
2724 and produce an error if it couldn't. For compatibility we also have
2725 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2726 don't allocate anything because they don't need to.
2729 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2731 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2733 *) Rebuild of the memory allocation routines used by OpenSSL code and
2734 possibly others as well. The purpose is to make an interface that
2735 provide hooks so anyone can build a separate set of allocation and
2736 deallocation routines to be used by OpenSSL, for example memory
2737 pool implementations, or something else, which was previously hard
2738 since Malloc(), Realloc() and Free() were defined as macros having
2739 the values malloc, realloc and free, respectively (except for Win32
2740 compilations). The same is provided for memory debugging code.
2741 OpenSSL already comes with functionality to find memory leaks, but
2742 this gives people a chance to debug other memory problems.
2744 With these changes, a new set of functions and macros have appeared:
2746 CRYPTO_set_mem_debug_functions() [F]
2747 CRYPTO_get_mem_debug_functions() [F]
2748 CRYPTO_dbg_set_options() [F]
2749 CRYPTO_dbg_get_options() [F]
2750 CRYPTO_malloc_debug_init() [M]
2752 The memory debug functions are NULL by default, unless the library
2753 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2754 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2755 gives the standard debugging functions that come with OpenSSL) or
2756 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2757 provided by the library user) must be used. When the standard
2758 debugging functions are used, CRYPTO_dbg_set_options can be used to
2759 request additional information:
2760 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2761 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2763 Also, things like CRYPTO_set_mem_functions will always give the
2764 expected result (the new set of functions is used for allocation
2765 and deallocation) at all times, regardless of platform and compiler
2768 To finish it up, some functions that were never use in any other
2769 way than through macros have a new API and new semantic:
2772 CRYPTO_dbg_realloc()
2775 All macros of value have retained their old syntax.
2776 [Richard Levitte and Bodo Moeller]
2778 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2779 ordering of SMIMECapabilities wasn't in "strength order" and there
2780 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2784 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2785 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2786 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2788 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2789 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2790 functionality to handle multipart/signed properly) and a utility
2791 called 'smime' to call all this stuff. This is based on code I
2792 originally wrote for Celo who have kindly allowed it to be
2793 included in OpenSSL.
2796 *) Add variants des_set_key_checked and des_set_key_unchecked of
2797 des_set_key (aka des_key_sched). Global variable des_check_key
2798 decides which of these is called by des_set_key; this way
2799 des_check_key behaves as it always did, but applications and
2800 the library itself, which was buggy for des_check_key == 1,
2801 have a cleaner way to pick the version they need.
2804 *) New function PKCS12_newpass() which changes the password of a
2808 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2809 dynamic mix. In both cases the ids can be used as an index into the
2810 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2811 functions so they accept a list of the field values and the
2812 application doesn't need to directly manipulate the X509_TRUST
2816 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2820 *) Modify the way the V3 extension code looks up extensions. This now
2821 works in a similar way to the object code: we have some "standard"
2822 extensions in a static table which is searched with OBJ_bsearch()
2823 and the application can add dynamic ones if needed. The file
2824 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2825 updated whenever a new extension is added to the core code and kept
2826 in ext_nid order. There is a simple program 'tabtest.c' which checks
2827 this. New extensions are not added too often so this file can readily
2828 be maintained manually.
2830 There are two big advantages in doing things this way. The extensions
2831 can be looked up immediately and no longer need to be "added" using
2832 X509V3_add_standard_extensions(): this function now does nothing.
2833 [Side note: I get *lots* of email saying the extension code doesn't
2834 work because people forget to call this function]
2835 Also no dynamic allocation is done unless new extensions are added:
2836 so if we don't add custom extensions there is no need to call
2837 X509V3_EXT_cleanup().
2840 *) Modify enc utility's salting as follows: make salting the default. Add a
2841 magic header, so unsalted files fail gracefully instead of just decrypting
2842 to garbage. This is because not salting is a big security hole, so people
2843 should be discouraged from doing it.
2846 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2847 digest to be passed on the command line but it only used this
2848 parameter when signing a certificate. Modified so all relevant
2849 operations are affected by the digest parameter including the
2850 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2851 DSA key was used because it didn't fix the digest.
2854 *) Initial certificate chain verify code. Currently tests the untrusted
2855 certificates for consistency with the verify purpose (which is set
2856 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2858 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2859 this is because it will reject chains with invalid extensions whereas
2860 every previous version of OpenSSL and SSLeay made no checks at all.
2862 Trust code: checks the root CA for the relevant trust settings. Trust
2863 settings have an initial value consistent with the verify purpose: e.g.
2864 if the verify purpose is for SSL client use it expects the CA to be
2865 trusted for SSL client use. However the default value can be changed to
2866 permit custom trust settings: one example of this would be to only trust
2867 certificates from a specific "secure" set of CAs.
2869 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2870 which should be used for version portability: especially since the
2871 verify structure is likely to change more often now.
2873 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2874 to set them. If not set then assume SSL clients will verify SSL servers
2877 Two new options to the verify program: -untrusted allows a set of
2878 untrusted certificates to be passed in and -purpose which sets the
2879 intended purpose of the certificate. If a purpose is set then the
2880 new chain verify code is used to check extension consistency.
2883 *) Support for the authority information access extension.
2886 *) Modify RSA and DSA PEM read routines to transparently handle
2887 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2888 public keys in a format compatible with certificate
2889 SubjectPublicKeyInfo structures. Unfortunately there were already
2890 functions called *_PublicKey_* which used various odd formats so
2891 these are retained for compatibility: however the DSA variants were
2892 never in a public release so they have been deleted. Changed dsa/rsa
2893 utilities to handle the new format: note no releases ever handled public
2894 keys so we should be OK.
2896 The primary motivation for this change is to avoid the same fiasco
2897 that dogs private keys: there are several incompatible private key
2898 formats some of which are standard and some OpenSSL specific and
2899 require various evil hacks to allow partial transparent handling and
2900 even then it doesn't work with DER formats. Given the option anything
2901 other than PKCS#8 should be dumped: but the other formats have to
2902 stay in the name of compatibility.
2904 With public keys and the benefit of hindsight one standard format
2905 is used which works with EVP_PKEY, RSA or DSA structures: though
2906 it clearly returns an error if you try to read the wrong kind of key.
2908 Added a -pubkey option to the 'x509' utility to output the public key.
2909 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2910 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2911 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2912 that do the same as the EVP_PKEY_assign_*() except they up the
2913 reference count of the added key (they don't "swallow" the
2917 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2918 CRLs would fail if the file contained no certificates or no CRLs:
2919 added a new function to read in both types and return the number
2920 read: this means that if none are read it will be an error. The
2921 DER versions of the certificate and CRL reader would always fail
2922 because it isn't possible to mix certificates and CRLs in DER format
2923 without choking one or the other routine. Changed this to just read
2924 a certificate: this is the best we can do. Also modified the code
2925 in apps/verify.c to take notice of return codes: it was previously
2926 attempting to read in certificates from NULL pointers and ignoring
2927 any errors: this is one reason why the cert and CRL reader seemed
2928 to work. It doesn't check return codes from the default certificate
2929 routines: these may well fail if the certificates aren't installed.
2932 *) Code to support otherName option in GeneralName.
2935 *) First update to verify code. Change the verify utility
2936 so it warns if it is passed a self signed certificate:
2937 for consistency with the normal behaviour. X509_verify
2938 has been modified to it will now verify a self signed
2939 certificate if *exactly* the same certificate appears
2940 in the store: it was previously impossible to trust a
2941 single self signed certificate. This means that:
2942 openssl verify ss.pem
2943 now gives a warning about a self signed certificate but
2944 openssl verify -CAfile ss.pem ss.pem
2948 *) For servers, store verify_result in SSL_SESSION data structure
2949 (and add it to external session representation).
2950 This is needed when client certificate verifications fails,
2951 but an application-provided verification callback (set by
2952 SSL_CTX_set_cert_verify_callback) allows accepting the session
2953 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2954 but returns 1): When the session is reused, we have to set
2955 ssl->verify_result to the appropriate error code to avoid
2957 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2959 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2960 case in PKCS7_dataInit() where the signed PKCS7 structure
2961 didn't contain any existing data because it was being created.
2962 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2964 *) Add a salt to the key derivation routines in enc.c. This
2965 forms the first 8 bytes of the encrypted file. Also add a
2966 -S option to allow a salt to be input on the command line.
2969 *) New function X509_cmp(). Oddly enough there wasn't a function
2970 to compare two certificates. We do this by working out the SHA1
2971 hash and comparing that. X509_cmp() will be needed by the trust
2975 *) SSL_get1_session() is like SSL_get_session(), but increments
2976 the reference count in the SSL_SESSION returned.
2977 [Geoff Thorpe <geoff@eu.c2.net>]
2979 *) Fix for 'req': it was adding a null to request attributes.
2980 Also change the X509_LOOKUP and X509_INFO code to handle
2981 certificate auxiliary information.
2984 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2988 *) Add the possibility to add extra information to the memory leak
2989 detecting output, to form tracebacks, showing from where each
2990 allocation was originated: CRYPTO_push_info("constant string") adds
2991 the string plus current file name and line number to a per-thread
2992 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2993 is like calling CYRPTO_pop_info() until the stack is empty.
2994 Also updated memory leak detection code to be multi-thread-safe.
2997 *) Add options -text and -noout to pkcs7 utility and delete the
2998 encryption options which never did anything. Update docs.
3001 *) Add options to some of the utilities to allow the pass phrase
3002 to be included on either the command line (not recommended on
3003 OSes like Unix) or read from the environment. Update the
3004 manpages and fix a few bugs.
3007 *) Add a few manpages for some of the openssl commands.
3010 *) Fix the -revoke option in ca. It was freeing up memory twice,
3011 leaking and not finding already revoked certificates.
3014 *) Extensive changes to support certificate auxiliary information.
3015 This involves the use of X509_CERT_AUX structure and X509_AUX
3016 functions. An X509_AUX function such as PEM_read_X509_AUX()
3017 can still read in a certificate file in the usual way but it
3018 will also read in any additional "auxiliary information". By
3019 doing things this way a fair degree of compatibility can be
3020 retained: existing certificates can have this information added
3021 using the new 'x509' options.
3023 Current auxiliary information includes an "alias" and some trust
3024 settings. The trust settings will ultimately be used in enhanced
3025 certificate chain verification routines: currently a certificate
3026 can only be trusted if it is self signed and then it is trusted
3030 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3031 The problem was that one of the replacement routines had not been working
3032 since SSLeay releases. For now the offending routine has been replaced
3033 with non-optimised assembler. Even so, this now gives around 95%
3034 performance improvement for 1024 bit RSA signs.
3037 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3038 handling. Most clients have the effective key size in bits equal to
3039 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3040 A few however don't do this and instead use the size of the decrypted key
3041 to determine the RC2 key length and the AlgorithmIdentifier to determine
3042 the effective key length. In this case the effective key length can still
3043 be 40 bits but the key length can be 168 bits for example. This is fixed
3044 by manually forcing an RC2 key into the EVP_PKEY structure because the
3045 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3046 the key length and effective key length are equal.
3049 *) Add a bunch of functions that should simplify the creation of
3050 X509_NAME structures. Now you should be able to do:
3051 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3052 and have it automatically work out the correct field type and fill in
3053 the structures. The more adventurous can try:
3054 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3055 and it will (hopefully) work out the correct multibyte encoding.
3058 *) Change the 'req' utility to use the new field handling and multibyte
3059 copy routines. Before the DN field creation was handled in an ad hoc
3060 way in req, ca, and x509 which was rather broken and didn't support
3061 BMPStrings or UTF8Strings. Since some software doesn't implement
3062 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3063 using the dirstring_type option. See the new comment in the default
3064 openssl.cnf for more info.
3067 *) Make crypto/rand/md_rand.c more robust:
3068 - Assure unique random numbers after fork().
3069 - Make sure that concurrent threads access the global counter and
3070 md serializably so that we never lose entropy in them
3071 or use exactly the same state in multiple threads.
3072 Access to the large state is not always serializable because
3073 the additional locking could be a performance killer, and
3074 md should be large enough anyway.
3077 *) New file apps/app_rand.c with commonly needed functionality
3078 for handling the random seed file.
3080 Use the random seed file in some applications that previously did not:
3082 dsaparam -genkey (which also ignored its '-rand' option),
3085 x509 (when signing).
3086 Except on systems with /dev/urandom, it is crucial to have a random
3087 seed file at least for key creation, DSA signing, and for DH exchanges;
3088 for RSA signatures we could do without one.
3090 gendh and gendsa (unlike genrsa) used to read only the first byte
3091 of each file listed in the '-rand' option. The function as previously
3092 found in genrsa is now in app_rand.c and is used by all programs
3093 that support '-rand'.
3096 *) In RAND_write_file, use mode 0600 for creating files;
3097 don't just chmod when it may be too late.
3100 *) Report an error from X509_STORE_load_locations
3101 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3104 *) New function ASN1_mbstring_copy() this copies a string in either
3105 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3106 into an ASN1_STRING type. A mask of permissible types is passed
3107 and it chooses the "minimal" type to use or an error if not type
3111 *) Add function equivalents to the various macros in asn1.h. The old
3112 macros are retained with an M_ prefix. Code inside the library can
3113 use the M_ macros. External code (including the openssl utility)
3114 should *NOT* in order to be "shared library friendly".
3117 *) Add various functions that can check a certificate's extensions
3118 to see if it usable for various purposes such as SSL client,
3119 server or S/MIME and CAs of these types. This is currently
3120 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3121 verification. Also added a -purpose flag to x509 utility to
3122 print out all the purposes.
3125 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3129 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3130 for, obtain and decode and extension and obtain its critical flag.
3131 This allows all the necessary extension code to be handled in a
3132 single function call.
3135 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3136 platforms. See crypto/rc4/rc4_enc.c for further details.
3139 *) New -noout option to asn1parse. This causes no output to be produced
3140 its main use is when combined with -strparse and -out to extract data
3141 from a file (which may not be in ASN.1 format).
3144 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3145 when producing the local key id.
3146 [Richard Levitte <levitte@stacken.kth.se>]
3148 *) New option -dhparam in s_server. This allows a DH parameter file to be
3149 stated explicitly. If it is not stated then it tries the first server
3150 certificate file. The previous behaviour hard coded the filename
3154 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3155 a public key to be input or output. For example:
3156 openssl rsa -in key.pem -pubout -out pubkey.pem
3157 Also added necessary DSA public key functions to handle this.
3160 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3161 in the message. This was handled by allowing
3162 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3163 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3165 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3166 to the end of the strings whereas this didn't. This would cause problems
3167 if strings read with d2i_ASN1_bytes() were later modified.
3168 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3170 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3171 data and it contains EOF it will end up returning an error. This is
3172 caused by input 46 bytes long. The cause is due to the way base64
3173 BIOs find the start of base64 encoded data. They do this by trying a
3174 trial decode on each line until they find one that works. When they
3175 do a flag is set and it starts again knowing it can pass all the
3176 data directly through the decoder. Unfortunately it doesn't reset
3177 the context it uses. This means that if EOF is reached an attempt
3178 is made to pass two EOFs through the context and this causes the
3179 resulting error. This can also cause other problems as well. As is
3180 usual with these problems it takes *ages* to find and the fix is
3181 trivial: move one line.
3182 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3184 *) Ugly workaround to get s_client and s_server working under Windows. The
3185 old code wouldn't work because it needed to select() on sockets and the
3186 tty (for keypresses and to see if data could be written). Win32 only
3187 supports select() on sockets so we select() with a 1s timeout on the
3188 sockets and then see if any characters are waiting to be read, if none
3189 are present then we retry, we also assume we can always write data to
3190 the tty. This isn't nice because the code then blocks until we've
3191 received a complete line of data and it is effectively polling the
3192 keyboard at 1s intervals: however it's quite a bit better than not
3193 working at all :-) A dedicated Windows application might handle this
3194 with an event loop for example.
3197 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3198 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3199 will be called when RSA_sign() and RSA_verify() are used. This is useful
3200 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3201 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3202 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3203 This necessitated the support of an extra signature type NID_md5_sha1
3204 for SSL signatures and modifications to the SSL library to use it instead
3205 of calling RSA_public_decrypt() and RSA_private_encrypt().
3208 *) Add new -verify -CAfile and -CApath options to the crl program, these
3209 will lookup a CRL issuers certificate and verify the signature in a
3210 similar way to the verify program. Tidy up the crl program so it
3211 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3212 less strict. It will now permit CRL extensions even if it is not
3213 a V2 CRL: this will allow it to tolerate some broken CRLs.
3216 *) Initialize all non-automatic variables each time one of the openssl
3217 sub-programs is started (this is necessary as they may be started
3218 multiple times from the "OpenSSL>" prompt).
3219 [Lennart Bang, Bodo Moeller]
3221 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3222 removing all other RSA functionality (this is what NO_RSA does). This
3223 is so (for example) those in the US can disable those operations covered
3224 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3228 *) Non-copying interface to BIO pairs.
3229 (still largely untested)
3232 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3233 ASCII string. This was handled independently in various places before.
3236 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3237 UTF8 strings a character at a time.
3240 *) Use client_version from client hello to select the protocol
3241 (s23_srvr.c) and for RSA client key exchange verification
3242 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3245 *) Add various utility functions to handle SPKACs, these were previously
3246 handled by poking round in the structure internals. Added new function
3247 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3248 print, verify and generate SPKACs. Based on an original idea from
3249 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3252 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3255 *) Allow the config file extension section to be overwritten on the
3256 command line. Based on an original idea from Massimiliano Pala
3257 <madwolf@comune.modena.it>. The new option is called -extensions
3258 and can be applied to ca, req and x509. Also -reqexts to override
3259 the request extensions in req and -crlexts to override the crl extensions
3263 *) Add new feature to the SPKAC handling in ca. Now you can include
3264 the same field multiple times by preceding it by "XXXX." for example:
3267 this is the same syntax as used in the req config file.
3270 *) Allow certificate extensions to be added to certificate requests. These
3271 are specified in a 'req_extensions' option of the req section of the
3272 config file. They can be printed out with the -text option to req but
3273 are otherwise ignored at present.
3276 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3277 data read consists of only the final block it would not decrypted because
3278 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3279 A misplaced 'break' also meant the decrypted final block might not be
3280 copied until the next read.
3283 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3284 a few extra parameters to the DH structure: these will be useful if
3285 for example we want the value of 'q' or implement X9.42 DH.
3288 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3289 provides hooks that allow the default DSA functions or functions on a
3290 "per key" basis to be replaced. This allows hardware acceleration and
3291 hardware key storage to be handled without major modification to the
3292 library. Also added low level modexp hooks and CRYPTO_EX structure and
3293 associated functions.
3296 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3297 as "read only": it can't be written to and the buffer it points to will
3298 not be freed. Reading from a read only BIO is much more efficient than
3299 a normal memory BIO. This was added because there are several times when
3300 an area of memory needs to be read from a BIO. The previous method was
3301 to create a memory BIO and write the data to it, this results in two
3302 copies of the data and an O(n^2) reading algorithm. There is a new
3303 function BIO_new_mem_buf() which creates a read only memory BIO from
3304 an area of memory. Also modified the PKCS#7 routines to use read only
3308 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3309 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3310 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3311 but a retry condition occured while trying to read the rest.
3314 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3315 NID_pkcs7_encrypted by default: this was wrong since this should almost
3316 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3317 the encrypted data type: this is a more sensible place to put it and it
3318 allows the PKCS#12 code to be tidied up that duplicated this
3322 *) Changed obj_dat.pl script so it takes its input and output files on
3323 the command line. This should avoid shell escape redirection problems
3327 *) Initial support for certificate extension requests, these are included
3328 in things like Xenroll certificate requests. Included functions to allow
3329 extensions to be obtained and added.
3332 *) -crlf option to s_client and s_server for sending newlines as
3333 CRLF (as required by many protocols).
3336 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3338 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3339 [Ralf S. Engelschall]
3341 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3342 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3344 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3348 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3349 DH parameters/keys (q is lost during that conversion, but the resulting
3350 DH parameters contain its length).
3352 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3353 much faster than DH_generate_parameters (which creates parameters
3354 where p = 2*q + 1), and also the smaller q makes DH computations
3355 much more efficient (160-bit exponentiation instead of 1024-bit
3356 exponentiation); so this provides a convenient way to support DHE
3357 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3358 utter importance to use
3359 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3361 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3362 when such DH parameters are used, because otherwise small subgroup
3363 attacks may become possible!
3366 *) Avoid memory leak in i2d_DHparams.
3369 *) Allow the -k option to be used more than once in the enc program:
3370 this allows the same encrypted message to be read by multiple recipients.
3373 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3374 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3375 it will always use the numerical form of the OID, even if it has a short
3379 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3380 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3381 otherwise bn_mod_exp was called. In the case of hardware keys for example
3382 no private key components need be present and it might store extra data
3383 in the RSA structure, which cannot be accessed from bn_mod_exp.
3384 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3385 private key operations.
3388 *) Added support for SPARC Linux.
3391 *) pem_password_cb function type incompatibly changed from
3392 typedef int pem_password_cb(char *buf, int size, int rwflag);
3394 ....(char *buf, int size, int rwflag, void *userdata);
3395 so that applications can pass data to their callbacks:
3396 The PEM[_ASN1]_{read,write}... functions and macros now take an
3397 additional void * argument, which is just handed through whenever
3398 the password callback is called.
3399 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3401 New function SSL_CTX_set_default_passwd_cb_userdata.
3403 Compatibility note: As many C implementations push function arguments
3404 onto the stack in reverse order, the new library version is likely to
3405 interoperate with programs that have been compiled with the old
3406 pem_password_cb definition (PEM_whatever takes some data that
3407 happens to be on the stack as its last argument, and the callback
3408 just ignores this garbage); but there is no guarantee whatsoever that
3411 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3412 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3413 problems not only on Windows, but also on some Unix platforms.
3414 To avoid problematic command lines, these definitions are now in an
3415 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3416 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3419 *) MIPS III/IV assembler module is reimplemented.
3422 *) More DES library cleanups: remove references to srand/rand and
3423 delete an unused file.
3426 *) Add support for the the free Netwide assembler (NASM) under Win32,
3427 since not many people have MASM (ml) and it can be hard to obtain.
3428 This is currently experimental but it seems to work OK and pass all
3429 the tests. Check out INSTALL.W32 for info.
3432 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3433 without temporary keys kept an extra copy of the server key,
3434 and connections with temporary keys did not free everything in case
3438 *) New function RSA_check_key and new openssl rsa option -check
3439 for verifying the consistency of RSA keys.
3440 [Ulf Moeller, Bodo Moeller]
3442 *) Various changes to make Win32 compile work:
3443 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3444 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3445 comparison" warnings.
3446 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3449 *) Add a debugging option to PKCS#5 v2 key generation function: when
3450 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3451 derived keys are printed to stderr.
3454 *) Copy the flags in ASN1_STRING_dup().
3455 [Roman E. Pavlov <pre@mo.msk.ru>]
3457 *) The x509 application mishandled signing requests containing DSA
3458 keys when the signing key was also DSA and the parameters didn't match.
3460 It was supposed to omit the parameters when they matched the signing key:
3461 the verifying software was then supposed to automatically use the CA's
3462 parameters if they were absent from the end user certificate.
3464 Omitting parameters is no longer recommended. The test was also
3465 the wrong way round! This was probably due to unusual behaviour in
3466 EVP_cmp_parameters() which returns 1 if the parameters match.
3467 This meant that parameters were omitted when they *didn't* match and
3468 the certificate was useless. Certificates signed with 'ca' didn't have
3470 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3472 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3473 The interface is as follows:
3474 Applications can use
3475 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3476 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3477 "off" is now the default.
3478 The library internally uses
3479 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3480 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3481 to disable memory-checking temporarily.
3483 Some inconsistent states that previously were possible (and were
3484 even the default) are now avoided.
3486 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3487 with each memory chunk allocated; this is occasionally more helpful
3488 than just having a counter.
3490 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3492 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3496 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3497 which largely parallels "options", but is for changing API behaviour,
3498 whereas "options" are about protocol behaviour.
3499 Initial "mode" flags are:
3501 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3502 a single record has been written.
3503 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3504 retries use the same buffer location.
3505 (But all of the contents must be
3509 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3512 *) Fix problems with no-hmac etc.
3513 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3515 *) New functions RSA_get_default_method(), RSA_set_method() and
3516 RSA_get_method(). These allows replacement of RSA_METHODs without having
3517 to mess around with the internals of an RSA structure.
3520 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3521 Also really enable memory leak checks in openssl.c and in some
3523 [Chad C. Mulligan, Bodo Moeller]
3525 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3526 up the length of negative integers. This has now been simplified to just
3527 store the length when it is first determined and use it later, rather
3528 than trying to keep track of where data is copied and updating it to
3530 [Steve Henson, reported by Brien Wheeler
3531 <bwheeler@authentica-security.com>]
3533 *) Add a new function PKCS7_signatureVerify. This allows the verification
3534 of a PKCS#7 signature but with the signing certificate passed to the
3535 function itself. This contrasts with PKCS7_dataVerify which assumes the
3536 certificate is present in the PKCS#7 structure. This isn't always the
3537 case: certificates can be omitted from a PKCS#7 structure and be
3538 distributed by "out of band" means (such as a certificate database).
3541 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3542 function prototypes in pem.h, also change util/mkdef.pl to add the
3543 necessary function names.
3546 *) mk1mf.pl (used by Windows builds) did not properly read the
3547 options set by Configure in the top level Makefile, and Configure
3548 was not even able to write more than one option correctly.
3549 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3552 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3553 file to be loaded from a BIO or FILE pointer. The BIO version will
3554 for example allow memory BIOs to contain config info.
3557 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3558 Whoever hopes to achieve shared-library compatibility across versions
3559 must use this, not the compile-time macro.
3560 (Exercise 0.9.4: Which is the minimum library version required by
3562 Note: All this applies only to multi-threaded programs, others don't
3566 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3567 through a BIO pair triggered the default case, i.e.
3568 SSLerr(...,SSL_R_UNKNOWN_STATE).
3571 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3572 can use the SSL library even if none of the specific BIOs is
3576 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3577 for the encoded length.
3578 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3580 *) Add initial documentation of the X509V3 functions.
3583 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3584 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3585 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3586 secure PKCS#8 private key format with a high iteration count.
3589 *) Fix determination of Perl interpreter: A perl or perl5
3590 _directory_ in $PATH was also accepted as the interpreter.
3591 [Ralf S. Engelschall]
3593 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3594 wrong with it but it was very old and did things like calling
3595 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3599 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3600 to use the new extension code.
3603 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3604 with macros. This should make it easier to change their form, add extra
3605 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3609 *) Add to configuration table a new entry that can specify an alternative
3610 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3611 according to Mark Crispin <MRC@Panda.COM>.
3615 *) DES CBC did not update the IV. Weird.
3618 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3619 Changing the behaviour of the former might break existing programs --
3620 where IV updating is needed, des_ncbc_encrypt can be used.
3623 *) When bntest is run from "make test" it drives bc to check its
3624 calculations, as well as internally checking them. If an internal check
3625 fails, it needs to cause bc to give a non-zero result or make test carries
3626 on without noticing the failure. Fixed.
3629 *) DES library cleanups.
3632 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3633 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3634 ciphers. NOTE: although the key derivation function has been verified
3635 against some published test vectors it has not been extensively tested
3636 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3640 *) Instead of "mkdir -p", which is not fully portable, use new
3641 Perl script "util/mkdir-p.pl".
3644 *) Rewrite the way password based encryption (PBE) is handled. It used to
3645 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3646 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3647 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3648 the 'parameter' field of the AlgorithmIdentifier is passed to the
3649 underlying key generation function so it must do its own ASN1 parsing.
3650 This has also changed the EVP_PBE_CipherInit() function which now has a
3651 'parameter' argument instead of literal salt and iteration count values
3652 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3655 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3656 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3657 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3658 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3659 value was just used as a "magic string" and not used directly its
3660 value doesn't matter.
3663 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3667 *) "linux-sparc64" configuration (ultrapenguin).
3668 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3669 "linux-sparc" configuration.
3670 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3672 *) config now generates no-xxx options for missing ciphers.
3675 *) Support the EBCDIC character set (work in progress).
3676 File ebcdic.c not yet included because it has a different license.
3677 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3679 *) Support BS2000/OSD-POSIX.
3680 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3682 *) Make callbacks for key generation use void * instead of char *.
3685 *) Make S/MIME samples compile (not yet tested).
3688 *) Additional typesafe stacks.
3691 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3695 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3697 *) New configuration variant "sco5-gcc".
3699 *) Updated some demos.
3700 [Sean O Riordain, Wade Scholine]
3702 *) Add missing BIO_free at exit of pkcs12 application.
3705 *) Fix memory leak in conf.c.
3708 *) Updates for Win32 to assembler version of MD5.
3711 *) Set #! path to perl in apps/der_chop to where we found it
3712 instead of using a fixed path.
3715 *) SHA library changes for irix64-mips4-cc.
3718 *) Improvements for VMS support.
3722 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3724 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3725 This also avoids the problems with SC4.2 and unpatched SC5.
3726 [Andy Polyakov <appro@fy.chalmers.se>]
3728 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3729 These are required because of the typesafe stack would otherwise break
3730 existing code. If old code used a structure member which used to be STACK
3731 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3732 sk_num or sk_value it would produce an error because the num, data members
3733 are not present in STACK_OF. Now it just produces a warning. sk_set
3734 replaces the old method of assigning a value to sk_value
3735 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3736 that does this will no longer work (and should use sk_set instead) but
3737 this could be regarded as a "questionable" behaviour anyway.
3740 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3741 correctly handle encrypted S/MIME data.
3744 *) Change type of various DES function arguments from des_cblock
3745 (which means, in function argument declarations, pointer to char)
3746 to des_cblock * (meaning pointer to array with 8 char elements),
3747 which allows the compiler to do more typechecking; it was like
3748 that back in SSLeay, but with lots of ugly casts.
3750 Introduce new type const_des_cblock.
3753 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3754 problems: find RecipientInfo structure that matches recipient certificate
3755 and initialise the ASN1 structures properly based on passed cipher.
3758 *) Belatedly make the BN tests actually check the results.
3761 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3762 to and from BNs: it was completely broken. New compilation option
3763 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3764 key elements as negative integers.
3767 *) Reorganize and speed up MD5.
3768 [Andy Polyakov <appro@fy.chalmers.se>]
3771 [Richard Levitte <richard@levitte.org>]
3773 *) New option -out to asn1parse to allow the parsed structure to be
3774 output to a file. This is most useful when combined with the -strparse
3775 option to examine the output of things like OCTET STRINGS.
3778 *) Make SSL library a little more fool-proof by not requiring any longer
3779 that SSL_set_{accept,connect}_state be called before
3780 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3781 in many applications because usually everything *appeared* to work as
3782 intended anyway -- now it really works as intended).
3785 *) Move openssl.cnf out of lib/.
3788 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3789 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3790 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3791 [Ralf S. Engelschall]
3793 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3794 handle PKCS#7 enveloped data properly.
3795 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3797 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3798 copying pointers. The cert_st handling is changed by this in
3799 various ways (and thus what used to be known as ctx->default_cert
3800 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3801 any longer when s->cert does not give us what we need).
3802 ssl_cert_instantiate becomes obsolete by this change.
3803 As soon as we've got the new code right (possibly it already is?),
3804 we have solved a couple of bugs of the earlier code where s->cert
3805 was used as if it could not have been shared with other SSL structures.
3807 Note that using the SSL API in certain dirty ways now will result
3808 in different behaviour than observed with earlier library versions:
3809 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3810 does not influence s as it used to.
3812 In order to clean up things more thoroughly, inside SSL_SESSION
3813 we don't use CERT any longer, but a new structure SESS_CERT
3814 that holds per-session data (if available); currently, this is
3815 the peer's certificate chain and, for clients, the server's certificate
3816 and temporary key. CERT holds only those values that can have
3817 meaningful defaults in an SSL_CTX.
3820 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3821 from the internal representation. Various PKCS#7 fixes: remove some
3822 evil casts and set the enc_dig_alg field properly based on the signing
3826 *) Allow PKCS#12 password to be set from the command line or the
3827 environment. Let 'ca' get its config file name from the environment
3828 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3832 *) Allow certificate policies extension to use an IA5STRING for the
3833 organization field. This is contrary to the PKIX definition but
3834 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3838 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3839 without disallowing inline assembler and the like for non-pedantic builds.
3842 *) Support Borland C++ builder.
3843 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3848 *) SHA-1 cleanups and performance enhancements.
3849 [Andy Polyakov <appro@fy.chalmers.se>]
3851 *) Sparc v8plus assembler for the bignum library.
3852 [Andy Polyakov <appro@fy.chalmers.se>]
3854 *) Accept any -xxx and +xxx compiler options in Configure.
3857 *) Update HPUX configuration.
3860 *) Add missing sk_<type>_unshift() function to safestack.h
3861 [Ralf S. Engelschall]
3863 *) New function SSL_CTX_use_certificate_chain_file that sets the
3864 "extra_cert"s in addition to the certificate. (This makes sense
3865 only for "PEM" format files, as chains as a whole are not
3869 *) Support verify_depth from the SSL API.
3870 x509_vfy.c had what can be considered an off-by-one-error:
3871 Its depth (which was not part of the external interface)
3872 was actually counting the number of certificates in a chain;
3873 now it really counts the depth.
3876 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3877 instead of X509err, which often resulted in confusing error
3878 messages since the error codes are not globally unique
3879 (e.g. an alleged error in ssl3_accept when a certificate
3880 didn't match the private key).
3882 *) New function SSL_CTX_set_session_id_context that allows to set a default
3883 value (so that you don't need SSL_set_session_id_context for each
3884 connection using the SSL_CTX).
3887 *) OAEP decoding bug fix.
3890 *) Support INSTALL_PREFIX for package builders, as proposed by
3894 *) New Configure options "threads" and "no-threads". For systems
3895 where the proper compiler options are known (currently Solaris
3896 and Linux), "threads" is the default.
3899 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3902 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3903 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3904 such as /usr/local/bin.
3907 *) "make linux-shared" to build shared libraries.
3908 [Niels Poppe <niels@netbox.org>]
3910 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3913 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3914 extension adding in x509 utility.
3917 *) Remove NOPROTO sections and error code comments.
3920 *) Partial rewrite of the DEF file generator to now parse the ANSI
3924 *) New Configure options --prefix=DIR and --openssldir=DIR.
3927 *) Complete rewrite of the error code script(s). It is all now handled
3928 by one script at the top level which handles error code gathering,
3929 header rewriting and C source file generation. It should be much better
3930 than the old method: it now uses a modified version of Ulf's parser to
3931 read the ANSI prototypes in all header files (thus the old K&R definitions
3932 aren't needed for error creation any more) and do a better job of
3933 translating function codes into names. The old 'ASN1 error code imbedded
3934 in a comment' is no longer necessary and it doesn't use .err files which
3935 have now been deleted. Also the error code call doesn't have to appear all
3936 on one line (which resulted in some large lines...).
3939 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3942 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3943 0 (which usually indicates a closed connection), but continue reading.
3946 *) Fix some race conditions.
3949 *) Add support for CRL distribution points extension. Add Certificate
3950 Policies and CRL distribution points documentation.
3953 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3956 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3957 8 of keying material. Merlin has also confirmed interop with this fix
3958 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3959 [Merlin Hughes <merlin@baltimore.ie>]
3961 *) Fix lots of warnings.
3962 [Richard Levitte <levitte@stacken.kth.se>]
3964 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3965 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3966 [Richard Levitte <levitte@stacken.kth.se>]
3968 *) Fix problems with sizeof(long) == 8.
3969 [Andy Polyakov <appro@fy.chalmers.se>]
3971 *) Change functions to ANSI C.
3974 *) Fix typos in error codes.
3975 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3977 *) Remove defunct assembler files from Configure.
3980 *) SPARC v8 assembler BIGNUM implementation.
3981 [Andy Polyakov <appro@fy.chalmers.se>]
3983 *) Support for Certificate Policies extension: both print and set.
3984 Various additions to support the r2i method this uses.
3987 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3988 return a const string when you are expecting an allocated buffer.
3991 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3992 types DirectoryString and DisplayText.
3995 *) Add code to allow r2i extensions to access the configuration database,
3996 add an LHASH database driver and add several ctx helper functions.
3999 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4000 fail when they extended the size of a BIGNUM.
4003 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4004 support typesafe stack.
4007 *) Fix typo in SSL_[gs]et_options().
4008 [Nils Frostberg <nils@medcom.se>]
4010 *) Delete various functions and files that belonged to the (now obsolete)
4011 old X509V3 handling code.
4014 *) New Configure option "rsaref".
4017 *) Don't auto-generate pem.h.
4020 *) Introduce type-safe ASN.1 SETs.
4023 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4024 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4026 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4027 that links with OpenSSL (well at least cause lots of warnings), but fear
4028 not: the conversion is trivial, and it eliminates loads of evil casts. A
4029 few STACKed things have been converted already. Feel free to convert more.
4030 In the fullness of time, I'll do away with the STACK type altogether.
4033 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4034 specified in <certfile> by updating the entry in the index.txt file.
4035 This way one no longer has to edit the index.txt file manually for
4036 revoking a certificate. The -revoke option does the gory details now.
4037 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4039 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4040 `-text' option at all and this way the `-noout -text' combination was
4041 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4042 [Ralf S. Engelschall]
4044 *) Make sure a corresponding plain text error message exists for the
4045 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4046 verify callback function determined that a certificate was revoked.
4047 [Ralf S. Engelschall]
4049 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4050 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4051 all available cipers including rc5, which was forgotten until now.
4052 In order to let the testing shell script know which algorithms
4053 are available, a new (up to now undocumented) command
4054 "openssl list-cipher-commands" is used.
4057 *) Bugfix: s_client occasionally would sleep in select() when
4058 it should have checked SSL_pending() first.
4061 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4062 the raw DSA values prior to ASN.1 encoding.
4065 *) Tweaks to Configure
4066 [Niels Poppe <niels@netbox.org>]
4068 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4072 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4075 *) New config option to avoid instructions that are illegal on the 80386.
4076 The default code is faster, but requires at least a 486.
4079 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4080 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4081 same as SSL2_VERSION anyway.
4084 *) New "-showcerts" option for s_client.
4087 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4088 application. Various cleanups and fixes.
4091 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4092 modify error routines to work internally. Add error codes and PBE init
4093 to library startup routines.
4096 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4097 packing functions to asn1 and evp. Changed function names and error
4098 codes along the way.
4101 *) PKCS12 integration: and so it begins... First of several patches to
4102 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4103 objects to objects.h
4106 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4107 and display support for Thawte strong extranet extension.
4110 *) Add LinuxPPC support.
4111 [Jeff Dubrule <igor@pobox.org>]
4113 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4114 bn_div_words in alpha.s.
4115 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4117 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4118 OAEP isn't supported when OpenSSL is built with RSAref.
4119 [Ulf Moeller <ulf@fitug.de>]
4121 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4122 so they no longer are missing under -DNOPROTO.
4123 [Soren S. Jorvang <soren@t.dk>]
4126 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4128 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4129 doesn't work when the session is reused. Coming soon!
4132 *) Fix a security hole, that allows sessions to be reused in the wrong
4133 context thus bypassing client cert protection! All software that uses
4134 client certs and session caches in multiple contexts NEEDS PATCHING to
4135 allow session reuse! A fuller solution is in the works.
4136 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4138 *) Some more source tree cleanups (removed obsolete files
4139 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4140 permission on "config" script to be executable) and a fix for the INSTALL
4142 [Ulf Moeller <ulf@fitug.de>]
4144 *) Remove some legacy and erroneous uses of malloc, free instead of
4146 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4148 *) Make rsa_oaep_test return non-zero on error.
4149 [Ulf Moeller <ulf@fitug.de>]
4151 *) Add support for native Solaris shared libraries. Configure
4152 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4153 if someone would make that last step automatic.
4154 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4156 *) ctx_size was not built with the right compiler during "make links". Fixed.
4159 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4160 except NULL ciphers". This means the default cipher list will no longer
4161 enable NULL ciphers. They need to be specifically enabled e.g. with
4162 the string "DEFAULT:eNULL".
4165 *) Fix to RSA private encryption routines: if p < q then it would
4166 occasionally produce an invalid result. This will only happen with
4167 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4170 *) Be less restrictive and allow also `perl util/perlpath.pl
4171 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4172 because this way one can also use an interpreter named `perl5' (which is
4173 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4174 installed as `perl').
4175 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4177 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4178 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4180 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4181 advapi32.lib to Win32 build and change the pem test comparision
4182 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4183 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4184 and crypto/des/ede_cbcm_enc.c.
4187 *) DES quad checksum was broken on big-endian architectures. Fixed.
4190 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4191 Win32 test batch file so it (might) work again. The Win32 test batch file
4192 is horrible: I feel ill....
4195 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4196 in e_os.h. Audit of header files to check ANSI and non ANSI
4197 sections: 10 functions were absent from non ANSI section and not exported
4198 from Windows DLLs. Fixed up libeay.num for new functions.
4201 *) Make `openssl version' output lines consistent.
4202 [Ralf S. Engelschall]
4204 *) Fix Win32 symbol export lists for BIO functions: Added
4205 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4206 to ms/libeay{16,32}.def.
4207 [Ralf S. Engelschall]
4209 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4210 fine under Unix and passes some trivial tests I've now added. But the
4211 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4212 added to make sure no one expects that this stuff really works in the
4213 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4214 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4216 [Ralf S. Engelschall]
4218 *) Fix the generation of two part addresses in perl.
4219 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4221 *) Add config entry for Linux on MIPS.
4222 [John Tobey <jtobey@channel1.com>]
4224 *) Make links whenever Configure is run, unless we are on Windoze.
4227 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4228 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4232 *) Add a useful kludge to allow package maintainers to specify compiler and
4233 other platforms details on the command line without having to patch the
4234 Configure script everytime: One now can use ``perl Configure
4235 <id>:<details>'', i.e. platform ids are allowed to have details appended
4236 to them (seperated by colons). This is treated as there would be a static
4237 pre-configured entry in Configure's %table under key <id> with value
4238 <details> and ``perl Configure <id>'' is called. So, when you want to
4239 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4240 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4241 now, which overrides the FreeBSD-elf entry on-the-fly.
4242 [Ralf S. Engelschall]
4244 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4247 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4248 on the `perl Configure ...' command line. This way one can compile
4249 OpenSSL libraries with Position Independent Code (PIC) which is needed
4250 for linking it into DSOs.
4251 [Ralf S. Engelschall]
4253 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4257 *) Cleaned up the LICENSE document: The official contact for any license
4258 questions now is the OpenSSL core team under openssl-core@openssl.org.
4259 And add a paragraph about the dual-license situation to make sure people
4260 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4261 to the OpenSSL toolkit.
4262 [Ralf S. Engelschall]
4264 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4265 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4266 Additonally cleaned up the `make links' target: Remove unnecessary
4267 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4268 to speed processing and no longer clutter the display with confusing
4269 stuff. Instead only the actually done links are displayed.
4270 [Ralf S. Engelschall]
4272 *) Permit null encryption ciphersuites, used for authentication only. It used
4273 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4274 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4278 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4279 signed attributes when verifying signatures (this would break them),
4280 the detached data encoding was wrong and public keys obtained using
4281 X509_get_pubkey() weren't freed.
4284 *) Add text documentation for the BUFFER functions. Also added a work around
4285 to a Win95 console bug. This was triggered by the password read stuff: the
4286 last character typed gets carried over to the next fread(). If you were
4287 generating a new cert request using 'req' for example then the last
4288 character of the passphrase would be CR which would then enter the first
4292 *) Added the new `Includes OpenSSL Cryptography Software' button as
4293 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4294 button and can be used by applications based on OpenSSL to show the
4295 relationship to the OpenSSL project.
4296 [Ralf S. Engelschall]
4298 *) Remove confusing variables in function signatures in files
4299 ssl/ssl_lib.c and ssl/ssl.h.
4300 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4302 *) Don't install bss_file.c under PREFIX/include/
4303 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4305 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4306 functions that return function pointers and has support for NT specific
4307 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4308 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4309 unsigned to signed types: this was killing the Win32 compile.
4312 *) Add new certificate file to stack functions,
4313 SSL_add_dir_cert_subjects_to_stack() and
4314 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4315 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4316 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4317 This means that Apache-SSL and similar packages don't have to mess around
4318 to add as many CAs as they want to the preferred list.
4321 *) Experiment with doxygen documentation. Currently only partially applied to
4323 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4324 openssl.doxy as the configuration file.
4327 *) Get rid of remaining C++-style comments which strict C compilers hate.
4328 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4330 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4331 compiled in by default: it has problems with large keys.
4334 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4335 DH private keys and/or callback functions which directly correspond to
4336 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4337 is needed for applications which have to configure certificates on a
4338 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4340 For the RSA certificate situation is makes no difference, but
4341 for the DSA certificate situation this fixes the "no shared cipher"
4342 problem where the OpenSSL cipher selection procedure failed because the
4343 temporary keys were not overtaken from the context and the API provided
4344 no way to reconfigure them.
4345 The new functions now let applications reconfigure the stuff and they
4346 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4347 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4348 non-public-API function ssl_cert_instantiate() is used as a helper
4349 function and also to reduce code redundancy inside ssl_rsa.c.
4350 [Ralf S. Engelschall]
4352 *) Move s_server -dcert and -dkey options out of the undocumented feature
4353 area because they are useful for the DSA situation and should be
4354 recognized by the users.
4355 [Ralf S. Engelschall]
4357 *) Fix the cipher decision scheme for export ciphers: the export bits are
4358 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4359 SSL_EXP_MASK. So, the original variable has to be used instead of the
4360 already masked variable.
4361 [Richard Levitte <levitte@stacken.kth.se>]
4363 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4364 [Richard Levitte <levitte@stacken.kth.se>]
4366 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4367 from `int' to `unsigned int' because it's a length and initialized by
4368 EVP_DigestFinal() which expects an `unsigned int *'.
4369 [Richard Levitte <levitte@stacken.kth.se>]
4371 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4372 script. Instead use the usual Shell->Perl transition trick.
4373 [Ralf S. Engelschall]
4375 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4376 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4377 -noout -modulus' as it's already the case for `openssl rsa -noout
4378 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4379 currently the public key is printed (a decision which was already done by
4380 `openssl dsa -modulus' in the past) which serves a similar purpose.
4381 Additionally the NO_RSA no longer completely removes the whole -modulus
4382 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4384 [Ralf S. Engelschall]
4386 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4387 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4388 [Arne Ansper <arne@ats.cyber.ee>]
4390 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4391 to be added. Now both 'req' and 'ca' can use new objects defined in the
4395 *) Add cool BIO that does syslog (or event log on NT).
4396 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4398 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4399 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4400 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4401 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4404 *) Add preliminary config info for new extension code.
4407 *) Make RSA_NO_PADDING really use no padding.
4408 [Ulf Moeller <ulf@fitug.de>]
4410 *) Generate errors when private/public key check is done.
4413 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4414 for some CRL extensions and new objects added.
4417 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4418 key usage extension and fuller support for authority key id.
4421 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4422 padding method for RSA, which is recommended for new applications in PKCS
4423 #1 v2.0 (RFC 2437, October 1998).
4424 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4425 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4426 against Bleichbacher's attack on RSA.
4427 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4430 *) Updates to the new SSL compression code
4431 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4433 *) Fix so that the version number in the master secret, when passed
4434 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4435 (because the server will not accept higher), that the version number
4436 is 0x03,0x01, not 0x03,0x00
4437 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4439 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4440 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4441 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4444 *) Support for RAW extensions where an arbitrary extension can be
4445 created by including its DER encoding. See apps/openssl.cnf for
4449 *) Make sure latest Perl versions don't interpret some generated C array
4450 code as Perl array code in the crypto/err/err_genc.pl script.
4451 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4453 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4454 not many people have the assembler. Various Win32 compilation fixes and
4455 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4459 *) Modify configure script 'Configure' to automatically create crypto/date.h
4460 file under Win32 and also build pem.h from pem.org. New script
4461 util/mkfiles.pl to create the MINFO file on environments that can't do a
4462 'make files': perl util/mkfiles.pl >MINFO should work.
4465 *) Major rework of DES function declarations, in the pursuit of correctness
4466 and purity. As a result, many evil casts evaporated, and some weirdness,
4467 too. You may find this causes warnings in your code. Zapping your evil
4468 casts will probably fix them. Mostly.
4471 *) Fix for a typo in asn1.h. Bug fix to object creation script
4472 obj_dat.pl. It considered a zero in an object definition to mean
4473 "end of object": none of the objects in objects.h have any zeros
4474 so it wasn't spotted.
4475 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4477 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4478 Masking (CBCM). In the absence of test vectors, the best I have been able
4479 to do is check that the decrypt undoes the encrypt, so far. Send me test
4480 vectors if you have them.
4483 *) Correct calculation of key length for export ciphers (too much space was
4484 allocated for null ciphers). This has not been tested!
4487 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4488 message is now correct (it understands "crypto" and "ssl" on its
4489 command line). There is also now an "update" option. This will update
4490 the util/ssleay.num and util/libeay.num files with any new functions.
4492 perl util/mkdef.pl crypto ssl update
4493 it will update them.
4496 *) Overhauled the Perl interface (perl/*):
4497 - ported BN stuff to OpenSSL's different BN library
4498 - made the perl/ source tree CVS-aware
4499 - renamed the package from SSLeay to OpenSSL (the files still contain
4500 their history because I've copied them in the repository)
4501 - removed obsolete files (the test scripts will be replaced
4502 by better Test::Harness variants in the future)
4503 [Ralf S. Engelschall]
4505 *) First cut for a very conservative source tree cleanup:
4506 1. merge various obsolete readme texts into doc/ssleay.txt
4507 where we collect the old documents and readme texts.
4508 2. remove the first part of files where I'm already sure that we no
4509 longer need them because of three reasons: either they are just temporary
4510 files which were left by Eric or they are preserved original files where
4511 I've verified that the diff is also available in the CVS via "cvs diff
4512 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4513 the crypto/md/ stuff).
4514 [Ralf S. Engelschall]
4516 *) More extension code. Incomplete support for subject and issuer alt
4517 name, issuer and authority key id. Change the i2v function parameters
4518 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4519 what that's for :-) Fix to ASN1 macro which messed up
4520 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4523 *) Preliminary support for ENUMERATED type. This is largely copied from the
4527 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4528 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4530 *) Make sure `make rehash' target really finds the `openssl' program.
4531 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4533 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4534 like to hear about it if this slows down other processors.
4537 *) Add CygWin32 platform information to Configure script.
4538 [Alan Batie <batie@aahz.jf.intel.com>]
4540 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4541 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4543 *) New program nseq to manipulate netscape certificate sequences
4546 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4550 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4551 but the BN code had some problems that would cause failures when
4552 doing certificate verification and some other functions.
4553 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4555 *) Add ASN1 and PEM code to support netscape certificate sequences.
4558 *) Add ASN1 and PEM code to support netscape certificate sequences.
4561 *) Add several PKIX and private extended key usage OIDs.
4564 *) Modify the 'ca' program to handle the new extension code. Modify
4565 openssl.cnf for new extension format, add comments.
4568 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
4569 and add a sample to openssl.cnf so req -x509 now adds appropriate
4573 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
4574 error code, add initial support to X509_print() and x509 application.
4577 *) Takes a deep breath and start addding X509 V3 extension support code. Add
4578 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
4579 stuff is currently isolated and isn't even compiled yet.
4582 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
4583 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
4584 Removed the versions check from X509 routines when loading extensions:
4585 this allows certain broken certificates that don't set the version
4586 properly to be processed.
4589 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
4590 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
4591 can still be regenerated with "make depend".
4594 *) Spelling mistake in C version of CAST-128.
4595 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
4597 *) Changes to the error generation code. The perl script err-code.pl
4598 now reads in the old error codes and retains the old numbers, only
4599 adding new ones if necessary. It also only changes the .err files if new
4600 codes are added. The makefiles have been modified to only insert errors
4601 when needed (to avoid needlessly modifying header files). This is done
4602 by only inserting errors if the .err file is newer than the auto generated
4603 C file. To rebuild all the error codes from scratch (the old behaviour)
4604 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
4605 or delete all the .err files.
4608 *) CAST-128 was incorrectly implemented for short keys. The C version has
4609 been fixed, but is untested. The assembler versions are also fixed, but
4610 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
4611 to regenerate it if needed.
4612 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4613 Hagino <itojun@kame.net>]
4615 *) File was opened incorrectly in randfile.c.
4616 [Ulf Möller <ulf@fitug.de>]
4618 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4619 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4620 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4621 al: it's just almost always a UTCTime. Note this patch adds new error
4622 codes so do a "make errors" if there are problems.
4625 *) Correct Linux 1 recognition in config.
4626 [Ulf Möller <ulf@fitug.de>]
4628 *) Remove pointless MD5 hash when using DSA keys in ca.
4629 [Anonymous <nobody@replay.com>]
4631 *) Generate an error if given an empty string as a cert directory. Also
4632 generate an error if handed NULL (previously returned 0 to indicate an
4633 error, but didn't set one).
4634 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4636 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4639 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4640 parameters. This was causing a warning which killed off the Win32 compile.
4643 *) Remove C++ style comments from crypto/bn/bn_local.h.
4644 [Neil Costigan <neil.costigan@celocom.com>]
4646 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4647 based on a text string, looking up short and long names and finally
4648 "dot" format. The "dot" format stuff didn't work. Added new function
4649 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4650 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4651 OID is not part of the table.
4654 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4655 X509_LOOKUP_by_alias().
4658 *) Sort openssl functions by name.
4661 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4662 encryption from sample DSA keys (in case anyone is interested the password
4666 *) Make _all_ *_free functions accept a NULL pointer.
4667 [Frans Heymans <fheymans@isaserver.be>]
4669 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4671 [Anonymous <nobody@replay.com>]
4673 *) s_server should send the CAfile as acceptable CAs, not its own cert.
4674 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4676 *) Don't blow it for numeric -newkey arguments to apps/req.
4677 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4679 *) Temp key "for export" tests were wrong in s3_srvr.c.
4680 [Anonymous <nobody@replay.com>]
4682 *) Add prototype for temp key callback functions
4683 SSL_CTX_set_tmp_{rsa,dh}_callback().
4686 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4687 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4690 *) X509_name_add_entry() freed the wrong thing after an error.
4691 [Arne Ansper <arne@ats.cyber.ee>]
4693 *) rsa_eay.c would attempt to free a NULL context.
4694 [Arne Ansper <arne@ats.cyber.ee>]
4696 *) BIO_s_socket() had a broken should_retry() on Windoze.
4697 [Arne Ansper <arne@ats.cyber.ee>]
4699 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4700 [Arne Ansper <arne@ats.cyber.ee>]
4702 *) Make sure the already existing X509_STORE->depth variable is initialized
4703 in X509_STORE_new(), but document the fact that this variable is still
4704 unused in the certificate verification process.
4705 [Ralf S. Engelschall]
4707 *) Fix the various library and apps files to free up pkeys obtained from
4708 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4711 *) Fix reference counting in X509_PUBKEY_get(). This makes
4712 demos/maurice/example2.c work, amongst others, probably.
4713 [Steve Henson and Ben Laurie]
4715 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4716 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4717 are no longer created. This way we have a single and consistent command
4718 line interface `openssl <command>', similar to `cvs <command>'.
4719 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4721 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4722 BIT STRING wrapper always have zero unused bits.
4725 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4728 *) Make the top-level INSTALL documentation easier to understand.
4731 *) Makefiles updated to exit if an error occurs in a sub-directory
4732 make (including if user presses ^C) [Paul Sutton]
4734 *) Make Montgomery context stuff explicit in RSA data structure.
4737 *) Fix build order of pem and err to allow for generated pem.h.
4740 *) Fix renumbering bug in X509_NAME_delete_entry().
4743 *) Enhanced the err-ins.pl script so it makes the error library number
4744 global and can add a library name. This is needed for external ASN1 and
4745 other error libraries.
4748 *) Fixed sk_insert which never worked properly.
4751 *) Fix ASN1 macros so they can handle indefinite length construted
4752 EXPLICIT tags. Some non standard certificates use these: they can now
4756 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4757 into a single doc/ssleay.txt bundle. This way the information is still
4758 preserved but no longer messes up this directory. Now it's new room for
4759 the new set of documenation files.
4760 [Ralf S. Engelschall]
4762 *) SETs were incorrectly DER encoded. This was a major pain, because they
4763 shared code with SEQUENCEs, which aren't coded the same. This means that
4764 almost everything to do with SETs or SEQUENCEs has either changed name or
4765 number of arguments.
4766 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4768 *) Fix test data to work with the above.
4771 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4772 was already fixed by Eric for 0.9.1 it seems.
4773 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4775 *) Autodetect FreeBSD3.
4778 *) Fix various bugs in Configure. This affects the following platforms:
4782 unixware-2.0-pentium
4786 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4787 before they are needed.
4790 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4794 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4796 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4797 changed SSLeay to OpenSSL in version strings.
4798 [Ralf S. Engelschall]
4800 *) Some fixups to the top-level documents.
4803 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4804 because the symlink to include/ was missing.
4805 [Ralf S. Engelschall]
4807 *) Incorporated the popular no-RSA/DSA-only patches
4808 which allow to compile a RSA-free SSLeay.
4809 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4811 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4812 when "ssleay" is still not found.
4813 [Ralf S. Engelschall]
4815 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4816 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4818 *) Updated the README file.
4819 [Ralf S. Engelschall]
4821 *) Added various .cvsignore files in the CVS repository subdirs
4822 to make a "cvs update" really silent.
4823 [Ralf S. Engelschall]
4825 *) Recompiled the error-definition header files and added
4826 missing symbols to the Win32 linker tables.
4827 [Ralf S. Engelschall]
4829 *) Cleaned up the top-level documents;
4830 o new files: CHANGES and LICENSE
4831 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4832 o merged COPYRIGHT into LICENSE
4833 o removed obsolete TODO file
4834 o renamed MICROSOFT to INSTALL.W32
4835 [Ralf S. Engelschall]
4837 *) Removed dummy files from the 0.9.1b source tree:
4838 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4839 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4840 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4841 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4842 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4843 [Ralf S. Engelschall]
4845 *) Added various platform portability fixes.
4848 *) The Genesis of the OpenSSL rpject:
4849 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4850 Young and Tim J. Hudson created while they were working for C2Net until
4852 [The OpenSSL Project]
4855 Changes between 0.9.0b and 0.9.1b [not released]
4857 *) Updated a few CA certificates under certs/
4860 *) Changed some BIGNUM api stuff.
4863 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4864 DGUX x86, Linux Alpha, etc.
4867 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4868 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4872 *) Add -strparse option to asn1pars program which parses nested
4874 [Dr Stephen Henson <shenson@bigfoot.com>]
4876 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4879 *) DSA fix for "ca" program.
4882 *) Added "-genkey" option to "dsaparam" program.
4885 *) Added RIPE MD160 (rmd160) message digest.
4888 *) Added -a (all) option to "ssleay version" command.
4891 *) Added PLATFORM define which is the id given to Configure.
4894 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4897 *) Extended the ASN.1 parser routines.
4900 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4903 *) Added a BN_CTX to the BN library.
4906 *) Fixed the weak key values in DES library
4909 *) Changed API in EVP library for cipher aliases.
4912 *) Added support for RC2/64bit cipher.
4915 *) Converted the lhash library to the crypto/mem.c functions.
4918 *) Added more recognized ASN.1 object ids.
4921 *) Added more RSA padding checks for SSL/TLS.
4924 *) Added BIO proxy/filter functionality.
4927 *) Added extra_certs to SSL_CTX which can be used
4928 send extra CA certificates to the client in the CA cert chain sending
4929 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4932 *) Now Fortezza is denied in the authentication phase because
4933 this is key exchange mechanism is not supported by SSLeay at all.
4936 *) Additional PKCS1 checks.
4939 *) Support the string "TLSv1" for all TLS v1 ciphers.
4942 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4943 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4946 *) Fixed a few memory leaks.
4949 *) Fixed various code and comment typos.
4952 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4953 bytes sent in the client random.
4954 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob