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 +) Added the OS2-EMX target.
15 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
17 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
18 to support NCONF routines in extension code. New function CONF_set_nconf()
19 to allow functions which take an NCONF to also handle the old LHASH
20 structure: this means that the old CONF compatible routines can be
21 retained (in particular wrt extensions) without having to duplicate the
22 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
25 *) Handle special case when X509_NAME is empty in X509 printing routines.
28 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
29 positive and less than q.
32 +) Enhance the general user interface with mechanisms for inner control
33 and with pssibilities to have yes/no kind of prompts.
36 *) Change all calls to low level digest routines in the library and
37 applications to use EVP. Add missing calls to HMAC_cleanup() and
38 don't assume HMAC_CTX can be copied using memcpy().
39 [Verdon Walker <VWalker@novell.com>, Steve Henson]
41 +) Add the possibility to control engines through control names but with
42 arbitrary arguments instead of just a string.
43 Change the key loaders to take a UI_METHOD instead of a callback
44 function pointer. NOTE: this breaks binary compatibility with earlier
45 versions of OpenSSL [engine].
46 Addapt the nCipher code for these new conditions and add a card insertion
50 +) Enhance the general user interface with mechanisms to better support
51 dialog box interfaces, application-defined prompts, the possibility
52 to use defaults (for example default passwords from somewhere else)
53 and interrupts/cancelations.
56 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
57 used: it isn't thread safe and the add_lock_callback should handle
59 [Paul Rose <Paul.Rose@bridge.com>]
61 *) Verify that incoming data obeys the block size in
62 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
65 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
66 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
70 [Ulf Möller, Bodo Möller]
72 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
73 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5
74 when fixing the server behaviour for backwards-compatible 'client
75 hello' messages. (Note that the attack is impractical against
76 SSL 3.0 and TLS 1.0 anyway because length and version checking
77 means that the probability of guessing a valid ciphertext is
78 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
81 Before 0.9.5, the countermeasure (hide the error by generating a
82 random 'decryption result') did not work properly because
83 ERR_clear_error() was missing, meaning that SSL_get_error() would
84 detect the supposedly ignored error.
86 Both problems are now fixed.
89 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
90 (previously it was 1024).
93 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
94 tidy up some unecessarily weird code in 'sk_new()').
95 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
97 +) Change the key loading routines for ENGINEs to use the same kind
98 callback (pem_password_cb) as all other routines that need this
102 *) Fix for compatibility mode trust settings: ignore trust settings
103 unless some valid trust or reject settings are present.
106 *) Fix for blowfish EVP: its a variable length cipher.
109 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
110 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
111 than this minimum value is recommended.
114 +) New random seeder for OpenVMS, using the system process statistics
115 that are easily reachable.
118 +) Windows apparently can't transparently handle global
119 variables defined in DLLs. Initialisations such as:
121 const ASN1_ITEM *it = &ASN1_INTEGER_it;
123 wont compile. This is used by the any applications that need to
124 delcare their own ASN1 modules. This was fixed by adding the option
125 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
126 needed for static libraries under Win32.
129 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
130 setting of purpose and trust fields. New X509_STORE trust and
131 purpose functions and tidy up setting in other SSL functions.
134 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
135 structure. These are inherited by X509_STORE_CTX when it is
136 initialised. This allows various defaults to be set in the
137 X509_STORE structure (such as flags for CRL checking and custom
138 purpose or trust settings) for functions which only use X509_STORE_CTX
139 internally such as S/MIME.
141 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
142 trust settings if they are not set in X509_STORE. This allows X509_STORE
143 purposes and trust (in S/MIME for example) to override any set by default.
145 Add command line options for CRL checking to smime, s_client and s_server
149 +) Initial CRL based revocation checking. If the CRL checking flag(s)
150 are set then the CRL is looked up in the X509_STORE structure and
151 its validity and signature checked, then if the certificate is found
152 in the CRL the verify fails with a revoked error.
154 Various new CRL related callbacks added to X509_STORE_CTX structure.
156 Command line options added to 'verify' application to support this.
158 This needs some additional work, such as being able to handle multiple
159 CRLs with different times, extension based lookup (rather than just
160 by subject name) and ultimately more complete V2 CRL extension
164 +) Add a general user interface API (crypto/ui/). This is designed
165 to replace things like des_read_password and friends (backward
166 compatibility functions using this new API are provided).
167 The purpose is to remove prompting functions from the DES code
168 section as well as provide for prompting through dialog boxes in
169 a window system and the like.
172 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
173 in the current directory if neither $RANDFILE nor $HOME was set.
174 RAND_file_name() in 0.9.6a returned NULL in this case. This has
175 caused some confusion to Windows users who haven't defined $HOME.
176 Thus RAND_file_name() is changed again: e_os.h can define a
177 DEFAULT_HOME, which will be used if $HOME is not set.
178 For Windows, we use "C:"; on other platforms, we still require
179 environment variables.
181 +) Add "ex_data" support to ENGINE so implementations can add state at a
182 per-structure level rather than having to store it globally.
185 +) Make it possible for ENGINE structures to be copied when retrieved by
186 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
187 This causes the "original" ENGINE structure to act like a template,
188 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
189 operational state can be localised to each ENGINE structure, despite the
190 fact they all share the same "methods". New ENGINE structures returned in
191 this case have no functional references and the return value is the single
192 structural reference. This matches the single structural reference returned
193 by ENGINE_by_id() normally, when it is incremented on the pre-existing
197 *) Fix various bugs related to DSA S/MIME verification. Handle missing
198 parameters in DSA public key structures and return an error in the
199 DSA routines if parameters are absent.
202 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
203 needs to match any other type at all we need to manually clear the
207 +) Changes to the "openssl engine" utility to include;
208 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
209 about an ENGINE's available control commands.
210 - executing control commands from command line arguments using the
211 '-pre' and '-post' switches. '-post' is only used if '-t' is
212 specified and the ENGINE is successfully initialised. The syntax for
213 the individual commands are colon-separated, for example;
214 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
217 +) New dynamic control command support for ENGINEs. ENGINEs can now
218 declare their own commands (numbers), names (strings), descriptions,
219 and input types for run-time discovery by calling applications. A
220 subset of these commands are implicitly classed as "executable"
221 depending on their input type, and only these can be invoked through
222 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
223 can be based on user input, config files, etc). The distinction is
224 that "executable" commands cannot return anything other than a boolean
225 result and can only support numeric or string input, whereas some
226 discoverable commands may only be for direct use through
227 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
228 pointers, or other custom uses. The "executable" commands are to
229 support parameterisations of ENGINE behaviour that can be
230 unambiguously defined by ENGINEs and used consistently across any
231 OpenSSL-based application. Commands have been added to all the
232 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
233 control over shared-library paths without source code alterations.
236 +) Changed all ENGINE implementations to dynamically allocate their
237 ENGINEs rather than declaring them statically. Apart from this being
238 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
239 this also allows the implementations to compile without using the
240 internal engine_int.h header.
243 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
244 'const' value. Any code that should be able to modify a RAND_METHOD
245 should already have non-const pointers to it (ie. they should only
246 modify their own ones).
249 +) Made a variety of little tweaks to the ENGINE code.
250 - "atalla" and "ubsec" string definitions were moved from header files
251 to C code. "nuron" string definitions were placed in variables
252 rather than hard-coded - allowing parameterisation of these values
253 later on via ctrl() commands.
254 - Removed unused "#if 0"'d code.
255 - Fixed engine list iteration code so it uses ENGINE_free() to release
256 structural references.
257 - Constified the RAND_METHOD element of ENGINE structures.
258 - Constified various get/set functions as appropriate and added
259 missing functions (including a catch-all ENGINE_cpy that duplicates
260 all ENGINE values onto a new ENGINE except reference counts/state).
261 - Removed NULL parameter checks in get/set functions. Setting a method
262 or function to NULL is a way of cancelling out a previously set
263 value. Passing a NULL ENGINE parameter is just plain stupid anyway
264 and doesn't justify the extra error symbols and code.
265 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
266 flags from engine_int.h to engine.h.
267 - Changed prototypes for ENGINE handler functions (init(), finish(),
268 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
271 *) Move 'if (!initialized) RAND_poll()' into regions protected by
272 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
273 having multiple threads call RAND_poll() concurrently.
276 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
277 combination of a flag and a thread ID variable.
278 Otherwise while one thread is in ssleay_rand_bytes (which sets the
279 flag), *other* threads can enter ssleay_add_bytes without obeying
280 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock
281 that they do not hold after the first thread unsets add_do_not_lock).
284 +) Implement binary inversion algorithm for BN_mod_inverse in addition
285 to the algorithm using long divison. The binary algorithm can be
286 used only if the modulus is odd. On 32-bit systems, it is faster
287 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
288 roughly 5-15% for 256-bit moduli), so we use it only for moduli
289 up to 450 bits. In 64-bit environments, the binary algorithm
290 appears to be advantageous for much longer moduli; here we use it
291 for moduli up to 2048 bits.
294 *) Change bctest again: '-x' expressions are not available in all
298 -) OpenSSL 0.9.6a released [5 Apr 2001]
300 *) Fix a couple of memory leaks in PKCS7_dataDecode()
301 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
303 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
304 the default extension for executables, if any. Also, make the perl
305 scripts that use symlink() to test if it really exists and use "cp"
306 if it doesn't. All this made OpenSSL compilable and installable in
310 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
311 could not support the combine flag in choice fields.
314 -) Fix for asn1_GetSequence() for indefinite length constructed data.
315 If SEQUENCE is length is indefinite just set c->slen to the total
316 amount of data available.
317 [Steve Henson, reported by shige@FreeBSD.org]
318 [This change does not apply to 0.9.7.]
320 *) Change bctest to avoid here-documents inside command substitution
321 (workaround for FreeBSD /bin/sh bug).
322 For compatibility with Ultrix, avoid shell functions (introduced
323 in the bctest version that searches along $PATH).
326 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
327 with des_encrypt() defined on some operating systems, like Solaris
331 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
332 On the Importance of Eliminating Errors in Cryptographic
333 Computations, J. Cryptology 14 (2001) 2, 101-119,
334 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
337 *) MIPS assembler BIGNUM division bug fix.
340 *) Disabled incorrect Alpha assembler code.
343 -) Fix PKCS#7 decode routines so they correctly update the length
344 after reading an EOC for the EXPLICIT tag.
346 [This change does not apply to 0.9.7.]
348 *) Fix bug in PKCS#12 key generation routines. This was triggered
349 if a 3DES key was generated with a 0 initial byte. Include
350 PKCS12_BROKEN_KEYGEN compilation option to retain the old
351 (but broken) behaviour.
354 *) Enhance bctest to search for a working bc along $PATH and print
356 [Tim Rice <tim@multitalents.net> via Richard Levitte]
358 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
359 extensions from a certificate request to the certificate.
362 +) Allow multiple 'certopt' and 'nameopt' options to be separated
363 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
364 file: this allows the display of the certificate about to be
365 signed to be customised, to allow certain fields to be included
366 or excluded and extension details. The old system didn't display
367 multicharacter strings properly, omitted fields not in the policy
368 and couldn't display additional details such as extensions.
371 *) Fix memory leaks in err.c: free err_data string if necessary;
372 don't write to the wrong index in ERR_set_error_data.
375 +) Function EC_POINTs_mul for simultaneous scalar multiplication
376 of an arbitrary number of elliptic curve points, optionally
377 including the generator defined for the EC_GROUP.
378 EC_POINT_mul is a simple wrapper function for the typical case
379 that the point list has just one item (besides the optional
383 +) First EC_METHODs for curves over GF(p):
385 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
386 operations and provides various method functions that can also
387 operate with faster implementations of modular arithmetic.
389 EC_GFp_mont_method() reuses most functions that are part of
390 EC_GFp_simple_method, but uses Montgomery arithmetic.
392 [Bodo Moeller; point addition and point doubling
393 implementation directly derived from source code provided by
394 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
396 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
399 Curves are EC_GROUP objects (with an optional group generator)
400 based on EC_METHODs that are built into the library.
402 Points are EC_POINT objects based on EC_GROUP objects.
404 Most of the framework would be able to handle curves over arbitrary
405 finite fields, but as there are no obvious types for fields other
406 than GF(p), some functions are limited to that for now.
409 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
410 that the file contains a complete HTTP response.
413 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
414 change the def and num file printf format specifier from "%-40sXXX"
415 to "%-39s XXX". The latter will always guarantee a space after the
416 field while the former will cause them to run together if the field
417 is 40 of more characters long.
420 +) Constify the cipher and digest 'method' functions and structures
421 and modify related functions to take constant EVP_MD and EVP_CIPHER
425 *) Implement ssl23_peek (analogous to ssl23_read), which previously
429 *) Replace rdtsc with _emit statements for VC++ version 5.
430 [Jeremy Cooper <jeremy@baymoo.org>]
432 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
433 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
436 +) Modify EVP_Digest*() routines so they now return values. Although the
437 internal software routines can never fail additional hardware versions
441 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
443 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
444 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
447 ERR_R_NESTED_ASN1_ERROR
449 ERR_R_MISSING_ASN1_EOS
450 were 4 .. 9, conflicting with
451 ERR_LIB_RSA (= ERR_R_RSA_LIB)
453 ERR_LIB_PEM (= ERR_R_PEM_LIB).
454 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
456 Add new error code 'ERR_R_INTERNAL_ERROR'.
459 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
463 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
464 sets the subject name for a new request or supersedes the
465 subject name in a given request. Formats that can be parsed are
466 'CN=Some Name, OU=myOU, C=IT'
468 'CN=Some Name/OU=myOU/C=IT'.
470 Add options '-batch' and '-verbose' to 'openssl req'.
471 [Massimiliano Pala <madwolf@hackmasters.net>]
473 *) Make it possible to reuse SSLv2 sessions.
476 +) Introduce the possibility to access global variables through
477 functions on platform were that's the best way to handle exporting
478 global variables in shared libraries. To enable this functionality,
479 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
480 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
481 is normally done by Configure or something similar).
483 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
484 in the source file (foo.c) like this:
486 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
487 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
489 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
490 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
492 OPENSSL_DECLARE_GLOBAL(int,foo);
493 #define foo OPENSSL_GLOBAL_REF(foo)
494 OPENSSL_DECLARE_GLOBAL(double,bar);
495 #define bar OPENSSL_GLOBAL_REF(bar)
497 The #defines are very important, and therefore so is including the
498 header file everywere where the defined globals are used.
500 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
501 of ASN.1 items, but that structure is a bt different.
503 The largest change is in util/mkdef.pl which has been enhanced with
504 better and easier to understand logic to choose which symbols should
505 go into the Windows .def files as well as a number of fixes and code
506 cleanup (among others, algorithm keywords are now sorted
507 lexicographically to avoid constant rewrites).
510 *) In copy_email() check for >= 0 as a return value for
511 X509_NAME_get_index_by_NID() since 0 is a valid index.
512 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
514 +) In BN_div() keep a copy of the sign of 'num' before writing the
515 result to 'rm' because if rm==num the value will be overwritten
516 and produce the wrong result if 'num' is negative: this caused
517 problems with BN_mod() and BN_nnmod().
520 +) Function OCSP_request_verify(). This checks the signature on an
521 OCSP request and verifies the signer certificate. The signer
522 certificate is just checked for a generic purpose and OCSP request
526 +) Add OCSP_check_validity() function to check the validity of OCSP
527 responses. OCSP responses are prepared in real time and may only
528 be a few seconds old. Simply checking that the current time lies
529 between thisUpdate and nextUpdate max reject otherwise valid responses
530 caused by either OCSP responder or client clock innacuracy. Instead
531 we allow thisUpdate and nextUpdate to fall within a certain period of
532 the current time. The age of the response can also optionally be
533 checked. Two new options -validity_period and -status_age added to
537 +) If signature or public key algorithm is unrecognized print out its
538 OID rather that just UNKOWN.
541 *) Avoid coredump with unsupported or invalid public keys by checking if
542 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
543 PKCS7_verify() fails with non detached data.
546 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
547 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
548 ID to be generated from the issuer certificate alone which can then be
549 passed to OCSP_id_issuer_cmp().
552 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
553 ASN1 modules to export functions returning ASN1_ITEM pointers
554 instead of the ASN1_ITEM structures themselves. This adds several
555 new macros which allow the underlying ASN1 function/structure to
556 be accessed transparently. As a result code should not use ASN1_ITEM
557 references directly (such as &X509_it) but instead use the relevant
558 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
559 use of the new ASN1 code on platforms where exporting structures
560 is problematical (for example in shared libraries) but exporting
561 functions returning pointers to structures is not.
564 +) Add support for overriding the generation of SSL/TLS session IDs.
565 These callbacks can be registered either in an SSL_CTX or per SSL.
566 The purpose of this is to allow applications to control, if they wish,
567 the arbitrary values chosen for use as session IDs, particularly as it
568 can be useful for session caching in multiple-server environments. A
569 command-line switch for testing this (and any client code that wishes
570 to use such a feature) has been added to "s_server".
571 [Geoff Thorpe, Lutz Jaenicke]
573 +) Modify mkdef.pl to recognise and parse prprocessor conditionals
574 of the form '#if defined(...) || defined(...) || ...' and
575 '#if !defined(...) && !defined(...) && ...'. This also avoids
576 the growing number of special cases it was previously handling.
579 *) Don't use getenv in library functions when run as setuid/setgid.
580 New function OPENSSL_issetugid().
583 +) Make all configuration macros available for application by making
584 sure they are available in opensslconf.h, by giving them names starting
585 with "OPENSSL_" to avoid conflicts with other packages and by making
586 sure e_os2.h will cover all platform-specific cases together with
588 Additionally, it is now possible to define configuration/platform-
589 specific names (called "system identities"). In the C code, these
590 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
591 macro with the name beginning with "OPENSSL_SYS_", which is determined
592 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
596 +) New option -set_serial to 'req' and 'x509' this allows the serial
597 number to use to be specified on the command line. Previously self
598 signed certificates were hard coded with serial number 0 and the
599 CA options of 'x509' had to use a serial number in a file which was
603 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
604 due to incorrect handling of multi-threading:
606 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
608 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
610 3. Count how many times MemCheck_off() has been called so that
611 nested use can be treated correctly. This also avoids
612 inband-signalling in the previous code (which relied on the
613 assumption that thread ID 0 is impossible).
616 +) New options to 'ca' utility to support V2 CRL entry extensions.
617 Currently CRL reason, invalidity date and hold instruction are
618 supported. Add new CRL extensions to V3 code and some new objects.
621 *) Add "-rand" option also to s_client and s_server.
624 +) New function EVP_CIPHER_CTX_set_padding() this is used to
625 disable standard block padding (aka PKCS#5 padding) in the EVP
626 API, which was previously mandatory. This means that the data is
627 not padded in any way and so the total length much be a multiple
628 of the block size, otherwise an error occurs.
631 +) Initial (incomplete) OCSP SSL support.
634 *) Fix CPU detection on Irix 6.x.
635 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
636 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
638 +) New function OCSP_parse_url(). This splits up a URL into its host,
639 port and path components: primarily to parse OCSP URLs. New -url
640 option to ocsp utility.
643 +) New nonce behavior. The return value of OCSP_check_nonce() now
644 reflects the various checks performed. Applications can decide
645 whether to tolerate certain situations such as an absent nonce
646 in a response when one was present in a request: the ocsp application
647 just prints out a warning. New function OCSP_add1_basic_nonce()
648 this is to allow responders to include a nonce in a response even if
649 the request is nonce-less.
652 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
654 [This change does not apply to 0.9.7.]
656 *) Use the cached encoding of an X509_NAME structure rather than
657 copying it. This is apparently the reason for the libsafe "errors"
658 but the code is actually correct.
661 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
662 skipped when using openssl x509 multiple times on a single input file,
663 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
666 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
667 set string type: to handle setting ASN1_TIME structures. Fix ca
668 utility to correctly initialize revocation date of CRLs.
671 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
672 the clients preferred ciphersuites and rather use its own preferences.
673 Should help to work around M$ SGC (Server Gated Cryptography) bug in
674 Internet Explorer by ensuring unchanged hash method during stepup.
677 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
678 to aes and add a new 'exist' option to print out symbols that don't
682 +) Additional options to ocsp utility to allow flags to be set and
683 additional certificates supplied.
686 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
687 OCSP client a number of certificate to only verify the response
691 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
692 Bleichenbacher's DSA attack.
693 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
694 to be set and top=0 forces the highest bit to be set; top=-1 is new
695 and leaves the highest bit random.
696 [Ulf Moeller, Bodo Moeller]
698 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
699 handle the new API. Currently only ECB, CBC modes supported. Add new
700 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
701 for TLS" draft-ietf-tls-ciphersuite-03.txt.
702 [Ben Laurie, Steve Henson]
704 *) In the NCONF_...-based implementations for CONF_... queries
705 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
706 a temporary CONF structure with the data component set to NULL
707 (which gives segmentation faults in lh_retrieve).
708 Instead, use NULL for the CONF pointer in CONF_get_string and
709 CONF_get_number (which may use environment variables) and directly
710 return NULL from CONF_get_section.
713 *) Fix potential buffer overrun for EBCDIC.
716 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
720 +) Functions for OCSP responders. OCSP_request_onereq_count(),
721 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
722 extract information from a certificate request. OCSP_response_create()
723 creates a response and optionally adds a basic response structure.
724 OCSP_basic_add1_status() adds a complete single response to a basic
725 reponse and returns the OCSP_SINGLERESP structure just added (to allow
726 extensions to be included for example). OCSP_basic_add1_cert() adds a
727 certificate to a basic response and OCSP_basic_sign() signs a basic
728 response with various flags. New helper functions ASN1_TIME_check()
729 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
730 (converts ASN1_TIME to GeneralizedTime).
733 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
734 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
735 structure from a certificate. X509_pubkey_digest() digests tha public_key
736 contents: this is used in various key identifiers.
739 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
740 keyUsage if basicConstraints absent for a CA.
743 *) Make SMIME_write_PKCS7() write mail header values with a format that
744 is more generally accepted (no spaces before the semicolon), since
745 some programs can't parse those values properly otherwise. Also make
746 sure BIO's that break lines after each write do not create invalid
750 +) Make sk_sort() tolerate a NULL argument.
751 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
753 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
754 passed by the function are trusted implicitly. If any of them signed the
755 reponse then it is assumed to be valid and is not verified.
758 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
759 macros previously used would not encode an empty SEQUENCE OF
760 and break the signature.
762 [This change does not apply to 0.9.7.]
764 *) Zero the premaster secret after deriving the master secret in
768 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
769 to data. This was previously part of the PKCS7 ASN1 code. This
770 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
771 [Steve Henson, reported by Kenneth R. Robinette
772 <support@securenetterm.com>]
774 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
775 routines: without these tracing memory leaks is very painful.
776 Fix leaks in PKCS12 and PKCS7 routines.
779 *) Add some EVP_add_digest_alias registrations (as found in
780 OpenSSL_add_all_digests()) to SSL_library_init()
781 aka OpenSSL_add_ssl_algorithms(). This provides improved
782 compatibility with peers using X.509 certificates
783 with unconventional AlgorithmIdentifier OIDs.
786 *) Fix for Irix with NO_ASM.
787 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
789 *) ./config script fixes.
790 [Ulf Moeller, Richard Levitte]
792 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
793 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
794 effectively meant GeneralizedTime would never be used. Now it
795 is initialised to -1 but X509_time_adj() now has to check the value
796 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
797 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
798 [Steve Henson, reported by Kenneth R. Robinette
799 <support@securenetterm.com>]
801 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
802 result in a zero length in the ASN1_INTEGER structure which was
803 not consistent with the structure when d2i_ASN1_INTEGER() was used
804 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
805 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
806 where it did not print out a minus for negative ASN1_INTEGER.
809 *) Fix 'openssl passwd -1'.
812 +) Add summary printout to ocsp utility. The various functions which
813 convert status values to strings have been renamed to:
814 OCSP_response_status_str(), OCSP_cert_status_str() and
815 OCSP_crl_reason_str() and are no longer static. New options
816 to verify nonce values and to disable verification. OCSP response
817 printout format cleaned up.
820 +) Add additional OCSP certificate checks. These are those specified
821 in RFC2560. This consists of two separate checks: the CA of the
822 certificate being checked must either be the OCSP signer certificate
823 or the issuer of the OCSP signer certificate. In the latter case the
824 OCSP signer certificate must contain the OCSP signing extended key
825 usage. This check is performed by attempting to match the OCSP
826 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
827 in the OCSP_CERTID structures of the response.
830 +) Initial OCSP certificate verification added to OCSP_basic_verify()
831 and related routines. This uses the standard OpenSSL certificate
832 verify routines to perform initial checks (just CA validity) and
833 to obtain the certificate chain. Then additional checks will be
834 performed on the chain. Currently the root CA is checked to see
835 if it is explicitly trusted for OCSP signing. This is used to set
836 a root CA as a global signing root: that is any certificate that
837 chains to that CA is an acceptable OCSP signing certificate.
840 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
841 extensions from a separate configuration file.
842 As when reading extensions from the main configuration file,
843 the '-extensions ...' option may be used for specifying the
845 [Massimiliano Pala <madwolf@comune.modena.it>]
847 *) Change PKCS12_key_gen_asc() so it can cope with non null
848 terminated strings whose length is passed in the passlen
849 parameter, for example from PEM callbacks. This was done
850 by adding an extra length parameter to asc2uni().
851 [Steve Henson, reported by <oddissey@samsung.co.kr>]
853 +) New OCSP utility. Allows OCSP requests to be generated or
854 read. The request can be sent to a responder and the output
855 parsed, outputed or printed in text form. Not complete yet:
856 still needs to check the OCSP response validity.
859 +) New subcommands for 'openssl ca':
860 'openssl ca -status <serial>' prints the status of the cert with
861 the given serial number (according to the index file).
862 'openssl ca -updatedb' updates the expiry status of certificates
864 [Massimiliano Pala <madwolf@comune.modena.it>]
866 +) New '-newreq-nodes' command option to CA.pl. This is like
867 '-newreq', but calls 'openssl req' with the '-nodes' option
868 so that the resulting key is not encrypted.
869 [Damien Miller <djm@mindrot.org>]
871 +) New configuration for the GNU Hurd.
872 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
874 +) Initial code to implement OCSP basic response verify. This
875 is currently incomplete. Currently just finds the signer's
876 certificate and verifies the signature on the response.
879 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
880 value of OPENSSLDIR. This is available via the new '-d' option
881 to 'openssl version', and is also included in 'openssl version -a'.
884 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
885 call failed, free the DSA structure.
888 +) Allowing defining memory allocation callbacks that will be given
889 file name and line number information in additional arguments
890 (a const char* and an int). The basic functionality remains, as
891 well as the original possibility to just replace malloc(),
892 realloc() and free() by functions that do not know about these
893 additional arguments. To register and find out the current
894 settings for extended allocation functions, the following
895 functions are provided:
897 CRYPTO_set_mem_ex_functions
898 CRYPTO_set_locked_mem_ex_functions
899 CRYPTO_get_mem_ex_functions
900 CRYPTO_get_locked_mem_ex_functions
902 These work the same way as CRYPTO_set_mem_functions and friends.
903 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
904 extended allocation function is enabled.
905 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
906 a conventional allocation function is enabled.
907 [Richard Levitte, Bodo Moeller]
909 *) Fix to uni2asc() to cope with zero length Unicode strings.
910 These are present in some PKCS#12 files.
913 +) Finish off removing the remaining LHASH function pointer casts.
914 There should no longer be any prototype-casting required when using
915 the LHASH abstraction, and any casts that remain are "bugs". See
916 the callback types and macros at the head of lhash.h for details
917 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
920 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
921 If /dev/[u]random devices are not available or do not return enough
922 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
924 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
925 /etc/entropy will be queried once each in this sequence, quering stops
926 when enough entropy was collected without querying more sockets.
929 +) Change the Unix RAND_poll() variant to be able to poll several
930 random devices, as specified by DEVRANDOM, until a sufficient amount
931 of data has been collected. We spend at most 10 ms on each file
932 (select timeout) and read in non-blocking mode. DEVRANDOM now
933 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
934 (previously it was just the string "/dev/urandom"), so on typical
935 platforms the 10 ms delay will never occur.
936 Also separate out the Unix variant to its own file, rand_unix.c.
937 For VMS, there's a currently-empty rand_vms.c.
940 +) Move OCSP client related routines to ocsp_cl.c. These
941 provide utility functions which an application needing
942 to issue a request to an OCSP responder and analyse the
943 response will typically need: as opposed to those which an
944 OCSP responder itself would need which will be added later.
946 OCSP_request_sign() signs an OCSP request with an API similar
947 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
948 response. OCSP_response_get1_basic() extracts basic response
949 from response. OCSP_resp_find_status(): finds and extracts status
950 information from an OCSP_CERTID structure (which will be created
951 when the request structure is built). These are built from lower
952 level functions which work on OCSP_SINGLERESP structures but
953 wont normally be used unless the application wishes to examine
954 extensions in the OCSP response for example.
956 Replace nonce routines with a pair of functions.
957 OCSP_request_add1_nonce() adds a nonce value and optionally
958 generates a random value. OCSP_check_nonce() checks the
959 validity of the nonce in an OCSP response.
962 +) Change function OCSP_request_add() to OCSP_request_add0_id().
963 This doesn't copy the supplied OCSP_CERTID and avoids the
964 need to free up the newly created id. Change return type
965 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
966 This can then be used to add extensions to the request.
967 Deleted OCSP_request_new(), since most of its functionality
968 is now in OCSP_REQUEST_new() (and the case insensitive name
969 clash) apart from the ability to set the request name which
970 will be added elsewhere.
973 +) Update OCSP API. Remove obsolete extensions argument from
974 various functions. Extensions are now handled using the new
975 OCSP extension code. New simple OCSP HTTP function which
976 can be used to send requests and parse the response.
979 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
980 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
981 uses the special reorder version of SET OF to sort the attributes
982 and reorder them to match the encoded order. This resolves a long
983 standing problem: a verify on a PKCS7 structure just after signing
984 it used to fail because the attribute order did not match the
985 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
986 it uses the received order. This is necessary to tolerate some broken
987 software that does not order SET OF. This is handled by encoding
988 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
989 to produce the required SET OF.
992 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
993 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
994 files to get correct declarations of the ASN.1 item variables.
997 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
998 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
999 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1000 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1001 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1002 ASN1_ITEM and no wrapper functions.
1005 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1006 replace the old function pointer based I/O routines. Change most of
1007 the *_d2i_bio() and *_d2i_fp() functions to use these.
1010 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1011 lines, recognice more "algorithms" that can be deselected, and make
1012 it complain about algorithm deselection that isn't recognised.
1015 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1016 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1017 to use new functions. Add NO_ASN1_OLD which can be set to remove
1018 some old style ASN1 functions: this can be used to determine if old
1019 code will still work when these eventually go away.
1022 +) New extension functions for OCSP structures, these follow the
1023 same conventions as certificates and CRLs.
1026 +) New function X509V3_add1_i2d(). This automatically encodes and
1027 adds an extension. Its behaviour can be customised with various
1028 flags to append, replace or delete. Various wrappers added for
1029 certifcates and CRLs.
1032 +) Fix to avoid calling the underlying ASN1 print routine when
1033 an extension cannot be parsed. Correct a typo in the
1034 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1037 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1038 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1039 when writing a 32767 byte record.
1040 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1042 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1043 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1045 (RSA objects have a reference count access to which is protected
1046 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1047 so they are meant to be shared between threads.)
1048 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1049 "Reddie, Steven" <Steven.Reddie@ca.com>]
1051 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1052 entries for variables.
1055 *) Fix a deadlock in CRYPTO_mem_leaks().
1058 +) Add functionality to apps/openssl.c for detecting locking
1059 problems: As the program is single-threaded, all we have
1060 to do is register a locking callback using an array for
1061 storing which locks are currently held by the program.
1064 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1065 SSL_get_ex_data_X509_STORE_idx(), which is used in
1066 ssl_verify_cert_chain() and thus can be called at any time
1067 during TLS/SSL handshakes so that thread-safety is essential.
1068 Unfortunately, the ex_data design is not at all suited
1069 for multi-threaded use, so it probably should be abolished.
1072 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1073 [Broadcom, tweaked and integrated by Geoff Thorpe]
1075 +) Move common extension printing code to new function
1076 X509V3_print_extensions(). Reorganise OCSP print routines and
1077 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1080 +) New function X509_signature_print() to remove duplication in some
1084 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1085 set (this was treated exactly the same as SET OF previously). This
1086 is used to reorder the STACK representing the structure to match the
1087 encoding. This will be used to get round a problem where a PKCS7
1088 structure which was signed could not be verified because the STACK
1089 order did not reflect the encoded order.
1092 +) Reimplement the OCSP ASN1 module using the new code.
1095 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1096 for its ASN1 operations. The old style function pointers still exist
1097 for now but they will eventually go away.
1100 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1101 completely replaces the old ASN1 functionality with a table driven
1102 encoder and decoder which interprets an ASN1_ITEM structure describing
1103 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1104 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1105 has also been converted to the new form.
1108 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1109 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1110 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1111 for negative moduli.
1114 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1115 of not touching the result's sign bit.
1118 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1122 +) Changed the LHASH code to use prototypes for callbacks, and created
1123 macros to declare and implement thin (optionally static) functions
1124 that provide type-safety and avoid function pointer casting for the
1125 type-specific callbacks.
1128 *) Use better test patterns in bntest.
1131 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1133 [Veers Staats <staatsvr@asc.hpc.mil>,
1134 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1136 *) rand_win.c fix for Borland C.
1139 *) BN_rshift bugfix for n == 0.
1142 +) Reformat the FAQ so the different questions and answers can be divided
1143 in sections depending on the subject.
1146 +) Have the zlib compression code load ZLIB.DLL dynamically under
1150 +) New function BN_mod_sqrt for computing square roots modulo a prime
1151 (using the probabilistic Tonelli-Shanks algorithm unless
1152 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1153 be handled deterministically).
1154 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1156 *) Add a 'bctest' script that checks for some known 'bc' bugs
1157 so that 'make test' does not abort just because 'bc' is broken.
1160 *) Store verify_result within SSL_SESSION also for client side to
1161 avoid potential security hole. (Re-used sessions on the client side
1162 always resulted in verify_result==X509_V_OK, not using the original
1163 result of the server certificate verification.)
1166 +) Make BN_mod_inverse faster by explicitly handling small quotients
1167 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1168 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1171 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1172 SSL3_RT_APPLICATION_DATA, return 0.
1173 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1177 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1178 releases, have been re-implemented by renaming the previous
1179 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1180 and ssl3_read_internal, respectively, and adding 'peek' parameters
1181 to them. The new ssl[23]_{read,peek} functions are calls to
1182 ssl[23]_read_internal with the 'peek' flag set appropriately.
1183 A 'peek' parameter has also been added to ssl3_read_bytes, which
1184 does the actual work for ssl3_read_internal.
1187 +) New function BN_kronecker.
1190 +) Fix BN_gcd so that it works on negative inputs; the result is
1191 positive unless both parameters are zero.
1192 Previously something reasonably close to an infinite loop was
1193 possible because numbers could be growing instead of shrinking
1194 in the implementation of Euclid's algorithm.
1197 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1198 sign of the number in question.
1200 Fix BN_is_word(a,w) to work correctly for w == 0.
1202 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1203 because its test if the absolute value of 'a' equals 'w'.
1204 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1205 it exists mostly for use in the implementations of BN_is_zero(),
1206 BN_is_one(), and BN_is_word().
1209 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1210 the method-specific "init()" handler. Also clean up ex_data after
1211 calling the method-specific "finish()" handler. Previously, this was
1212 happening the other way round.
1215 +) New function BN_swap.
1218 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1219 the exponentiation functions are more likely to produce reasonable
1220 results on negative inputs.
1223 +) Change BN_mod_mul so that the result is always non-negative.
1224 Previously, it could be negative if one of the factors was negative;
1225 I don't think anyone really wanted that behaviour.
1228 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1229 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1230 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1231 and add new functions:
1240 BN_mod_lshift1_quick
1244 These functions always generate non-negative results.
1246 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1247 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1249 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1250 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1251 be reduced modulo m.
1252 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1254 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1255 was actually never needed) and in BN_mul(). The removal in BN_mul()
1256 required a small change in bn_mul_part_recursive() and the addition
1257 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1258 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1259 bn_sub_words() and bn_add_words() except they take arrays with
1263 +) In 'openssl passwd', verify passwords read from the terminal
1264 unless the '-salt' option is used (which usually means that
1265 verification would just waste user's time since the resulting
1266 hash is going to be compared with some given password hash)
1267 or the new '-noverify' option is used.
1269 This is an incompatible change, but it does not affect
1270 non-interactive use of 'openssl passwd' (passwords on the command
1271 line, '-stdin' option, '-in ...' option) and thus should not
1275 +) Remove all references to RSAref, since there's no more need for it.
1278 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1279 The previous value, 12, was not always sufficient for BN_mod_exp().
1282 +) Make DSO load along a path given through an environment variable
1283 (SHLIB_PATH) with shl_load().
1286 +) Constify the ENGINE code as a result of BIGNUM constification.
1287 Also constify the RSA code and most things related to it. In a
1288 few places, most notable in the depth of the ASN.1 code, ugly
1289 casts back to non-const were required (to be solved at a later
1293 +) Make it so the openssl application has all engines loaded by default.
1296 +) Constify the BIGNUM routines a little more.
1299 *) Make sure that shared libraries get the internal name engine with
1300 the full version number and not just 0. This should mark the
1301 shared libraries as not backward compatible. Of course, this should
1302 be changed again when we can guarantee backward binary compatibility.
1305 +) Add the following functions:
1307 ENGINE_load_cswift()
1309 ENGINE_load_atalla()
1311 ENGINE_load_builtin_engines()
1313 That way, an application can itself choose if external engines that
1314 are built-in in OpenSSL shall ever be used or not. The benefit is
1315 that applications won't have to be linked with libdl or other dso
1316 libraries unless it's really needed.
1318 Changed 'openssl engine' to load all engines on demand.
1319 Changed the engine header files to avoid the duplication of some
1320 declarations (they differed!).
1323 +) 'openssl engine' can now list capabilities.
1326 +) Better error reporting in 'openssl engine'.
1329 +) Never call load_dh_param(NULL) in s_server.
1332 +) Add engine application. It can currently list engines by name and
1333 identity, and test if they are actually available.
1336 +) Add support for shared libraries under Irix.
1337 [Albert Chin-A-Young <china@thewrittenword.com>]
1339 +) Improve RPM specification file by forcing symbolic linking and making
1340 sure the installed documentation is also owned by root.root.
1341 [Damien Miller <djm@mindrot.org>]
1343 +) Add configuration option to build on Linux on both big-endian and
1345 [Ralf Baechle <ralf@uni-koblenz.de>]
1347 +) Give the OpenSSL applications more possibilities to make use of
1348 keys (public as well as private) handled by engines.
1351 +) Add OCSP code that comes from CertCo.
1354 +) Add VMS support for the Rijndael code.
1357 +) Added untested support for Nuron crypto accelerator.
1360 +) Add support for external cryptographic devices. This code was
1361 previously distributed separately as the "engine" branch.
1362 [Geoff Thorpe, Richard Levitte]
1364 +) Rework the filename-translation in the DSO code. It is now possible to
1365 have far greater control over how a "name" is turned into a filename
1366 depending on the operating environment and any oddities about the
1367 different shared library filenames on each system.
1370 +) Support threads on FreeBSD-elf in Configure.
1373 +) Add the possibility to create shared libraries on HP-UX
1376 +) Fix for SHA1 assembly problem with MASM: it produces
1377 warnings about corrupt line number information when assembling
1378 with debugging information. This is caused by the overlapping
1380 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1383 NCONF_get_number() has no error checking at all. As a replacement,
1384 NCONF_get_number_e() is defined (_e for "error checking") and is
1385 promoted strongly. The old NCONF_get_number is kept around for
1386 binary backward compatibility.
1387 Make it possible for methods to load from something other than a BIO,
1388 by providing a function pointer that is given a name instead of a BIO.
1389 For example, this could be used to load configuration data from an
1393 *) Fix typo in get_cert_by_subject() in by_dir.c
1394 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1396 *) Rework the system to generate shared libraries:
1398 - Make note of the expected extension for the shared libraries and
1399 if there is a need for symbolic links from for example libcrypto.so.0
1400 to libcrypto.so.0.9.7. There is extended info in Configure for
1403 - Make as few rebuilds of the shared libraries as possible.
1405 - Still avoid linking the OpenSSL programs with the shared libraries.
1407 - When installing, install the shared libraries separately from the
1411 +) Fix for non blocking accept BIOs. Added new I/O special reason
1412 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1413 with non blocking I/O was not possible because no retry code was
1414 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1418 +) Added the beginnings of Rijndael support.
1421 +) Fix for bug in DirectoryString mask setting. Add support for
1422 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1423 to allow certificate printing to more controllable, additional
1424 'certopt' option to 'x509' to allow new printing options to be
1428 +) Clean old EAY MD5 hack from e_os.h.
1431 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1433 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1434 and not in SSL_clear because the latter is also used by the
1435 accept/connect functions; previously, the settings made by
1436 SSL_set_read_ahead would be lost during the handshake.
1437 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1439 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1440 Previously, it would create entries for disableed algorithms no
1444 *) Added several new manual pages for SSL_* function.
1447 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1449 *) In ssl23_get_client_hello, generate an error message when faced
1450 with an initial SSL 3.0/TLS record that is too small to contain the
1451 first two bytes of the ClientHello message, i.e. client_version.
1452 (Note that this is a pathologic case that probably has never happened
1453 in real life.) The previous approach was to use the version number
1454 from the record header as a substitute; but our protocol choice
1455 should not depend on that one because it is not authenticated
1456 by the Finished messages.
1459 *) More robust randomness gathering functions for Windows.
1460 [Jeffrey Altman <jaltman@columbia.edu>]
1462 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1463 not set then we don't setup the error code for issuer check errors
1464 to avoid possibly overwriting other errors which the callback does
1465 handle. If an application does set the flag then we assume it knows
1466 what it is doing and can handle the new informational codes
1470 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1471 a general "ANY" type, as such it should be able to decode anything
1472 including tagged types. However it didn't check the class so it would
1473 wrongly interpret tagged types in the same way as their universal
1474 counterpart and unknown types were just rejected. Changed so that the
1475 tagged and unknown types are handled in the same way as a SEQUENCE:
1476 that is the encoding is stored intact. There is also a new type
1477 "V_ASN1_OTHER" which is used when the class is not universal, in this
1478 case we have no idea what the actual type is so we just lump them all
1482 *) On VMS, stdout may very well lead to a file that is written to
1483 in a record-oriented fashion. That means that every write() will
1484 write a separate record, which will be read separately by the
1485 programs trying to read from it. This can be very confusing.
1487 The solution is to put a BIO filter in the way that will buffer
1488 text until a linefeed is reached, and then write everything a
1489 line at a time, so every record written will be an actual line,
1490 not chunks of lines and not (usually doesn't happen, but I've
1491 seen it once) several lines in one record. BIO_f_linebuffer() is
1494 Currently, it's a VMS-only method, because that's where it has
1495 been tested well enough.
1498 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1499 it can return incorrect results.
1500 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1501 but it was in 0.9.6-beta[12].)
1504 *) Disable the check for content being present when verifying detached
1505 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1506 include zero length content when signing messages.
1509 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1510 BIO_ctrl (for BIO pairs).
1513 *) Add DSO method for VMS.
1516 *) Bug fix: Montgomery multiplication could produce results with the
1520 *) Add RPM specification openssl.spec and modify it to build three
1521 packages. The default package contains applications, application
1522 documentation and run-time libraries. The devel package contains
1523 include files, static libraries and function documentation. The
1524 doc package contains the contents of the doc directory. The original
1525 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1528 *) Add a large number of documentation files for many SSL routines.
1529 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1531 *) Add a configuration entry for Sony News 4.
1532 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1534 *) Don't set the two most significant bits to one when generating a
1535 random number < q in the DSA library.
1538 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1539 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1540 the underlying transport is blocking) if a handshake took place.
1541 (The default behaviour is needed by applications such as s_client
1542 and s_server that use select() to determine when to use SSL_read;
1543 but for applications that know in advance when to expect data, it
1544 just makes things more complicated.)
1547 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1551 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1552 work better on such systems.
1553 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1555 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1556 Update PKCS12_parse() so it copies the friendlyName and the
1557 keyid to the certificates aux info.
1560 *) Fix bug in PKCS7_verify() which caused an infinite loop
1561 if there was more than one signature.
1562 [Sven Uszpelkat <su@celocom.de>]
1564 *) Major change in util/mkdef.pl to include extra information
1565 about each symbol, as well as presentig variables as well
1566 as functions. This change means that there's n more need
1567 to rebuild the .num files when some algorithms are excluded.
1570 *) Allow the verify time to be set by an application,
1571 rather than always using the current time.
1574 *) Phase 2 verify code reorganisation. The certificate
1575 verify code now looks up an issuer certificate by a
1576 number of criteria: subject name, authority key id
1577 and key usage. It also verifies self signed certificates
1578 by the same criteria. The main comparison function is
1579 X509_check_issued() which performs these checks.
1581 Lot of changes were necessary in order to support this
1582 without completely rewriting the lookup code.
1584 Authority and subject key identifier are now cached.
1586 The LHASH 'certs' is X509_STORE has now been replaced
1587 by a STACK_OF(X509_OBJECT). This is mainly because an
1588 LHASH can't store or retrieve multiple objects with
1589 the same hash value.
1591 As a result various functions (which were all internal
1592 use only) have changed to handle the new X509_STORE
1593 structure. This will break anything that messed round
1594 with X509_STORE internally.
1596 The functions X509_STORE_add_cert() now checks for an
1597 exact match, rather than just subject name.
1599 The X509_STORE API doesn't directly support the retrieval
1600 of multiple certificates matching a given criteria, however
1601 this can be worked round by performing a lookup first
1602 (which will fill the cache with candidate certificates)
1603 and then examining the cache for matches. This is probably
1604 the best we can do without throwing out X509_LOOKUP
1605 entirely (maybe later...).
1607 The X509_VERIFY_CTX structure has been enhanced considerably.
1609 All certificate lookup operations now go via a get_issuer()
1610 callback. Although this currently uses an X509_STORE it
1611 can be replaced by custom lookups. This is a simple way
1612 to bypass the X509_STORE hackery necessary to make this
1613 work and makes it possible to use more efficient techniques
1614 in future. A very simple version which uses a simple
1615 STACK for its trusted certificate store is also provided
1616 using X509_STORE_CTX_trusted_stack().
1618 The verify_cb() and verify() callbacks now have equivalents
1619 in the X509_STORE_CTX structure.
1621 X509_STORE_CTX also has a 'flags' field which can be used
1622 to customise the verify behaviour.
1625 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1626 excludes S/MIME capabilities.
1629 *) When a certificate request is read in keep a copy of the
1630 original encoding of the signed data and use it when outputing
1631 again. Signatures then use the original encoding rather than
1632 a decoded, encoded version which may cause problems if the
1633 request is improperly encoded.
1636 *) For consistency with other BIO_puts implementations, call
1637 buffer_write(b, ...) directly in buffer_puts instead of calling
1640 In BIO_puts, increment b->num_write as in BIO_write.
1641 [Peter.Sylvester@EdelWeb.fr]
1643 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1644 BN_zero, we may not return a BIGNUM with an array consisting of
1648 *) Avoid calling abort() from within the library when problems are
1649 detected, except if preprocessor symbols have been defined
1650 (such as REF_CHECK, BN_DEBUG etc.).
1653 *) New openssl application 'rsautl'. This utility can be
1654 used for low level RSA operations. DER public key
1655 BIO/fp routines also added.
1658 *) New Configure entry and patches for compiling on QNX 4.
1659 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1661 *) A demo state-machine implementation was sponsored by
1662 Nuron (http://www.nuron.com/) and is now available in
1663 demos/state_machine.
1666 *) New options added to the 'dgst' utility for signature
1667 generation and verification.
1670 *) Unrecognized PKCS#7 content types are now handled via a
1671 catch all ASN1_TYPE structure. This allows unsupported
1672 types to be stored as a "blob" and an application can
1673 encode and decode it manually.
1676 *) Fix various signed/unsigned issues to make a_strex.c
1678 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1680 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1681 length if passed a buffer. ASN1_INTEGER_to_BN failed
1682 if passed a NULL BN and its argument was negative.
1683 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1685 *) Modification to PKCS#7 encoding routines to output definite
1686 length encoding. Since currently the whole structures are in
1687 memory there's not real point in using indefinite length
1688 constructed encoding. However if OpenSSL is compiled with
1689 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1692 *) Added BIO_vprintf() and BIO_vsnprintf().
1695 *) Added more prefixes to parse for in the the strings written
1696 through a logging bio, to cover all the levels that are available
1697 through syslog. The prefixes are now:
1699 PANIC, EMERG, EMR => LOG_EMERG
1700 ALERT, ALR => LOG_ALERT
1701 CRIT, CRI => LOG_CRIT
1702 ERROR, ERR => LOG_ERR
1703 WARNING, WARN, WAR => LOG_WARNING
1704 NOTICE, NOTE, NOT => LOG_NOTICE
1705 INFO, INF => LOG_INFO
1706 DEBUG, DBG => LOG_DEBUG
1708 and as before, if none of those prefixes are present at the
1709 beginning of the string, LOG_ERR is chosen.
1711 On Win32, the LOG_* levels are mapped according to this:
1713 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1714 LOG_WARNING => EVENTLOG_WARNING_TYPE
1715 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1719 *) Made it possible to reconfigure with just the configuration
1720 argument "reconf" or "reconfigure". The command line arguments
1721 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1722 and are retrieved from there when reconfiguring.
1726 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1728 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1731 *) The obj_dat.pl script was messing up the sorting of object
1732 names. The reason was that it compared the quoted version
1733 of strings as a result "OCSP" > "OCSP Signing" because
1734 " > SPACE. Changed script to store unquoted versions of
1735 names and add quotes on output. It was also omitting some
1736 names from the lookup table if they were given a default
1737 value (that is if SN is missing it is given the same
1738 value as LN and vice versa), these are now added on the
1739 grounds that if an object has a name we should be able to
1740 look it up. Finally added warning output when duplicate
1741 short or long names are found.
1744 *) Changes needed for Tandem NSK.
1745 [Scott Uroff <scott@xypro.com>]
1747 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1748 RSA_padding_check_SSLv23(), special padding was never detected
1749 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1750 version rollback attacks was not effective.
1752 In s23_clnt.c, don't use special rollback-attack detection padding
1753 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1754 client; similarly, in s23_srvr.c, don't do the rollback check if
1755 SSL 2.0 is the only protocol enabled in the server.
1758 *) Make it possible to get hexdumps of unprintable data with 'openssl
1759 asn1parse'. By implication, the functions ASN1_parse_dump() and
1760 BIO_dump_indent() are added.
1763 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1764 these print out strings and name structures based on various
1765 flags including RFC2253 support and proper handling of
1766 multibyte characters. Added options to the 'x509' utility
1767 to allow the various flags to be set.
1770 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1771 Also change the functions X509_cmp_current_time() and
1772 X509_gmtime_adj() work with an ASN1_TIME structure,
1773 this will enable certificates using GeneralizedTime in validity
1774 dates to be checked.
1777 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1778 negative public key encodings) on by default,
1779 NO_NEG_PUBKEY_BUG can be set to disable it.
1782 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1783 content octets. An i2c_ASN1_OBJECT is unnecessary because
1784 the encoding can be trivially obtained from the structure.
1787 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1788 not read locks (CRYPTO_r_[un]lock).
1791 *) A first attempt at creating official support for shared
1792 libraries through configuration. I've kept it so the
1793 default is static libraries only, and the OpenSSL programs
1794 are always statically linked for now, but there are
1795 preparations for dynamic linking in place.
1796 This has been tested on Linux and True64.
1799 *) Randomness polling function for Win9x, as described in:
1800 Peter Gutmann, Software Generation of Practically Strong
1804 *) Fix so PRNG is seeded in req if using an already existing
1808 *) New options to smime application. -inform and -outform
1809 allow alternative formats for the S/MIME message including
1810 PEM and DER. The -content option allows the content to be
1811 specified separately. This should allow things like Netscape
1812 form signing output easier to verify.
1815 *) Fix the ASN1 encoding of tags using the 'long form'.
1818 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1819 STRING types. These convert content octets to and from the
1820 underlying type. The actual tag and length octets are
1821 already assumed to have been read in and checked. These
1822 are needed because all other string types have virtually
1823 identical handling apart from the tag. By having versions
1824 of the ASN1 functions that just operate on content octets
1825 IMPLICIT tagging can be handled properly. It also allows
1826 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1827 and ASN1_INTEGER are identical apart from the tag.
1830 *) Change the handling of OID objects as follows:
1832 - New object identifiers are inserted in objects.txt, following
1833 the syntax given in objects.README.
1834 - objects.pl is used to process obj_mac.num and create a new
1836 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1839 This is currently kind of a hack, and the perl code in objects.pl
1840 isn't very elegant, but it works as I intended. The simplest way
1841 to check that it worked correctly is to look in obj_dat.h and
1842 check the array nid_objs and make sure the objects haven't moved
1843 around (this is important!). Additions are OK, as well as
1844 consistent name changes.
1847 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1850 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1851 The given file adds to whatever has already been seeded into the
1852 random pool through the RANDFILE configuration file option or
1853 environment variable, or the default random state file.
1856 *) mkstack.pl now sorts each macro group into lexical order.
1857 Previously the output order depended on the order the files
1858 appeared in the directory, resulting in needless rewriting
1862 *) Patches to make OpenSSL compile under Win32 again. Mostly
1863 work arounds for the VC++ problem that it treats func() as
1864 func(void). Also stripped out the parts of mkdef.pl that
1865 added extra typesafe functions: these no longer exist.
1868 *) Reorganisation of the stack code. The macros are now all
1869 collected in safestack.h . Each macro is defined in terms of
1870 a "stack macro" of the form SKM_<name>(type, a, b). The
1871 DEBUG_SAFESTACK is now handled in terms of function casts,
1872 this has the advantage of retaining type safety without the
1873 use of additional functions. If DEBUG_SAFESTACK is not defined
1874 then the non typesafe macros are used instead. Also modified the
1875 mkstack.pl script to handle the new form. Needs testing to see
1876 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1877 the default if no major problems. Similar behaviour for ASN1_SET_OF
1878 and PKCS12_STACK_OF.
1881 *) When some versions of IIS use the 'NET' form of private key the
1882 key derivation algorithm is different. Normally MD5(password) is
1883 used as a 128 bit RC4 key. In the modified case
1884 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1885 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1886 as the old Netscape_RSA functions except they have an additional
1887 'sgckey' parameter which uses the modified algorithm. Also added
1888 an -sgckey command line option to the rsa utility. Thanks to
1889 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1890 algorithm to openssl-dev.
1893 *) The evp_local.h macros were using 'c.##kname' which resulted in
1894 invalid expansion on some systems (SCO 5.0.5 for example).
1895 Corrected to 'c.kname'.
1896 [Phillip Porch <root@theporch.com>]
1898 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1899 a STACK of email addresses from a certificate or request, these look
1900 in the subject name and the subject alternative name extensions and
1901 omit any duplicate addresses.
1904 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1905 This makes DSA verification about 2 % faster.
1908 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1909 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1910 plus overhead for 1024 bit moduli).
1911 This makes exponentiations about 0.5 % faster for 1024 bit
1912 exponents (as measured by "openssl speed rsa2048").
1915 *) Rename memory handling macros to avoid conflicts with other
1917 Malloc => OPENSSL_malloc
1918 Malloc_locked => OPENSSL_malloc_locked
1919 Realloc => OPENSSL_realloc
1920 Free => OPENSSL_free
1923 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1924 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1927 *) CygWin32 support.
1928 [John Jarvie <jjarvie@newsguy.com>]
1930 *) The type-safe stack code has been rejigged. It is now only compiled
1931 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1932 by default all type-specific stack functions are "#define"d back to
1933 standard stack functions. This results in more streamlined output
1934 but retains the type-safety checking possibilities of the original
1938 *) The STACK code has been cleaned up, and certain type declarations
1939 that didn't make a lot of sense have been brought in line. This has
1940 also involved a cleanup of sorts in safestack.h to more correctly
1941 map type-safe stack functions onto their plain stack counterparts.
1942 This work has also resulted in a variety of "const"ifications of
1943 lots of the code, especially "_cmp" operations which should normally
1944 be prototyped with "const" parameters anyway.
1947 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1948 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1949 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1950 where all of 'md' is used each time the PRNG is used, but 'state'
1951 is used only indexed by a cyclic counter. As entropy may not be
1952 well distributed from the beginning, 'md' is important as a
1953 chaining variable. However, the output function chains only half
1954 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1955 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1956 in all of 'state' being rewritten, with the new values depending
1957 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1960 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1961 the handshake is continued after ssl_verify_cert_chain();
1962 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1963 can lead to 'unexplainable' connection aborts later.
1964 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1966 *) Major EVP API cipher revision.
1967 Add hooks for extra EVP features. This allows various cipher
1968 parameters to be set in the EVP interface. Support added for variable
1969 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1970 setting of RC2 and RC5 parameters.
1972 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1975 Remove lots of duplicated code from the EVP library. For example *every*
1976 cipher init() function handles the 'iv' in the same way according to the
1977 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1978 for CFB and OFB modes they zero ctx->num.
1980 New functionality allows removal of S/MIME code RC2 hack.
1982 Most of the routines have the same form and so can be declared in terms
1985 By shifting this to the top level EVP_CipherInit() it can be removed from
1986 all individual ciphers. If the cipher wants to handle IVs or keys
1987 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1990 Change lots of functions like EVP_EncryptUpdate() to now return a
1991 value: although software versions of the algorithms cannot fail
1992 any installed hardware versions can.
1995 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1996 this option is set, tolerate broken clients that send the negotiated
1997 protocol version number instead of the requested protocol version
2001 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2002 i.e. non-zero for export ciphersuites, zero otherwise.
2003 Previous versions had this flag inverted, inconsistent with
2004 rsa_tmp_cb (..._TMP_RSA_CB).
2005 [Bodo Moeller; problem reported by Amit Chopra]
2007 *) Add missing DSA library text string. Work around for some IIS
2008 key files with invalid SEQUENCE encoding.
2011 *) Add a document (doc/standards.txt) that list all kinds of standards
2012 and so on that are implemented in OpenSSL.
2015 *) Enhance c_rehash script. Old version would mishandle certificates
2016 with the same subject name hash and wouldn't handle CRLs at all.
2017 Added -fingerprint option to crl utility, to support new c_rehash
2021 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2024 *) Fix for SSL server purpose checking. Server checking was
2025 rejecting certificates which had extended key usage present
2026 but no ssl client purpose.
2027 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2029 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2030 is a little unclear about how a blank password is handled.
2031 Since the password in encoded as a BMPString with terminating
2032 double NULL a zero length password would end up as just the
2033 double NULL. However no password at all is different and is
2034 handled differently in the PKCS#12 key generation code. NS
2035 treats a blank password as zero length. MSIE treats it as no
2036 password on export: but it will try both on import. We now do
2037 the same: PKCS12_parse() tries zero length and no password if
2038 the password is set to "" or NULL (NULL is now a valid password:
2039 it wasn't before) as does the pkcs12 application.
2042 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2043 perror when PEM_read_bio_X509_REQ fails, the error message must
2044 be obtained from the error queue.
2047 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2048 it in ERR_remove_state if appropriate, and change ERR_get_state
2049 accordingly to avoid race conditions (this is necessary because
2050 thread_hash is no longer constant once set).
2053 *) Bugfix for linux-elf makefile.one.
2056 *) RSA_get_default_method() will now cause a default
2057 RSA_METHOD to be chosen if one doesn't exist already.
2058 Previously this was only set during a call to RSA_new()
2059 or RSA_new_method(NULL) meaning it was possible for
2060 RSA_get_default_method() to return NULL.
2063 *) Added native name translation to the existing DSO code
2064 that will convert (if the flag to do so is set) filenames
2065 that are sufficiently small and have no path information
2066 into a canonical native form. Eg. "blah" converted to
2067 "libblah.so" or "blah.dll" etc.
2070 *) New function ERR_error_string_n(e, buf, len) which is like
2071 ERR_error_string(e, buf), but writes at most 'len' bytes
2072 including the 0 terminator. For ERR_error_string_n, 'buf'
2074 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2076 *) CONF library reworked to become more general. A new CONF
2077 configuration file reader "class" is implemented as well as a
2078 new functions (NCONF_*, for "New CONF") to handle it. The now
2079 old CONF_* functions are still there, but are reimplemented to
2080 work in terms of the new functions. Also, a set of functions
2081 to handle the internal storage of the configuration data is
2082 provided to make it easier to write new configuration file
2083 reader "classes" (I can definitely see something reading a
2084 configuration file in XML format, for example), called _CONF_*,
2085 or "the configuration storage API"...
2087 The new configuration file reading functions are:
2089 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2090 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2092 NCONF_default, NCONF_WIN32
2094 NCONF_dump_fp, NCONF_dump_bio
2096 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2097 NCONF_new creates a new CONF object. This works in the same way
2098 as other interfaces in OpenSSL, like the BIO interface.
2099 NCONF_dump_* dump the internal storage of the configuration file,
2100 which is useful for debugging. All other functions take the same
2101 arguments as the old CONF_* functions wth the exception of the
2102 first that must be a `CONF *' instead of a `LHASH *'.
2104 To make it easer to use the new classes with the old CONF_* functions,
2105 the function CONF_set_default_method is provided.
2108 *) Add '-tls1' option to 'openssl ciphers', which was already
2109 mentioned in the documentation but had not been implemented.
2110 (This option is not yet really useful because even the additional
2111 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2114 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2115 OpenSSL-based applications) load shared libraries and bind to
2116 them in a portable way.
2117 [Geoff Thorpe, with contributions from Richard Levitte]
2119 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2121 *) Make sure _lrotl and _lrotr are only used with MSVC.
2123 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2124 (the default implementation of RAND_status).
2126 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2127 to '-clrext' (= clear extensions), as intended and documented.
2128 [Bodo Moeller; inconsistency pointed out by Michael Attili
2131 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2132 was larger than the MD block size.
2133 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2135 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2136 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2137 using the passed key: if the passed key was a private key the result
2138 of X509_print(), for example, would be to print out all the private key
2142 *) des_quad_cksum() byte order bug fix.
2143 [Ulf Möller, using the problem description in krb4-0.9.7, where
2144 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2146 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2148 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2150 *) For easily testing in shell scripts whether some command
2151 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2152 returns with exit code 0 iff no command of the given name is available.
2153 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2154 the output goes to stdout and nothing is printed to stderr.
2155 Additional arguments are always ignored.
2157 Since for each cipher there is a command of the same name,
2158 the 'no-cipher' compilation switches can be tested this way.
2160 ('openssl no-XXX' is not able to detect pseudo-commands such
2161 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2164 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2167 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2168 is set; it will be thrown away anyway because each handshake creates
2170 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2171 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2172 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2173 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2176 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2177 'Q' and 'R' lose their special meanings (quit/renegotiate).
2178 This is part of what -quiet does; unlike -quiet, -ign_eof
2179 does not suppress any output.
2182 *) Add compatibility options to the purpose and trust code. The
2183 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2184 accepts a certificate or CA, this was the previous behaviour,
2185 with all the associated security issues.
2187 X509_TRUST_COMPAT is the old trust behaviour: only and
2188 automatically trust self signed roots in certificate store. A
2189 new trust setting X509_TRUST_DEFAULT is used to specify that
2190 a purpose has no associated trust setting and it should instead
2191 use the value in the default purpose.
2194 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2195 and fix a memory leak.
2198 *) In util/mkerr.pl (which implements 'make errors'), preserve
2199 reason strings from the previous version of the .c file, as
2200 the default to have only downcase letters (and digits) in
2201 automatically generated reasons codes is not always appropriate.
2204 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2205 using strerror. Previously, ERR_reason_error_string() returned
2206 library names as reason strings for SYSerr; but SYSerr is a special
2207 case where small numbers are errno values, not library numbers.
2210 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2211 converts DSA parameters into DH parameters. (When creating parameters,
2212 DSA_generate_parameters is used.)
2215 *) Include 'length' (recommended exponent length) in C code generated
2216 by 'openssl dhparam -C'.
2219 *) The second argument to set_label in perlasm was already being used
2220 so couldn't be used as a "file scope" flag. Moved to third argument
2224 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2225 instead of RAND_bytes for encryption IVs and salts.
2228 *) Include RAND_status() into RAND_METHOD instead of implementing
2229 it only for md_rand.c Otherwise replacing the PRNG by calling
2230 RAND_set_rand_method would be impossible.
2233 *) Don't let DSA_generate_key() enter an infinite loop if the random
2234 number generation fails.
2237 *) New 'rand' application for creating pseudo-random output.
2240 *) Added configuration support for Linux/IA64
2241 [Rolf Haberrecker <rolf@suse.de>]
2243 *) Assembler module support for Mingw32.
2246 *) Shared library support for HPUX (in shlib/).
2247 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2249 *) Shared library support for Solaris gcc.
2250 [Lutz Behnke <behnke@trustcenter.de>]
2252 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2254 *) PKCS7_encrypt() was adding text MIME headers twice because they
2255 were added manually and by SMIME_crlf_copy().
2258 *) In bntest.c don't call BN_rand with zero bits argument.
2259 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2261 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2262 case was implemented. This caused BN_div_recp() to fail occasionally.
2265 *) Add an optional second argument to the set_label() in the perl
2266 assembly language builder. If this argument exists and is set
2267 to 1 it signals that the assembler should use a symbol whose
2268 scope is the entire file, not just the current function. This
2269 is needed with MASM which uses the format label:: for this scope.
2270 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2272 *) Change the ASN1 types so they are typedefs by default. Before
2273 almost all types were #define'd to ASN1_STRING which was causing
2274 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2278 *) Change names of new functions to the new get1/get0 naming
2279 convention: After 'get1', the caller owns a reference count
2280 and has to call ..._free; 'get0' returns a pointer to some
2281 data structure without incrementing reference counters.
2282 (Some of the existing 'get' functions increment a reference
2283 counter, some don't.)
2284 Similarly, 'set1' and 'add1' functions increase reference
2285 counters or duplicate objects.
2288 *) Allow for the possibility of temp RSA key generation failure:
2289 the code used to assume it always worked and crashed on failure.
2292 *) Fix potential buffer overrun problem in BIO_printf().
2293 [Ulf Möller, using public domain code by Patrick Powell; problem
2294 pointed out by David Sacerdote <das33@cornell.edu>]
2296 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2297 RAND_egd() and RAND_status(). In the command line application,
2298 the EGD socket can be specified like a seed file using RANDFILE
2302 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2303 Some CAs (e.g. Verisign) distribute certificates in this form.
2306 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2307 list to exclude them. This means that no special compilation option
2308 is needed to use anonymous DH: it just needs to be included in the
2312 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2313 EVP_MD_type. The old functionality is available in a new macro called
2314 EVP_MD_md(). Change code that uses it and update docs.
2317 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2318 where the 'void *' argument is replaced by a function pointer argument.
2319 Previously 'void *' was abused to point to functions, which works on
2320 many platforms, but is not correct. As these functions are usually
2321 called by macros defined in OpenSSL header files, most source code
2322 should work without changes.
2325 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2326 sections with information on -D... compiler switches used for
2327 compiling the library so that applications can see them. To enable
2328 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2329 must be defined. E.g.,
2330 #define OPENSSL_ALGORITHM_DEFINES
2331 #include <openssl/opensslconf.h>
2332 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2333 [Richard Levitte, Ulf and Bodo Möller]
2335 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2339 *) Change the 'other' type in certificate aux info to a STACK_OF
2340 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2341 the required ASN1 format: arbitrary types determined by an OID.
2344 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2345 argument to 'req'. This is not because the function is newer or
2346 better than others it just uses the work 'NEW' in the certificate
2347 request header lines. Some software needs this.
2350 *) Reorganise password command line arguments: now passwords can be
2351 obtained from various sources. Delete the PEM_cb function and make
2352 it the default behaviour: i.e. if the callback is NULL and the
2353 usrdata argument is not NULL interpret it as a null terminated pass
2354 phrase. If usrdata and the callback are NULL then the pass phrase
2355 is prompted for as usual.
2358 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2359 the support is automatically enabled. The resulting binaries will
2360 autodetect the card and use it if present.
2361 [Ben Laurie and Compaq Inc.]
2363 *) Work around for Netscape hang bug. This sends certificate request
2364 and server done in one record. Since this is perfectly legal in the
2365 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2366 the bugs/SSLv3 entry for more info.
2369 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2372 *) Add -rand argument to smime and pkcs12 applications and read/write
2376 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2379 *) Add command line password options to the remaining applications.
2382 *) Bug fix for BN_div_recp() for numerators with an even number of
2386 *) More tests in bntest.c, and changed test_bn output.
2389 *) ./config recognizes MacOS X now.
2392 *) Bug fix for BN_div() when the first words of num and divsor are
2393 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2396 *) Add support for various broken PKCS#8 formats, and command line
2397 options to produce them.
2400 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2401 get temporary BIGNUMs from a BN_CTX.
2404 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2408 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2409 include a #define from the old name to the new. The original intent
2410 was that statically linked binaries could for example just call
2411 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2412 link with digests. This never worked becayse SSLeay_add_all_digests()
2413 and SSLeay_add_all_ciphers() were in the same source file so calling
2414 one would link with the other. They are now in separate source files.
2417 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2420 *) Use a less unusual form of the Miller-Rabin primality test (it used
2421 a binary algorithm for exponentiation integrated into the Miller-Rabin
2422 loop, our standard modexp algorithms are faster).
2425 *) Support for the EBCDIC character set completed.
2426 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2428 *) Source code cleanups: use const where appropriate, eliminate casts,
2429 use void * instead of char * in lhash.
2432 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2433 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2434 this the server could overwrite ephemeral keys that the client
2438 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2439 using 50 iterations of the Rabin-Miller test.
2441 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2442 iterations of the Rabin-Miller test as required by the appendix
2443 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2444 As BN_is_prime_fasttest includes trial division, DSA parameter
2445 generation becomes much faster.
2447 This implies a change for the callback functions in DSA_is_prime
2448 and DSA_generate_parameters: The callback function is called once
2449 for each positive witness in the Rabin-Miller test, not just
2450 occasionally in the inner loop; and the parameters to the
2451 callback function now provide an iteration count for the outer
2452 loop rather than for the current invocation of the inner loop.
2453 DSA_generate_parameters additionally can call the callback
2454 function with an 'iteration count' of -1, meaning that a
2455 candidate has passed the trial division test (when q is generated
2456 from an application-provided seed, trial division is skipped).
2459 *) New function BN_is_prime_fasttest that optionally does trial
2460 division before starting the Rabin-Miller test and has
2461 an additional BN_CTX * argument (whereas BN_is_prime always
2462 has to allocate at least one BN_CTX).
2463 'callback(1, -1, cb_arg)' is called when a number has passed the
2464 trial division stage.
2467 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2471 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2474 *) New function BN_pseudo_rand().
2477 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2478 bignum version of BN_from_montgomery() with the working code from
2479 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2483 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2484 made it impossible to use the same SSL_SESSION data structure in
2485 SSL2 clients in multiple threads.
2488 *) The return value of RAND_load_file() no longer counts bytes obtained
2489 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2490 to seed the PRNG (previously an explicit byte count was required).
2491 [Ulf Möller, Bodo Möller]
2493 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2494 used (char *) instead of (void *) and had casts all over the place.
2497 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2500 *) Retain source code compatibility for BN_prime_checks macro:
2501 BN_is_prime(..., BN_prime_checks, ...) now uses
2502 BN_prime_checks_for_size to determine the appropriate number of
2503 Rabin-Miller iterations.
2506 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2507 DH_CHECK_P_NOT_SAFE_PRIME.
2508 (Check if this is true? OpenPGP calls them "strong".)
2511 *) Merge the functionality of "dh" and "gendh" programs into a new program
2512 "dhparam". The old programs are retained for now but will handle DH keys
2513 (instead of parameters) in future.
2516 *) Make the ciphers, s_server and s_client programs check the return values
2517 when a new cipher list is set.
2520 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2521 ciphers. Before when the 56bit ciphers were enabled the sorting was
2524 The syntax for the cipher sorting has been extended to support sorting by
2525 cipher-strength (using the strength_bits hard coded in the tables).
2526 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2528 Fix a bug in the cipher-command parser: when supplying a cipher command
2529 string with an "undefined" symbol (neither command nor alphanumeric
2530 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2531 an error is flagged.
2533 Due to the strength-sorting extension, the code of the
2534 ssl_create_cipher_list() function was completely rearranged. I hope that
2535 the readability was also increased :-)
2536 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2538 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2539 for the first serial number and places 2 in the serial number file. This
2540 avoids problems when the root CA is created with serial number zero and
2541 the first user certificate has the same issuer name and serial number
2545 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2546 the new code. Add documentation for this stuff.
2549 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2550 X509_*() to X509at_*() on the grounds that they don't handle X509
2551 structures and behave in an analagous way to the X509v3 functions:
2552 they shouldn't be called directly but wrapper functions should be used
2555 So we also now have some wrapper functions that call the X509at functions
2556 when passed certificate requests. (TO DO: similar things can be done with
2557 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2558 things. Some of these need some d2i or i2d and print functionality
2559 because they handle more complex structures.)
2562 *) Add missing #ifndefs that caused missing symbols when building libssl
2563 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2564 NO_RSA in ssl/s2*.c.
2565 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2567 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2568 has a return value which indicates the quality of the random data
2569 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2570 error queue. New function RAND_pseudo_bytes() generates output that is
2571 guaranteed to be unique but not unpredictable. RAND_add is like
2572 RAND_seed, but takes an extra argument for an entropy estimate
2573 (RAND_seed always assumes full entropy).
2576 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2577 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2578 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2579 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2580 false-positive rate of at most 2^-80 for random input.
2583 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2586 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2587 in the 0.9.5 release), this returns the chain
2588 from an X509_CTX structure with a dup of the stack and all
2589 the X509 reference counts upped: so the stack will exist
2590 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2593 Also make SSL_SESSION_print() print out the verify return
2597 *) Add manpage for the pkcs12 command. Also change the default
2598 behaviour so MAC iteration counts are used unless the new
2599 -nomaciter option is used. This improves file security and
2600 only older versions of MSIE (4.0 for example) need it.
2603 *) Honor the no-xxx Configure options when creating .DEF files.
2606 *) Add PKCS#10 attributes to field table: challengePassword,
2607 unstructuredName and unstructuredAddress. These are taken from
2608 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2609 international characters are used.
2611 More changes to X509_ATTRIBUTE code: allow the setting of types
2612 based on strings. Remove the 'loc' parameter when adding
2613 attributes because these will be a SET OF encoding which is sorted
2617 *) Initial changes to the 'req' utility to allow request generation
2618 automation. This will allow an application to just generate a template
2619 file containing all the field values and have req construct the
2622 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2623 used all over the place including certificate requests and PKCS#7
2624 structures. They are currently handled manually where necessary with
2625 some primitive wrappers for PKCS#7. The new functions behave in a
2626 manner analogous to the X509 extension functions: they allow
2627 attributes to be looked up by NID and added.
2629 Later something similar to the X509V3 code would be desirable to
2630 automatically handle the encoding, decoding and printing of the
2631 more complex types. The string types like challengePassword can
2632 be handled by the string table functions.
2634 Also modified the multi byte string table handling. Now there is
2635 a 'global mask' which masks out certain types. The table itself
2636 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2637 is useful when for example there is only one permissible type
2638 (as in countryName) and using the mask might result in no valid
2642 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2643 SSL_get_peer_finished to allow applications to obtain the latest
2644 Finished messages sent to the peer or expected from the peer,
2645 respectively. (SSL_get_peer_finished is usually the Finished message
2646 actually received from the peer, otherwise the protocol will be aborted.)
2648 As the Finished message are message digests of the complete handshake
2649 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2650 be used for external authentication procedures when the authentication
2651 provided by SSL/TLS is not desired or is not enough.
2654 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2655 the host supports BWX extension and if Compaq C is present on the
2656 $PATH. Just exploiting of the BWX extension results in 20-30%
2657 performance kick for some algorithms, e.g. DES and RC4 to mention
2658 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2662 *) Add support for MS "fast SGC". This is arguably a violation of the
2663 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2664 weak crypto and after checking the certificate is SGC a second one
2665 with strong crypto. MS SGC stops the first handshake after receiving
2666 the server certificate message and sends a second client hello. Since
2667 a server will typically do all the time consuming operations before
2668 expecting any further messages from the client (server key exchange
2669 is the most expensive) there is little difference between the two.
2671 To get OpenSSL to support MS SGC we have to permit a second client
2672 hello message after we have sent server done. In addition we have to
2673 reset the MAC if we do get this second client hello.
2676 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2677 if a DER encoded private key is RSA or DSA traditional format. Changed
2678 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2679 format DER encoded private key. Newer code should use PKCS#8 format which
2680 has the key type encoded in the ASN1 structure. Added DER private key
2681 support to pkcs8 application.
2684 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2685 ciphersuites has been selected (as required by the SSL 3/TLS 1
2686 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2687 is set, we interpret this as a request to violate the specification
2688 (the worst that can happen is a handshake failure, and 'correct'
2689 behaviour would result in a handshake failure anyway).
2692 *) In SSL_CTX_add_session, take into account that there might be multiple
2693 SSL_SESSION structures with the same session ID (e.g. when two threads
2694 concurrently obtain them from an external cache).
2695 The internal cache can handle only one SSL_SESSION with a given ID,
2696 so if there's a conflict, we now throw out the old one to achieve
2700 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2701 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2702 some routines that use cipher OIDs: some ciphers do not have OIDs
2703 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2707 *) Simplify the trust setting structure and code. Now we just have
2708 two sequences of OIDs for trusted and rejected settings. These will
2709 typically have values the same as the extended key usage extension
2710 and any application specific purposes.
2712 The trust checking code now has a default behaviour: it will just
2713 check for an object with the same NID as the passed id. Functions can
2714 be provided to override either the default behaviour or the behaviour
2715 for a given id. SSL client, server and email already have functions
2716 in place for compatibility: they check the NID and also return "trusted"
2717 if the certificate is self signed.
2720 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2721 traditional format into an EVP_PKEY structure.
2724 *) Add a password callback function PEM_cb() which either prompts for
2725 a password if usr_data is NULL or otherwise assumes it is a null
2726 terminated password. Allow passwords to be passed on command line
2727 environment or config files in a few more utilities.
2730 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2731 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2732 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2733 Update documentation.
2736 *) Support for ASN1 "NULL" type. This could be handled before by using
2737 ASN1_TYPE but there wasn't any function that would try to read a NULL
2738 and produce an error if it couldn't. For compatibility we also have
2739 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2740 don't allocate anything because they don't need to.
2743 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2745 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2747 *) Rebuild of the memory allocation routines used by OpenSSL code and
2748 possibly others as well. The purpose is to make an interface that
2749 provide hooks so anyone can build a separate set of allocation and
2750 deallocation routines to be used by OpenSSL, for example memory
2751 pool implementations, or something else, which was previously hard
2752 since Malloc(), Realloc() and Free() were defined as macros having
2753 the values malloc, realloc and free, respectively (except for Win32
2754 compilations). The same is provided for memory debugging code.
2755 OpenSSL already comes with functionality to find memory leaks, but
2756 this gives people a chance to debug other memory problems.
2758 With these changes, a new set of functions and macros have appeared:
2760 CRYPTO_set_mem_debug_functions() [F]
2761 CRYPTO_get_mem_debug_functions() [F]
2762 CRYPTO_dbg_set_options() [F]
2763 CRYPTO_dbg_get_options() [F]
2764 CRYPTO_malloc_debug_init() [M]
2766 The memory debug functions are NULL by default, unless the library
2767 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2768 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2769 gives the standard debugging functions that come with OpenSSL) or
2770 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2771 provided by the library user) must be used. When the standard
2772 debugging functions are used, CRYPTO_dbg_set_options can be used to
2773 request additional information:
2774 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2775 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2777 Also, things like CRYPTO_set_mem_functions will always give the
2778 expected result (the new set of functions is used for allocation
2779 and deallocation) at all times, regardless of platform and compiler
2782 To finish it up, some functions that were never use in any other
2783 way than through macros have a new API and new semantic:
2786 CRYPTO_dbg_realloc()
2789 All macros of value have retained their old syntax.
2790 [Richard Levitte and Bodo Moeller]
2792 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2793 ordering of SMIMECapabilities wasn't in "strength order" and there
2794 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2798 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2799 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2800 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2802 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2803 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2804 functionality to handle multipart/signed properly) and a utility
2805 called 'smime' to call all this stuff. This is based on code I
2806 originally wrote for Celo who have kindly allowed it to be
2807 included in OpenSSL.
2810 *) Add variants des_set_key_checked and des_set_key_unchecked of
2811 des_set_key (aka des_key_sched). Global variable des_check_key
2812 decides which of these is called by des_set_key; this way
2813 des_check_key behaves as it always did, but applications and
2814 the library itself, which was buggy for des_check_key == 1,
2815 have a cleaner way to pick the version they need.
2818 *) New function PKCS12_newpass() which changes the password of a
2822 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2823 dynamic mix. In both cases the ids can be used as an index into the
2824 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2825 functions so they accept a list of the field values and the
2826 application doesn't need to directly manipulate the X509_TRUST
2830 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2834 *) Modify the way the V3 extension code looks up extensions. This now
2835 works in a similar way to the object code: we have some "standard"
2836 extensions in a static table which is searched with OBJ_bsearch()
2837 and the application can add dynamic ones if needed. The file
2838 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2839 updated whenever a new extension is added to the core code and kept
2840 in ext_nid order. There is a simple program 'tabtest.c' which checks
2841 this. New extensions are not added too often so this file can readily
2842 be maintained manually.
2844 There are two big advantages in doing things this way. The extensions
2845 can be looked up immediately and no longer need to be "added" using
2846 X509V3_add_standard_extensions(): this function now does nothing.
2847 [Side note: I get *lots* of email saying the extension code doesn't
2848 work because people forget to call this function]
2849 Also no dynamic allocation is done unless new extensions are added:
2850 so if we don't add custom extensions there is no need to call
2851 X509V3_EXT_cleanup().
2854 *) Modify enc utility's salting as follows: make salting the default. Add a
2855 magic header, so unsalted files fail gracefully instead of just decrypting
2856 to garbage. This is because not salting is a big security hole, so people
2857 should be discouraged from doing it.
2860 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2861 digest to be passed on the command line but it only used this
2862 parameter when signing a certificate. Modified so all relevant
2863 operations are affected by the digest parameter including the
2864 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2865 DSA key was used because it didn't fix the digest.
2868 *) Initial certificate chain verify code. Currently tests the untrusted
2869 certificates for consistency with the verify purpose (which is set
2870 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2872 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2873 this is because it will reject chains with invalid extensions whereas
2874 every previous version of OpenSSL and SSLeay made no checks at all.
2876 Trust code: checks the root CA for the relevant trust settings. Trust
2877 settings have an initial value consistent with the verify purpose: e.g.
2878 if the verify purpose is for SSL client use it expects the CA to be
2879 trusted for SSL client use. However the default value can be changed to
2880 permit custom trust settings: one example of this would be to only trust
2881 certificates from a specific "secure" set of CAs.
2883 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2884 which should be used for version portability: especially since the
2885 verify structure is likely to change more often now.
2887 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2888 to set them. If not set then assume SSL clients will verify SSL servers
2891 Two new options to the verify program: -untrusted allows a set of
2892 untrusted certificates to be passed in and -purpose which sets the
2893 intended purpose of the certificate. If a purpose is set then the
2894 new chain verify code is used to check extension consistency.
2897 *) Support for the authority information access extension.
2900 *) Modify RSA and DSA PEM read routines to transparently handle
2901 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2902 public keys in a format compatible with certificate
2903 SubjectPublicKeyInfo structures. Unfortunately there were already
2904 functions called *_PublicKey_* which used various odd formats so
2905 these are retained for compatibility: however the DSA variants were
2906 never in a public release so they have been deleted. Changed dsa/rsa
2907 utilities to handle the new format: note no releases ever handled public
2908 keys so we should be OK.
2910 The primary motivation for this change is to avoid the same fiasco
2911 that dogs private keys: there are several incompatible private key
2912 formats some of which are standard and some OpenSSL specific and
2913 require various evil hacks to allow partial transparent handling and
2914 even then it doesn't work with DER formats. Given the option anything
2915 other than PKCS#8 should be dumped: but the other formats have to
2916 stay in the name of compatibility.
2918 With public keys and the benefit of hindsight one standard format
2919 is used which works with EVP_PKEY, RSA or DSA structures: though
2920 it clearly returns an error if you try to read the wrong kind of key.
2922 Added a -pubkey option to the 'x509' utility to output the public key.
2923 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2924 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2925 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2926 that do the same as the EVP_PKEY_assign_*() except they up the
2927 reference count of the added key (they don't "swallow" the
2931 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2932 CRLs would fail if the file contained no certificates or no CRLs:
2933 added a new function to read in both types and return the number
2934 read: this means that if none are read it will be an error. The
2935 DER versions of the certificate and CRL reader would always fail
2936 because it isn't possible to mix certificates and CRLs in DER format
2937 without choking one or the other routine. Changed this to just read
2938 a certificate: this is the best we can do. Also modified the code
2939 in apps/verify.c to take notice of return codes: it was previously
2940 attempting to read in certificates from NULL pointers and ignoring
2941 any errors: this is one reason why the cert and CRL reader seemed
2942 to work. It doesn't check return codes from the default certificate
2943 routines: these may well fail if the certificates aren't installed.
2946 *) Code to support otherName option in GeneralName.
2949 *) First update to verify code. Change the verify utility
2950 so it warns if it is passed a self signed certificate:
2951 for consistency with the normal behaviour. X509_verify
2952 has been modified to it will now verify a self signed
2953 certificate if *exactly* the same certificate appears
2954 in the store: it was previously impossible to trust a
2955 single self signed certificate. This means that:
2956 openssl verify ss.pem
2957 now gives a warning about a self signed certificate but
2958 openssl verify -CAfile ss.pem ss.pem
2962 *) For servers, store verify_result in SSL_SESSION data structure
2963 (and add it to external session representation).
2964 This is needed when client certificate verifications fails,
2965 but an application-provided verification callback (set by
2966 SSL_CTX_set_cert_verify_callback) allows accepting the session
2967 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2968 but returns 1): When the session is reused, we have to set
2969 ssl->verify_result to the appropriate error code to avoid
2971 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2973 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2974 case in PKCS7_dataInit() where the signed PKCS7 structure
2975 didn't contain any existing data because it was being created.
2976 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2978 *) Add a salt to the key derivation routines in enc.c. This
2979 forms the first 8 bytes of the encrypted file. Also add a
2980 -S option to allow a salt to be input on the command line.
2983 *) New function X509_cmp(). Oddly enough there wasn't a function
2984 to compare two certificates. We do this by working out the SHA1
2985 hash and comparing that. X509_cmp() will be needed by the trust
2989 *) SSL_get1_session() is like SSL_get_session(), but increments
2990 the reference count in the SSL_SESSION returned.
2991 [Geoff Thorpe <geoff@eu.c2.net>]
2993 *) Fix for 'req': it was adding a null to request attributes.
2994 Also change the X509_LOOKUP and X509_INFO code to handle
2995 certificate auxiliary information.
2998 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3002 *) Add the possibility to add extra information to the memory leak
3003 detecting output, to form tracebacks, showing from where each
3004 allocation was originated: CRYPTO_push_info("constant string") adds
3005 the string plus current file name and line number to a per-thread
3006 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3007 is like calling CYRPTO_pop_info() until the stack is empty.
3008 Also updated memory leak detection code to be multi-thread-safe.
3011 *) Add options -text and -noout to pkcs7 utility and delete the
3012 encryption options which never did anything. Update docs.
3015 *) Add options to some of the utilities to allow the pass phrase
3016 to be included on either the command line (not recommended on
3017 OSes like Unix) or read from the environment. Update the
3018 manpages and fix a few bugs.
3021 *) Add a few manpages for some of the openssl commands.
3024 *) Fix the -revoke option in ca. It was freeing up memory twice,
3025 leaking and not finding already revoked certificates.
3028 *) Extensive changes to support certificate auxiliary information.
3029 This involves the use of X509_CERT_AUX structure and X509_AUX
3030 functions. An X509_AUX function such as PEM_read_X509_AUX()
3031 can still read in a certificate file in the usual way but it
3032 will also read in any additional "auxiliary information". By
3033 doing things this way a fair degree of compatibility can be
3034 retained: existing certificates can have this information added
3035 using the new 'x509' options.
3037 Current auxiliary information includes an "alias" and some trust
3038 settings. The trust settings will ultimately be used in enhanced
3039 certificate chain verification routines: currently a certificate
3040 can only be trusted if it is self signed and then it is trusted
3044 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3045 The problem was that one of the replacement routines had not been working
3046 since SSLeay releases. For now the offending routine has been replaced
3047 with non-optimised assembler. Even so, this now gives around 95%
3048 performance improvement for 1024 bit RSA signs.
3051 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3052 handling. Most clients have the effective key size in bits equal to
3053 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3054 A few however don't do this and instead use the size of the decrypted key
3055 to determine the RC2 key length and the AlgorithmIdentifier to determine
3056 the effective key length. In this case the effective key length can still
3057 be 40 bits but the key length can be 168 bits for example. This is fixed
3058 by manually forcing an RC2 key into the EVP_PKEY structure because the
3059 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3060 the key length and effective key length are equal.
3063 *) Add a bunch of functions that should simplify the creation of
3064 X509_NAME structures. Now you should be able to do:
3065 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3066 and have it automatically work out the correct field type and fill in
3067 the structures. The more adventurous can try:
3068 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3069 and it will (hopefully) work out the correct multibyte encoding.
3072 *) Change the 'req' utility to use the new field handling and multibyte
3073 copy routines. Before the DN field creation was handled in an ad hoc
3074 way in req, ca, and x509 which was rather broken and didn't support
3075 BMPStrings or UTF8Strings. Since some software doesn't implement
3076 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3077 using the dirstring_type option. See the new comment in the default
3078 openssl.cnf for more info.
3081 *) Make crypto/rand/md_rand.c more robust:
3082 - Assure unique random numbers after fork().
3083 - Make sure that concurrent threads access the global counter and
3084 md serializably so that we never lose entropy in them
3085 or use exactly the same state in multiple threads.
3086 Access to the large state is not always serializable because
3087 the additional locking could be a performance killer, and
3088 md should be large enough anyway.
3091 *) New file apps/app_rand.c with commonly needed functionality
3092 for handling the random seed file.
3094 Use the random seed file in some applications that previously did not:
3096 dsaparam -genkey (which also ignored its '-rand' option),
3099 x509 (when signing).
3100 Except on systems with /dev/urandom, it is crucial to have a random
3101 seed file at least for key creation, DSA signing, and for DH exchanges;
3102 for RSA signatures we could do without one.
3104 gendh and gendsa (unlike genrsa) used to read only the first byte
3105 of each file listed in the '-rand' option. The function as previously
3106 found in genrsa is now in app_rand.c and is used by all programs
3107 that support '-rand'.
3110 *) In RAND_write_file, use mode 0600 for creating files;
3111 don't just chmod when it may be too late.
3114 *) Report an error from X509_STORE_load_locations
3115 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3118 *) New function ASN1_mbstring_copy() this copies a string in either
3119 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3120 into an ASN1_STRING type. A mask of permissible types is passed
3121 and it chooses the "minimal" type to use or an error if not type
3125 *) Add function equivalents to the various macros in asn1.h. The old
3126 macros are retained with an M_ prefix. Code inside the library can
3127 use the M_ macros. External code (including the openssl utility)
3128 should *NOT* in order to be "shared library friendly".
3131 *) Add various functions that can check a certificate's extensions
3132 to see if it usable for various purposes such as SSL client,
3133 server or S/MIME and CAs of these types. This is currently
3134 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3135 verification. Also added a -purpose flag to x509 utility to
3136 print out all the purposes.
3139 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3143 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3144 for, obtain and decode and extension and obtain its critical flag.
3145 This allows all the necessary extension code to be handled in a
3146 single function call.
3149 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3150 platforms. See crypto/rc4/rc4_enc.c for further details.
3153 *) New -noout option to asn1parse. This causes no output to be produced
3154 its main use is when combined with -strparse and -out to extract data
3155 from a file (which may not be in ASN.1 format).
3158 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3159 when producing the local key id.
3160 [Richard Levitte <levitte@stacken.kth.se>]
3162 *) New option -dhparam in s_server. This allows a DH parameter file to be
3163 stated explicitly. If it is not stated then it tries the first server
3164 certificate file. The previous behaviour hard coded the filename
3168 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3169 a public key to be input or output. For example:
3170 openssl rsa -in key.pem -pubout -out pubkey.pem
3171 Also added necessary DSA public key functions to handle this.
3174 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3175 in the message. This was handled by allowing
3176 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3177 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3179 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3180 to the end of the strings whereas this didn't. This would cause problems
3181 if strings read with d2i_ASN1_bytes() were later modified.
3182 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3184 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3185 data and it contains EOF it will end up returning an error. This is
3186 caused by input 46 bytes long. The cause is due to the way base64
3187 BIOs find the start of base64 encoded data. They do this by trying a
3188 trial decode on each line until they find one that works. When they
3189 do a flag is set and it starts again knowing it can pass all the
3190 data directly through the decoder. Unfortunately it doesn't reset
3191 the context it uses. This means that if EOF is reached an attempt
3192 is made to pass two EOFs through the context and this causes the
3193 resulting error. This can also cause other problems as well. As is
3194 usual with these problems it takes *ages* to find and the fix is
3195 trivial: move one line.
3196 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3198 *) Ugly workaround to get s_client and s_server working under Windows. The
3199 old code wouldn't work because it needed to select() on sockets and the
3200 tty (for keypresses and to see if data could be written). Win32 only
3201 supports select() on sockets so we select() with a 1s timeout on the
3202 sockets and then see if any characters are waiting to be read, if none
3203 are present then we retry, we also assume we can always write data to
3204 the tty. This isn't nice because the code then blocks until we've
3205 received a complete line of data and it is effectively polling the
3206 keyboard at 1s intervals: however it's quite a bit better than not
3207 working at all :-) A dedicated Windows application might handle this
3208 with an event loop for example.
3211 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3212 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3213 will be called when RSA_sign() and RSA_verify() are used. This is useful
3214 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3215 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3216 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3217 This necessitated the support of an extra signature type NID_md5_sha1
3218 for SSL signatures and modifications to the SSL library to use it instead
3219 of calling RSA_public_decrypt() and RSA_private_encrypt().
3222 *) Add new -verify -CAfile and -CApath options to the crl program, these
3223 will lookup a CRL issuers certificate and verify the signature in a
3224 similar way to the verify program. Tidy up the crl program so it
3225 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3226 less strict. It will now permit CRL extensions even if it is not
3227 a V2 CRL: this will allow it to tolerate some broken CRLs.
3230 *) Initialize all non-automatic variables each time one of the openssl
3231 sub-programs is started (this is necessary as they may be started
3232 multiple times from the "OpenSSL>" prompt).
3233 [Lennart Bang, Bodo Moeller]
3235 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3236 removing all other RSA functionality (this is what NO_RSA does). This
3237 is so (for example) those in the US can disable those operations covered
3238 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3242 *) Non-copying interface to BIO pairs.
3243 (still largely untested)
3246 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3247 ASCII string. This was handled independently in various places before.
3250 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3251 UTF8 strings a character at a time.
3254 *) Use client_version from client hello to select the protocol
3255 (s23_srvr.c) and for RSA client key exchange verification
3256 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3259 *) Add various utility functions to handle SPKACs, these were previously
3260 handled by poking round in the structure internals. Added new function
3261 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3262 print, verify and generate SPKACs. Based on an original idea from
3263 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3266 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3269 *) Allow the config file extension section to be overwritten on the
3270 command line. Based on an original idea from Massimiliano Pala
3271 <madwolf@comune.modena.it>. The new option is called -extensions
3272 and can be applied to ca, req and x509. Also -reqexts to override
3273 the request extensions in req and -crlexts to override the crl extensions
3277 *) Add new feature to the SPKAC handling in ca. Now you can include
3278 the same field multiple times by preceding it by "XXXX." for example:
3281 this is the same syntax as used in the req config file.
3284 *) Allow certificate extensions to be added to certificate requests. These
3285 are specified in a 'req_extensions' option of the req section of the
3286 config file. They can be printed out with the -text option to req but
3287 are otherwise ignored at present.
3290 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3291 data read consists of only the final block it would not decrypted because
3292 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3293 A misplaced 'break' also meant the decrypted final block might not be
3294 copied until the next read.
3297 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3298 a few extra parameters to the DH structure: these will be useful if
3299 for example we want the value of 'q' or implement X9.42 DH.
3302 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3303 provides hooks that allow the default DSA functions or functions on a
3304 "per key" basis to be replaced. This allows hardware acceleration and
3305 hardware key storage to be handled without major modification to the
3306 library. Also added low level modexp hooks and CRYPTO_EX structure and
3307 associated functions.
3310 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3311 as "read only": it can't be written to and the buffer it points to will
3312 not be freed. Reading from a read only BIO is much more efficient than
3313 a normal memory BIO. This was added because there are several times when
3314 an area of memory needs to be read from a BIO. The previous method was
3315 to create a memory BIO and write the data to it, this results in two
3316 copies of the data and an O(n^2) reading algorithm. There is a new
3317 function BIO_new_mem_buf() which creates a read only memory BIO from
3318 an area of memory. Also modified the PKCS#7 routines to use read only
3322 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3323 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3324 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3325 but a retry condition occured while trying to read the rest.
3328 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3329 NID_pkcs7_encrypted by default: this was wrong since this should almost
3330 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3331 the encrypted data type: this is a more sensible place to put it and it
3332 allows the PKCS#12 code to be tidied up that duplicated this
3336 *) Changed obj_dat.pl script so it takes its input and output files on
3337 the command line. This should avoid shell escape redirection problems
3341 *) Initial support for certificate extension requests, these are included
3342 in things like Xenroll certificate requests. Included functions to allow
3343 extensions to be obtained and added.
3346 *) -crlf option to s_client and s_server for sending newlines as
3347 CRLF (as required by many protocols).
3350 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3352 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3353 [Ralf S. Engelschall]
3355 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3356 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3358 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3362 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3363 DH parameters/keys (q is lost during that conversion, but the resulting
3364 DH parameters contain its length).
3366 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3367 much faster than DH_generate_parameters (which creates parameters
3368 where p = 2*q + 1), and also the smaller q makes DH computations
3369 much more efficient (160-bit exponentiation instead of 1024-bit
3370 exponentiation); so this provides a convenient way to support DHE
3371 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3372 utter importance to use
3373 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3375 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3376 when such DH parameters are used, because otherwise small subgroup
3377 attacks may become possible!
3380 *) Avoid memory leak in i2d_DHparams.
3383 *) Allow the -k option to be used more than once in the enc program:
3384 this allows the same encrypted message to be read by multiple recipients.
3387 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3388 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3389 it will always use the numerical form of the OID, even if it has a short
3393 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3394 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3395 otherwise bn_mod_exp was called. In the case of hardware keys for example
3396 no private key components need be present and it might store extra data
3397 in the RSA structure, which cannot be accessed from bn_mod_exp.
3398 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3399 private key operations.
3402 *) Added support for SPARC Linux.
3405 *) pem_password_cb function type incompatibly changed from
3406 typedef int pem_password_cb(char *buf, int size, int rwflag);
3408 ....(char *buf, int size, int rwflag, void *userdata);
3409 so that applications can pass data to their callbacks:
3410 The PEM[_ASN1]_{read,write}... functions and macros now take an
3411 additional void * argument, which is just handed through whenever
3412 the password callback is called.
3413 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3415 New function SSL_CTX_set_default_passwd_cb_userdata.
3417 Compatibility note: As many C implementations push function arguments
3418 onto the stack in reverse order, the new library version is likely to
3419 interoperate with programs that have been compiled with the old
3420 pem_password_cb definition (PEM_whatever takes some data that
3421 happens to be on the stack as its last argument, and the callback
3422 just ignores this garbage); but there is no guarantee whatsoever that
3425 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3426 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3427 problems not only on Windows, but also on some Unix platforms.
3428 To avoid problematic command lines, these definitions are now in an
3429 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3430 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3433 *) MIPS III/IV assembler module is reimplemented.
3436 *) More DES library cleanups: remove references to srand/rand and
3437 delete an unused file.
3440 *) Add support for the the free Netwide assembler (NASM) under Win32,
3441 since not many people have MASM (ml) and it can be hard to obtain.
3442 This is currently experimental but it seems to work OK and pass all
3443 the tests. Check out INSTALL.W32 for info.
3446 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3447 without temporary keys kept an extra copy of the server key,
3448 and connections with temporary keys did not free everything in case
3452 *) New function RSA_check_key and new openssl rsa option -check
3453 for verifying the consistency of RSA keys.
3454 [Ulf Moeller, Bodo Moeller]
3456 *) Various changes to make Win32 compile work:
3457 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3458 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3459 comparison" warnings.
3460 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3463 *) Add a debugging option to PKCS#5 v2 key generation function: when
3464 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3465 derived keys are printed to stderr.
3468 *) Copy the flags in ASN1_STRING_dup().
3469 [Roman E. Pavlov <pre@mo.msk.ru>]
3471 *) The x509 application mishandled signing requests containing DSA
3472 keys when the signing key was also DSA and the parameters didn't match.
3474 It was supposed to omit the parameters when they matched the signing key:
3475 the verifying software was then supposed to automatically use the CA's
3476 parameters if they were absent from the end user certificate.
3478 Omitting parameters is no longer recommended. The test was also
3479 the wrong way round! This was probably due to unusual behaviour in
3480 EVP_cmp_parameters() which returns 1 if the parameters match.
3481 This meant that parameters were omitted when they *didn't* match and
3482 the certificate was useless. Certificates signed with 'ca' didn't have
3484 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3486 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3487 The interface is as follows:
3488 Applications can use
3489 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3490 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3491 "off" is now the default.
3492 The library internally uses
3493 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3494 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3495 to disable memory-checking temporarily.
3497 Some inconsistent states that previously were possible (and were
3498 even the default) are now avoided.
3500 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3501 with each memory chunk allocated; this is occasionally more helpful
3502 than just having a counter.
3504 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3506 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3510 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3511 which largely parallels "options", but is for changing API behaviour,
3512 whereas "options" are about protocol behaviour.
3513 Initial "mode" flags are:
3515 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3516 a single record has been written.
3517 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3518 retries use the same buffer location.
3519 (But all of the contents must be
3523 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3526 *) Fix problems with no-hmac etc.
3527 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3529 *) New functions RSA_get_default_method(), RSA_set_method() and
3530 RSA_get_method(). These allows replacement of RSA_METHODs without having
3531 to mess around with the internals of an RSA structure.
3534 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3535 Also really enable memory leak checks in openssl.c and in some
3537 [Chad C. Mulligan, Bodo Moeller]
3539 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3540 up the length of negative integers. This has now been simplified to just
3541 store the length when it is first determined and use it later, rather
3542 than trying to keep track of where data is copied and updating it to
3544 [Steve Henson, reported by Brien Wheeler
3545 <bwheeler@authentica-security.com>]
3547 *) Add a new function PKCS7_signatureVerify. This allows the verification
3548 of a PKCS#7 signature but with the signing certificate passed to the
3549 function itself. This contrasts with PKCS7_dataVerify which assumes the
3550 certificate is present in the PKCS#7 structure. This isn't always the
3551 case: certificates can be omitted from a PKCS#7 structure and be
3552 distributed by "out of band" means (such as a certificate database).
3555 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3556 function prototypes in pem.h, also change util/mkdef.pl to add the
3557 necessary function names.
3560 *) mk1mf.pl (used by Windows builds) did not properly read the
3561 options set by Configure in the top level Makefile, and Configure
3562 was not even able to write more than one option correctly.
3563 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3566 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3567 file to be loaded from a BIO or FILE pointer. The BIO version will
3568 for example allow memory BIOs to contain config info.
3571 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3572 Whoever hopes to achieve shared-library compatibility across versions
3573 must use this, not the compile-time macro.
3574 (Exercise 0.9.4: Which is the minimum library version required by
3576 Note: All this applies only to multi-threaded programs, others don't
3580 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3581 through a BIO pair triggered the default case, i.e.
3582 SSLerr(...,SSL_R_UNKNOWN_STATE).
3585 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3586 can use the SSL library even if none of the specific BIOs is
3590 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3591 for the encoded length.
3592 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3594 *) Add initial documentation of the X509V3 functions.
3597 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3598 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3599 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3600 secure PKCS#8 private key format with a high iteration count.
3603 *) Fix determination of Perl interpreter: A perl or perl5
3604 _directory_ in $PATH was also accepted as the interpreter.
3605 [Ralf S. Engelschall]
3607 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3608 wrong with it but it was very old and did things like calling
3609 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3613 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3614 to use the new extension code.
3617 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3618 with macros. This should make it easier to change their form, add extra
3619 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3623 *) Add to configuration table a new entry that can specify an alternative
3624 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3625 according to Mark Crispin <MRC@Panda.COM>.
3629 *) DES CBC did not update the IV. Weird.
3632 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3633 Changing the behaviour of the former might break existing programs --
3634 where IV updating is needed, des_ncbc_encrypt can be used.
3637 *) When bntest is run from "make test" it drives bc to check its
3638 calculations, as well as internally checking them. If an internal check
3639 fails, it needs to cause bc to give a non-zero result or make test carries
3640 on without noticing the failure. Fixed.
3643 *) DES library cleanups.
3646 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3647 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3648 ciphers. NOTE: although the key derivation function has been verified
3649 against some published test vectors it has not been extensively tested
3650 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3654 *) Instead of "mkdir -p", which is not fully portable, use new
3655 Perl script "util/mkdir-p.pl".
3658 *) Rewrite the way password based encryption (PBE) is handled. It used to
3659 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3660 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3661 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3662 the 'parameter' field of the AlgorithmIdentifier is passed to the
3663 underlying key generation function so it must do its own ASN1 parsing.
3664 This has also changed the EVP_PBE_CipherInit() function which now has a
3665 'parameter' argument instead of literal salt and iteration count values
3666 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3669 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3670 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3671 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3672 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3673 value was just used as a "magic string" and not used directly its
3674 value doesn't matter.
3677 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3681 *) "linux-sparc64" configuration (ultrapenguin).
3682 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3683 "linux-sparc" configuration.
3684 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3686 *) config now generates no-xxx options for missing ciphers.
3689 *) Support the EBCDIC character set (work in progress).
3690 File ebcdic.c not yet included because it has a different license.
3691 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3693 *) Support BS2000/OSD-POSIX.
3694 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3696 *) Make callbacks for key generation use void * instead of char *.
3699 *) Make S/MIME samples compile (not yet tested).
3702 *) Additional typesafe stacks.
3705 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3709 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3711 *) New configuration variant "sco5-gcc".
3713 *) Updated some demos.
3714 [Sean O Riordain, Wade Scholine]
3716 *) Add missing BIO_free at exit of pkcs12 application.
3719 *) Fix memory leak in conf.c.
3722 *) Updates for Win32 to assembler version of MD5.
3725 *) Set #! path to perl in apps/der_chop to where we found it
3726 instead of using a fixed path.
3729 *) SHA library changes for irix64-mips4-cc.
3732 *) Improvements for VMS support.
3736 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3738 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3739 This also avoids the problems with SC4.2 and unpatched SC5.
3740 [Andy Polyakov <appro@fy.chalmers.se>]
3742 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3743 These are required because of the typesafe stack would otherwise break
3744 existing code. If old code used a structure member which used to be STACK
3745 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3746 sk_num or sk_value it would produce an error because the num, data members
3747 are not present in STACK_OF. Now it just produces a warning. sk_set
3748 replaces the old method of assigning a value to sk_value
3749 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3750 that does this will no longer work (and should use sk_set instead) but
3751 this could be regarded as a "questionable" behaviour anyway.
3754 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3755 correctly handle encrypted S/MIME data.
3758 *) Change type of various DES function arguments from des_cblock
3759 (which means, in function argument declarations, pointer to char)
3760 to des_cblock * (meaning pointer to array with 8 char elements),
3761 which allows the compiler to do more typechecking; it was like
3762 that back in SSLeay, but with lots of ugly casts.
3764 Introduce new type const_des_cblock.
3767 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3768 problems: find RecipientInfo structure that matches recipient certificate
3769 and initialise the ASN1 structures properly based on passed cipher.
3772 *) Belatedly make the BN tests actually check the results.
3775 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3776 to and from BNs: it was completely broken. New compilation option
3777 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3778 key elements as negative integers.
3781 *) Reorganize and speed up MD5.
3782 [Andy Polyakov <appro@fy.chalmers.se>]
3785 [Richard Levitte <richard@levitte.org>]
3787 *) New option -out to asn1parse to allow the parsed structure to be
3788 output to a file. This is most useful when combined with the -strparse
3789 option to examine the output of things like OCTET STRINGS.
3792 *) Make SSL library a little more fool-proof by not requiring any longer
3793 that SSL_set_{accept,connect}_state be called before
3794 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3795 in many applications because usually everything *appeared* to work as
3796 intended anyway -- now it really works as intended).
3799 *) Move openssl.cnf out of lib/.
3802 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3803 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3804 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3805 [Ralf S. Engelschall]
3807 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3808 handle PKCS#7 enveloped data properly.
3809 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3811 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3812 copying pointers. The cert_st handling is changed by this in
3813 various ways (and thus what used to be known as ctx->default_cert
3814 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3815 any longer when s->cert does not give us what we need).
3816 ssl_cert_instantiate becomes obsolete by this change.
3817 As soon as we've got the new code right (possibly it already is?),
3818 we have solved a couple of bugs of the earlier code where s->cert
3819 was used as if it could not have been shared with other SSL structures.
3821 Note that using the SSL API in certain dirty ways now will result
3822 in different behaviour than observed with earlier library versions:
3823 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3824 does not influence s as it used to.
3826 In order to clean up things more thoroughly, inside SSL_SESSION
3827 we don't use CERT any longer, but a new structure SESS_CERT
3828 that holds per-session data (if available); currently, this is
3829 the peer's certificate chain and, for clients, the server's certificate
3830 and temporary key. CERT holds only those values that can have
3831 meaningful defaults in an SSL_CTX.
3834 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3835 from the internal representation. Various PKCS#7 fixes: remove some
3836 evil casts and set the enc_dig_alg field properly based on the signing
3840 *) Allow PKCS#12 password to be set from the command line or the
3841 environment. Let 'ca' get its config file name from the environment
3842 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3846 *) Allow certificate policies extension to use an IA5STRING for the
3847 organization field. This is contrary to the PKIX definition but
3848 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3852 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3853 without disallowing inline assembler and the like for non-pedantic builds.
3856 *) Support Borland C++ builder.
3857 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3862 *) SHA-1 cleanups and performance enhancements.
3863 [Andy Polyakov <appro@fy.chalmers.se>]
3865 *) Sparc v8plus assembler for the bignum library.
3866 [Andy Polyakov <appro@fy.chalmers.se>]
3868 *) Accept any -xxx and +xxx compiler options in Configure.
3871 *) Update HPUX configuration.
3874 *) Add missing sk_<type>_unshift() function to safestack.h
3875 [Ralf S. Engelschall]
3877 *) New function SSL_CTX_use_certificate_chain_file that sets the
3878 "extra_cert"s in addition to the certificate. (This makes sense
3879 only for "PEM" format files, as chains as a whole are not
3883 *) Support verify_depth from the SSL API.
3884 x509_vfy.c had what can be considered an off-by-one-error:
3885 Its depth (which was not part of the external interface)
3886 was actually counting the number of certificates in a chain;
3887 now it really counts the depth.
3890 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3891 instead of X509err, which often resulted in confusing error
3892 messages since the error codes are not globally unique
3893 (e.g. an alleged error in ssl3_accept when a certificate
3894 didn't match the private key).
3896 *) New function SSL_CTX_set_session_id_context that allows to set a default
3897 value (so that you don't need SSL_set_session_id_context for each
3898 connection using the SSL_CTX).
3901 *) OAEP decoding bug fix.
3904 *) Support INSTALL_PREFIX for package builders, as proposed by
3908 *) New Configure options "threads" and "no-threads". For systems
3909 where the proper compiler options are known (currently Solaris
3910 and Linux), "threads" is the default.
3913 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3916 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3917 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3918 such as /usr/local/bin.
3921 *) "make linux-shared" to build shared libraries.
3922 [Niels Poppe <niels@netbox.org>]
3924 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3927 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3928 extension adding in x509 utility.
3931 *) Remove NOPROTO sections and error code comments.
3934 *) Partial rewrite of the DEF file generator to now parse the ANSI
3938 *) New Configure options --prefix=DIR and --openssldir=DIR.
3941 *) Complete rewrite of the error code script(s). It is all now handled
3942 by one script at the top level which handles error code gathering,
3943 header rewriting and C source file generation. It should be much better
3944 than the old method: it now uses a modified version of Ulf's parser to
3945 read the ANSI prototypes in all header files (thus the old K&R definitions
3946 aren't needed for error creation any more) and do a better job of
3947 translating function codes into names. The old 'ASN1 error code imbedded
3948 in a comment' is no longer necessary and it doesn't use .err files which
3949 have now been deleted. Also the error code call doesn't have to appear all
3950 on one line (which resulted in some large lines...).
3953 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3956 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3957 0 (which usually indicates a closed connection), but continue reading.
3960 *) Fix some race conditions.
3963 *) Add support for CRL distribution points extension. Add Certificate
3964 Policies and CRL distribution points documentation.
3967 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3970 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3971 8 of keying material. Merlin has also confirmed interop with this fix
3972 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3973 [Merlin Hughes <merlin@baltimore.ie>]
3975 *) Fix lots of warnings.
3976 [Richard Levitte <levitte@stacken.kth.se>]
3978 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3979 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3980 [Richard Levitte <levitte@stacken.kth.se>]
3982 *) Fix problems with sizeof(long) == 8.
3983 [Andy Polyakov <appro@fy.chalmers.se>]
3985 *) Change functions to ANSI C.
3988 *) Fix typos in error codes.
3989 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3991 *) Remove defunct assembler files from Configure.
3994 *) SPARC v8 assembler BIGNUM implementation.
3995 [Andy Polyakov <appro@fy.chalmers.se>]
3997 *) Support for Certificate Policies extension: both print and set.
3998 Various additions to support the r2i method this uses.
4001 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4002 return a const string when you are expecting an allocated buffer.
4005 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4006 types DirectoryString and DisplayText.
4009 *) Add code to allow r2i extensions to access the configuration database,
4010 add an LHASH database driver and add several ctx helper functions.
4013 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4014 fail when they extended the size of a BIGNUM.
4017 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4018 support typesafe stack.
4021 *) Fix typo in SSL_[gs]et_options().
4022 [Nils Frostberg <nils@medcom.se>]
4024 *) Delete various functions and files that belonged to the (now obsolete)
4025 old X509V3 handling code.
4028 *) New Configure option "rsaref".
4031 *) Don't auto-generate pem.h.
4034 *) Introduce type-safe ASN.1 SETs.
4037 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4038 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4040 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4041 that links with OpenSSL (well at least cause lots of warnings), but fear
4042 not: the conversion is trivial, and it eliminates loads of evil casts. A
4043 few STACKed things have been converted already. Feel free to convert more.
4044 In the fullness of time, I'll do away with the STACK type altogether.
4047 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4048 specified in <certfile> by updating the entry in the index.txt file.
4049 This way one no longer has to edit the index.txt file manually for
4050 revoking a certificate. The -revoke option does the gory details now.
4051 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4053 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4054 `-text' option at all and this way the `-noout -text' combination was
4055 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4056 [Ralf S. Engelschall]
4058 *) Make sure a corresponding plain text error message exists for the
4059 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4060 verify callback function determined that a certificate was revoked.
4061 [Ralf S. Engelschall]
4063 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4064 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4065 all available cipers including rc5, which was forgotten until now.
4066 In order to let the testing shell script know which algorithms
4067 are available, a new (up to now undocumented) command
4068 "openssl list-cipher-commands" is used.
4071 *) Bugfix: s_client occasionally would sleep in select() when
4072 it should have checked SSL_pending() first.
4075 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4076 the raw DSA values prior to ASN.1 encoding.
4079 *) Tweaks to Configure
4080 [Niels Poppe <niels@netbox.org>]
4082 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4086 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4089 *) New config option to avoid instructions that are illegal on the 80386.
4090 The default code is faster, but requires at least a 486.
4093 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4094 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4095 same as SSL2_VERSION anyway.
4098 *) New "-showcerts" option for s_client.
4101 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4102 application. Various cleanups and fixes.
4105 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4106 modify error routines to work internally. Add error codes and PBE init
4107 to library startup routines.
4110 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4111 packing functions to asn1 and evp. Changed function names and error
4112 codes along the way.
4115 *) PKCS12 integration: and so it begins... First of several patches to
4116 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4117 objects to objects.h
4120 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4121 and display support for Thawte strong extranet extension.
4124 *) Add LinuxPPC support.
4125 [Jeff Dubrule <igor@pobox.org>]
4127 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4128 bn_div_words in alpha.s.
4129 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4131 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4132 OAEP isn't supported when OpenSSL is built with RSAref.
4133 [Ulf Moeller <ulf@fitug.de>]
4135 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4136 so they no longer are missing under -DNOPROTO.
4137 [Soren S. Jorvang <soren@t.dk>]
4140 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4142 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4143 doesn't work when the session is reused. Coming soon!
4146 *) Fix a security hole, that allows sessions to be reused in the wrong
4147 context thus bypassing client cert protection! All software that uses
4148 client certs and session caches in multiple contexts NEEDS PATCHING to
4149 allow session reuse! A fuller solution is in the works.
4150 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4152 *) Some more source tree cleanups (removed obsolete files
4153 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4154 permission on "config" script to be executable) and a fix for the INSTALL
4156 [Ulf Moeller <ulf@fitug.de>]
4158 *) Remove some legacy and erroneous uses of malloc, free instead of
4160 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4162 *) Make rsa_oaep_test return non-zero on error.
4163 [Ulf Moeller <ulf@fitug.de>]
4165 *) Add support for native Solaris shared libraries. Configure
4166 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4167 if someone would make that last step automatic.
4168 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4170 *) ctx_size was not built with the right compiler during "make links". Fixed.
4173 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4174 except NULL ciphers". This means the default cipher list will no longer
4175 enable NULL ciphers. They need to be specifically enabled e.g. with
4176 the string "DEFAULT:eNULL".
4179 *) Fix to RSA private encryption routines: if p < q then it would
4180 occasionally produce an invalid result. This will only happen with
4181 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4184 *) Be less restrictive and allow also `perl util/perlpath.pl
4185 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4186 because this way one can also use an interpreter named `perl5' (which is
4187 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4188 installed as `perl').
4189 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4191 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4192 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4194 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4195 advapi32.lib to Win32 build and change the pem test comparision
4196 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4197 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4198 and crypto/des/ede_cbcm_enc.c.
4201 *) DES quad checksum was broken on big-endian architectures. Fixed.
4204 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4205 Win32 test batch file so it (might) work again. The Win32 test batch file
4206 is horrible: I feel ill....
4209 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4210 in e_os.h. Audit of header files to check ANSI and non ANSI
4211 sections: 10 functions were absent from non ANSI section and not exported
4212 from Windows DLLs. Fixed up libeay.num for new functions.
4215 *) Make `openssl version' output lines consistent.
4216 [Ralf S. Engelschall]
4218 *) Fix Win32 symbol export lists for BIO functions: Added
4219 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4220 to ms/libeay{16,32}.def.
4221 [Ralf S. Engelschall]
4223 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4224 fine under Unix and passes some trivial tests I've now added. But the
4225 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4226 added to make sure no one expects that this stuff really works in the
4227 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4228 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4230 [Ralf S. Engelschall]
4232 *) Fix the generation of two part addresses in perl.
4233 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4235 *) Add config entry for Linux on MIPS.
4236 [John Tobey <jtobey@channel1.com>]
4238 *) Make links whenever Configure is run, unless we are on Windoze.
4241 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4242 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4246 *) Add a useful kludge to allow package maintainers to specify compiler and
4247 other platforms details on the command line without having to patch the
4248 Configure script everytime: One now can use ``perl Configure
4249 <id>:<details>'', i.e. platform ids are allowed to have details appended
4250 to them (seperated by colons). This is treated as there would be a static
4251 pre-configured entry in Configure's %table under key <id> with value
4252 <details> and ``perl Configure <id>'' is called. So, when you want to
4253 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4254 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4255 now, which overrides the FreeBSD-elf entry on-the-fly.
4256 [Ralf S. Engelschall]
4258 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4261 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4262 on the `perl Configure ...' command line. This way one can compile
4263 OpenSSL libraries with Position Independent Code (PIC) which is needed
4264 for linking it into DSOs.
4265 [Ralf S. Engelschall]
4267 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4271 *) Cleaned up the LICENSE document: The official contact for any license
4272 questions now is the OpenSSL core team under openssl-core@openssl.org.
4273 And add a paragraph about the dual-license situation to make sure people
4274 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4275 to the OpenSSL toolkit.
4276 [Ralf S. Engelschall]
4278 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4279 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4280 Additonally cleaned up the `make links' target: Remove unnecessary
4281 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4282 to speed processing and no longer clutter the display with confusing
4283 stuff. Instead only the actually done links are displayed.
4284 [Ralf S. Engelschall]
4286 *) Permit null encryption ciphersuites, used for authentication only. It used
4287 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4288 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4292 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4293 signed attributes when verifying signatures (this would break them),
4294 the detached data encoding was wrong and public keys obtained using
4295 X509_get_pubkey() weren't freed.
4298 *) Add text documentation for the BUFFER functions. Also added a work around
4299 to a Win95 console bug. This was triggered by the password read stuff: the
4300 last character typed gets carried over to the next fread(). If you were
4301 generating a new cert request using 'req' for example then the last
4302 character of the passphrase would be CR which would then enter the first
4306 *) Added the new `Includes OpenSSL Cryptography Software' button as
4307 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4308 button and can be used by applications based on OpenSSL to show the
4309 relationship to the OpenSSL project.
4310 [Ralf S. Engelschall]
4312 *) Remove confusing variables in function signatures in files
4313 ssl/ssl_lib.c and ssl/ssl.h.
4314 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4316 *) Don't install bss_file.c under PREFIX/include/
4317 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4319 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4320 functions that return function pointers and has support for NT specific
4321 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4322 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4323 unsigned to signed types: this was killing the Win32 compile.
4326 *) Add new certificate file to stack functions,
4327 SSL_add_dir_cert_subjects_to_stack() and
4328 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4329 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4330 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4331 This means that Apache-SSL and similar packages don't have to mess around
4332 to add as many CAs as they want to the preferred list.
4335 *) Experiment with doxygen documentation. Currently only partially applied to
4337 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4338 openssl.doxy as the configuration file.
4341 *) Get rid of remaining C++-style comments which strict C compilers hate.
4342 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4344 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4345 compiled in by default: it has problems with large keys.
4348 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4349 DH private keys and/or callback functions which directly correspond to
4350 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4351 is needed for applications which have to configure certificates on a
4352 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4354 For the RSA certificate situation is makes no difference, but
4355 for the DSA certificate situation this fixes the "no shared cipher"
4356 problem where the OpenSSL cipher selection procedure failed because the
4357 temporary keys were not overtaken from the context and the API provided
4358 no way to reconfigure them.
4359 The new functions now let applications reconfigure the stuff and they
4360 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4361 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4362 non-public-API function ssl_cert_instantiate() is used as a helper
4363 function and also to reduce code redundancy inside ssl_rsa.c.
4364 [Ralf S. Engelschall]
4366 *) Move s_server -dcert and -dkey options out of the undocumented feature
4367 area because they are useful for the DSA situation and should be
4368 recognized by the users.
4369 [Ralf S. Engelschall]
4371 *) Fix the cipher decision scheme for export ciphers: the export bits are
4372 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4373 SSL_EXP_MASK. So, the original variable has to be used instead of the
4374 already masked variable.
4375 [Richard Levitte <levitte@stacken.kth.se>]
4377 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4378 [Richard Levitte <levitte@stacken.kth.se>]
4380 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4381 from `int' to `unsigned int' because it's a length and initialized by
4382 EVP_DigestFinal() which expects an `unsigned int *'.
4383 [Richard Levitte <levitte@stacken.kth.se>]
4385 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4386 script. Instead use the usual Shell->Perl transition trick.
4387 [Ralf S. Engelschall]
4389 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4390 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4391 -noout -modulus' as it's already the case for `openssl rsa -noout
4392 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4393 currently the public key is printed (a decision which was already done by
4394 `openssl dsa -modulus' in the past) which serves a similar purpose.
4395 Additionally the NO_RSA no longer completely removes the whole -modulus
4396 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4398 [Ralf S. Engelschall]
4400 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4401 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4402 [Arne Ansper <arne@ats.cyber.ee>]
4404 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4405 to be added. Now both 'req' and 'ca' can use new objects defined in the
4409 *) Add cool BIO that does syslog (or event log on NT).
4410 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4412 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4413 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4414 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4415 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4418 *) Add preliminary config info for new extension code.
4421 *) Make RSA_NO_PADDING really use no padding.
4422 [Ulf Moeller <ulf@fitug.de>]
4424 *) Generate errors when private/public key check is done.
4427 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4428 for some CRL extensions and new objects added.
4431 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4432 key usage extension and fuller support for authority key id.
4435 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4436 padding method for RSA, which is recommended for new applications in PKCS
4437 #1 v2.0 (RFC 2437, October 1998).
4438 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4439 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4440 against Bleichbacher's attack on RSA.
4441 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4444 *) Updates to the new SSL compression code
4445 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4447 *) Fix so that the version number in the master secret, when passed
4448 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4449 (because the server will not accept higher), that the version number
4450 is 0x03,0x01, not 0x03,0x00
4451 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4453 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4454 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4455 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4458 *) Support for RAW extensions where an arbitrary extension can be
4459 created by including its DER encoding. See apps/openssl.cnf for
4463 *) Make sure latest Perl versions don't interpret some generated C array
4464 code as Perl array code in the crypto/err/err_genc.pl script.
4465 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4467 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4468 not many people have the assembler. Various Win32 compilation fixes and
4469 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4473 *) Modify configure script 'Configure' to automatically create crypto/date.h
4474 file under Win32 and also build pem.h from pem.org. New script
4475 util/mkfiles.pl to create the MINFO file on environments that can't do a
4476 'make files': perl util/mkfiles.pl >MINFO should work.
4479 *) Major rework of DES function declarations, in the pursuit of correctness
4480 and purity. As a result, many evil casts evaporated, and some weirdness,
4481 too. You may find this causes warnings in your code. Zapping your evil
4482 casts will probably fix them. Mostly.
4485 *) Fix for a typo in asn1.h. Bug fix to object creation script
4486 obj_dat.pl. It considered a zero in an object definition to mean
4487 "end of object": none of the objects in objects.h have any zeros
4488 so it wasn't spotted.
4489 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4491 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4492 Masking (CBCM). In the absence of test vectors, the best I have been able
4493 to do is check that the decrypt undoes the encrypt, so far. Send me test
4494 vectors if you have them.
4497 *) Correct calculation of key length for export ciphers (too much space was
4498 allocated for null ciphers). This has not been tested!
4501 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4502 message is now correct (it understands "crypto" and "ssl" on its
4503 command line). There is also now an "update" option. This will update
4504 the util/ssleay.num and util/libeay.num files with any new functions.
4506 perl util/mkdef.pl crypto ssl update
4507 it will update them.
4510 *) Overhauled the Perl interface (perl/*):
4511 - ported BN stuff to OpenSSL's different BN library
4512 - made the perl/ source tree CVS-aware
4513 - renamed the package from SSLeay to OpenSSL (the files still contain
4514 their history because I've copied them in the repository)
4515 - removed obsolete files (the test scripts will be replaced
4516 by better Test::Harness variants in the future)
4517 [Ralf S. Engelschall]
4519 *) First cut for a very conservative source tree cleanup:
4520 1. merge various obsolete readme texts into doc/ssleay.txt
4521 where we collect the old documents and readme texts.
4522 2. remove the first part of files where I'm already sure that we no
4523 longer need them because of three reasons: either they are just temporary
4524 files which were left by Eric or they are preserved original files where
4525 I've verified that the diff is also available in the CVS via "cvs diff
4526 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4527 the crypto/md/ stuff).
4528 [Ralf S. Engelschall]
4530 *) More extension code. Incomplete support for subject and issuer alt
4531 name, issuer and authority key id. Change the i2v function parameters
4532 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4533 what that's for :-) Fix to ASN1 macro which messed up
4534 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4537 *) Preliminary support for ENUMERATED type. This is largely copied from the
4541 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4542 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4544 *) Make sure `make rehash' target really finds the `openssl' program.
4545 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4547 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4548 like to hear about it if this slows down other processors.
4551 *) Add CygWin32 platform information to Configure script.
4552 [Alan Batie <batie@aahz.jf.intel.com>]
4554 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4555 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4557 *) New program nseq to manipulate netscape certificate sequences
4560 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4564 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4565 but the BN code had some problems that would cause failures when
4566 doing certificate verification and some other functions.
4567 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4569 *) Add ASN1 and PEM code to support netscape certificate sequences.
4572 *) Add ASN1 and PEM code to support netscape certificate sequences.
4575 *) Add several PKIX and private extended key usage OIDs.
4578 *) Modify the 'ca' program to handle the new extension code. Modify
4579 openssl.cnf for new extension format, add comments.
4582 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
4583 and add a sample to openssl.cnf so req -x509 now adds appropriate
4587 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
4588 error code, add initial support to X509_print() and x509 application.
4591 *) Takes a deep breath and start addding X509 V3 extension support code. Add
4592 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
4593 stuff is currently isolated and isn't even compiled yet.
4596 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
4597 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
4598 Removed the versions check from X509 routines when loading extensions:
4599 this allows certain broken certificates that don't set the version
4600 properly to be processed.
4603 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
4604 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
4605 can still be regenerated with "make depend".
4608 *) Spelling mistake in C version of CAST-128.
4609 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
4611 *) Changes to the error generation code. The perl script err-code.pl
4612 now reads in the old error codes and retains the old numbers, only
4613 adding new ones if necessary. It also only changes the .err files if new
4614 codes are added. The makefiles have been modified to only insert errors
4615 when needed (to avoid needlessly modifying header files). This is done
4616 by only inserting errors if the .err file is newer than the auto generated
4617 C file. To rebuild all the error codes from scratch (the old behaviour)
4618 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
4619 or delete all the .err files.
4622 *) CAST-128 was incorrectly implemented for short keys. The C version has
4623 been fixed, but is untested. The assembler versions are also fixed, but
4624 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
4625 to regenerate it if needed.
4626 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4627 Hagino <itojun@kame.net>]
4629 *) File was opened incorrectly in randfile.c.
4630 [Ulf Möller <ulf@fitug.de>]
4632 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4633 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4634 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4635 al: it's just almost always a UTCTime. Note this patch adds new error
4636 codes so do a "make errors" if there are problems.
4639 *) Correct Linux 1 recognition in config.
4640 [Ulf Möller <ulf@fitug.de>]
4642 *) Remove pointless MD5 hash when using DSA keys in ca.
4643 [Anonymous <nobody@replay.com>]
4645 *) Generate an error if given an empty string as a cert directory. Also
4646 generate an error if handed NULL (previously returned 0 to indicate an
4647 error, but didn't set one).
4648 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4650 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4653 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4654 parameters. This was causing a warning which killed off the Win32 compile.
4657 *) Remove C++ style comments from crypto/bn/bn_local.h.
4658 [Neil Costigan <neil.costigan@celocom.com>]
4660 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4661 based on a text string, looking up short and long names and finally
4662 "dot" format. The "dot" format stuff didn't work. Added new function
4663 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4664 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4665 OID is not part of the table.
4668 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4669 X509_LOOKUP_by_alias().
4672 *) Sort openssl functions by name.
4675 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4676 encryption from sample DSA keys (in case anyone is interested the password
4680 *) Make _all_ *_free functions accept a NULL pointer.
4681 [Frans Heymans <fheymans@isaserver.be>]
4683 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4685 [Anonymous <nobody@replay.com>]
4687 *) s_server should send the CAfile as acceptable CAs, not its own cert.
4688 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4690 *) Don't blow it for numeric -newkey arguments to apps/req.
4691 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4693 *) Temp key "for export" tests were wrong in s3_srvr.c.
4694 [Anonymous <nobody@replay.com>]
4696 *) Add prototype for temp key callback functions
4697 SSL_CTX_set_tmp_{rsa,dh}_callback().
4700 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4701 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4704 *) X509_name_add_entry() freed the wrong thing after an error.
4705 [Arne Ansper <arne@ats.cyber.ee>]
4707 *) rsa_eay.c would attempt to free a NULL context.
4708 [Arne Ansper <arne@ats.cyber.ee>]
4710 *) BIO_s_socket() had a broken should_retry() on Windoze.
4711 [Arne Ansper <arne@ats.cyber.ee>]
4713 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4714 [Arne Ansper <arne@ats.cyber.ee>]
4716 *) Make sure the already existing X509_STORE->depth variable is initialized
4717 in X509_STORE_new(), but document the fact that this variable is still
4718 unused in the certificate verification process.
4719 [Ralf S. Engelschall]
4721 *) Fix the various library and apps files to free up pkeys obtained from
4722 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4725 *) Fix reference counting in X509_PUBKEY_get(). This makes
4726 demos/maurice/example2.c work, amongst others, probably.
4727 [Steve Henson and Ben Laurie]
4729 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4730 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4731 are no longer created. This way we have a single and consistent command
4732 line interface `openssl <command>', similar to `cvs <command>'.
4733 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4735 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4736 BIT STRING wrapper always have zero unused bits.
4739 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4742 *) Make the top-level INSTALL documentation easier to understand.
4745 *) Makefiles updated to exit if an error occurs in a sub-directory
4746 make (including if user presses ^C) [Paul Sutton]
4748 *) Make Montgomery context stuff explicit in RSA data structure.
4751 *) Fix build order of pem and err to allow for generated pem.h.
4754 *) Fix renumbering bug in X509_NAME_delete_entry().
4757 *) Enhanced the err-ins.pl script so it makes the error library number
4758 global and can add a library name. This is needed for external ASN1 and
4759 other error libraries.
4762 *) Fixed sk_insert which never worked properly.
4765 *) Fix ASN1 macros so they can handle indefinite length construted
4766 EXPLICIT tags. Some non standard certificates use these: they can now
4770 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4771 into a single doc/ssleay.txt bundle. This way the information is still
4772 preserved but no longer messes up this directory. Now it's new room for
4773 the new set of documenation files.
4774 [Ralf S. Engelschall]
4776 *) SETs were incorrectly DER encoded. This was a major pain, because they
4777 shared code with SEQUENCEs, which aren't coded the same. This means that
4778 almost everything to do with SETs or SEQUENCEs has either changed name or
4779 number of arguments.
4780 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4782 *) Fix test data to work with the above.
4785 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4786 was already fixed by Eric for 0.9.1 it seems.
4787 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4789 *) Autodetect FreeBSD3.
4792 *) Fix various bugs in Configure. This affects the following platforms:
4796 unixware-2.0-pentium
4800 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4801 before they are needed.
4804 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4808 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4810 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4811 changed SSLeay to OpenSSL in version strings.
4812 [Ralf S. Engelschall]
4814 *) Some fixups to the top-level documents.
4817 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4818 because the symlink to include/ was missing.
4819 [Ralf S. Engelschall]
4821 *) Incorporated the popular no-RSA/DSA-only patches
4822 which allow to compile a RSA-free SSLeay.
4823 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4825 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4826 when "ssleay" is still not found.
4827 [Ralf S. Engelschall]
4829 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4830 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4832 *) Updated the README file.
4833 [Ralf S. Engelschall]
4835 *) Added various .cvsignore files in the CVS repository subdirs
4836 to make a "cvs update" really silent.
4837 [Ralf S. Engelschall]
4839 *) Recompiled the error-definition header files and added
4840 missing symbols to the Win32 linker tables.
4841 [Ralf S. Engelschall]
4843 *) Cleaned up the top-level documents;
4844 o new files: CHANGES and LICENSE
4845 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4846 o merged COPYRIGHT into LICENSE
4847 o removed obsolete TODO file
4848 o renamed MICROSOFT to INSTALL.W32
4849 [Ralf S. Engelschall]
4851 *) Removed dummy files from the 0.9.1b source tree:
4852 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4853 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4854 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4855 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4856 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4857 [Ralf S. Engelschall]
4859 *) Added various platform portability fixes.
4862 *) The Genesis of the OpenSSL rpject:
4863 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4864 Young and Tim J. Hudson created while they were working for C2Net until
4866 [The OpenSSL Project]
4869 Changes between 0.9.0b and 0.9.1b [not released]
4871 *) Updated a few CA certificates under certs/
4874 *) Changed some BIGNUM api stuff.
4877 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4878 DGUX x86, Linux Alpha, etc.
4881 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4882 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4886 *) Add -strparse option to asn1pars program which parses nested
4888 [Dr Stephen Henson <shenson@bigfoot.com>]
4890 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4893 *) DSA fix for "ca" program.
4896 *) Added "-genkey" option to "dsaparam" program.
4899 *) Added RIPE MD160 (rmd160) message digest.
4902 *) Added -a (all) option to "ssleay version" command.
4905 *) Added PLATFORM define which is the id given to Configure.
4908 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4911 *) Extended the ASN.1 parser routines.
4914 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4917 *) Added a BN_CTX to the BN library.
4920 *) Fixed the weak key values in DES library
4923 *) Changed API in EVP library for cipher aliases.
4926 *) Added support for RC2/64bit cipher.
4929 *) Converted the lhash library to the crypto/mem.c functions.
4932 *) Added more recognized ASN.1 object ids.
4935 *) Added more RSA padding checks for SSL/TLS.
4938 *) Added BIO proxy/filter functionality.
4941 *) Added extra_certs to SSL_CTX which can be used
4942 send extra CA certificates to the client in the CA cert chain sending
4943 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4946 *) Now Fortezza is denied in the authentication phase because
4947 this is key exchange mechanism is not supported by SSLeay at all.
4950 *) Additional PKCS1 checks.
4953 *) Support the string "TLSv1" for all TLS v1 ciphers.
4956 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4957 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4960 *) Fixed a few memory leaks.
4963 *) Fixed various code and comment typos.
4966 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4967 bytes sent in the client random.
4968 [Edward Bishop <ebishop@spyglass.com>]