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 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
15 to support NCONF routines in extension code. New function CONF_set_nconf()
16 to allow functions which take an NCONF to also handle the old LHASH
17 structure: this means that the old CONF compatible routines can be
18 retained (in particular wrt extensions) without having to duplicate the
19 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
22 *) Handle special case when X509_NAME is empty in X509 printing routines.
25 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
26 positive and less than q.
29 +) Enhance the general user interface with mechanisms for inner control
30 and with pssibilities to have yes/no kind of prompts.
33 *) Change all calls to low level digest routines in the library and
34 applications to use EVP. Add missing calls to HMAC_cleanup() and
35 don't assume HMAC_CTX can be copied using memcpy().
36 [Verdon Walker <VWalker@novell.com>, Steve Henson]
38 +) Add the possibility to control engines through control names but with
39 arbitrary arguments instead of just a string.
40 Change the key loaders to take a UI_METHOD instead of a callback
41 function pointer. NOTE: this breaks binary compatibility with earlier
42 versions of OpenSSL [engine].
43 Addapt the nCipher code for these new conditions and add a card insertion
47 +) Enhance the general user interface with mechanisms to better support
48 dialog box interfaces, application-defined prompts, the possibility
49 to use defaults (for example default passwords from somewhere else)
50 and interrupts/cancelations.
53 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
54 used: it isn't thread safe and the add_lock_callback should handle
56 [Paul Rose <Paul.Rose@bridge.com>]
58 *) Verify that incoming data obeys the block size in
59 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
62 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
63 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
67 [Ulf Möller, Bodo Möller]
69 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
70 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5
71 when fixing the server behaviour for backwards-compatible 'client
72 hello' messages. (Note that the attack is impractical against
73 SSL 3.0 and TLS 1.0 anyway because length and version checking
74 means that the probability of guessing a valid ciphertext is
75 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
78 Before 0.9.5, the countermeasure (hide the error by generating a
79 random 'decryption result') did not work properly because
80 ERR_clear_error() was missing, meaning that SSL_get_error() would
81 detect the supposedly ignored error.
83 Both problems are now fixed.
86 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
87 (previously it was 1024).
90 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
91 tidy up some unecessarily weird code in 'sk_new()').
92 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
94 +) Change the key loading routines for ENGINEs to use the same kind
95 callback (pem_password_cb) as all other routines that need this
99 *) Fix for compatibility mode trust settings: ignore trust settings
100 unless some valid trust or reject settings are present.
103 *) Fix for blowfish EVP: its a variable length cipher.
106 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
107 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
108 than this minimum value is recommended.
111 +) New random seeder for OpenVMS, using the system process statistics
112 that are easily reachable.
115 +) Windows apparently can't transparently handle global
116 variables defined in DLLs. Initialisations such as:
118 const ASN1_ITEM *it = &ASN1_INTEGER_it;
120 wont compile. This is used by the any applications that need to
121 delcare their own ASN1 modules. This was fixed by adding the option
122 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
123 needed for static libraries under Win32.
126 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
127 setting of purpose and trust fields. New X509_STORE trust and
128 purpose functions and tidy up setting in other SSL functions.
131 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
132 structure. These are inherited by X509_STORE_CTX when it is
133 initialised. This allows various defaults to be set in the
134 X509_STORE structure (such as flags for CRL checking and custom
135 purpose or trust settings) for functions which only use X509_STORE_CTX
136 internally such as S/MIME.
138 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
139 trust settings if they are not set in X509_STORE. This allows X509_STORE
140 purposes and trust (in S/MIME for example) to override any set by default.
142 Add command line options for CRL checking to smime, s_client and s_server
146 +) Initial CRL based revocation checking. If the CRL checking flag(s)
147 are set then the CRL is looked up in the X509_STORE structure and
148 its validity and signature checked, then if the certificate is found
149 in the CRL the verify fails with a revoked error.
151 Various new CRL related callbacks added to X509_STORE_CTX structure.
153 Command line options added to 'verify' application to support this.
155 This needs some additional work, such as being able to handle multiple
156 CRLs with different times, extension based lookup (rather than just
157 by subject name) and ultimately more complete V2 CRL extension
161 +) Add a general user interface API (crypto/ui/). This is designed
162 to replace things like des_read_password and friends (backward
163 compatibility functions using this new API are provided).
164 The purpose is to remove prompting functions from the DES code
165 section as well as provide for prompting through dialog boxes in
166 a window system and the like.
169 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
170 in the current directory if neither $RANDFILE nor $HOME was set.
171 RAND_file_name() in 0.9.6a returned NULL in this case. This has
172 caused some confusion to Windows users who haven't defined $HOME.
173 Thus RAND_file_name() is changed again: e_os.h can define a
174 DEFAULT_HOME, which will be used if $HOME is not set.
175 For Windows, we use "C:"; on other platforms, we still require
176 environment variables.
178 +) Add "ex_data" support to ENGINE so implementations can add state at a
179 per-structure level rather than having to store it globally.
182 +) Make it possible for ENGINE structures to be copied when retrieved by
183 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
184 This causes the "original" ENGINE structure to act like a template,
185 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
186 operational state can be localised to each ENGINE structure, despite the
187 fact they all share the same "methods". New ENGINE structures returned in
188 this case have no functional references and the return value is the single
189 structural reference. This matches the single structural reference returned
190 by ENGINE_by_id() normally, when it is incremented on the pre-existing
194 *) Fix various bugs related to DSA S/MIME verification. Handle missing
195 parameters in DSA public key structures and return an error in the
196 DSA routines if parameters are absent.
199 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
200 needs to match any other type at all we need to manually clear the
204 +) Changes to the "openssl engine" utility to include;
205 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
206 about an ENGINE's available control commands.
207 - executing control commands from command line arguments using the
208 '-pre' and '-post' switches. '-post' is only used if '-t' is
209 specified and the ENGINE is successfully initialised. The syntax for
210 the individual commands are colon-separated, for example;
211 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
214 +) New dynamic control command support for ENGINEs. ENGINEs can now
215 declare their own commands (numbers), names (strings), descriptions,
216 and input types for run-time discovery by calling applications. A
217 subset of these commands are implicitly classed as "executable"
218 depending on their input type, and only these can be invoked through
219 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
220 can be based on user input, config files, etc). The distinction is
221 that "executable" commands cannot return anything other than a boolean
222 result and can only support numeric or string input, whereas some
223 discoverable commands may only be for direct use through
224 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
225 pointers, or other custom uses. The "executable" commands are to
226 support parameterisations of ENGINE behaviour that can be
227 unambiguously defined by ENGINEs and used consistently across any
228 OpenSSL-based application. Commands have been added to all the
229 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
230 control over shared-library paths without source code alterations.
233 +) Changed all ENGINE implementations to dynamically allocate their
234 ENGINEs rather than declaring them statically. Apart from this being
235 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
236 this also allows the implementations to compile without using the
237 internal engine_int.h header.
240 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
241 'const' value. Any code that should be able to modify a RAND_METHOD
242 should already have non-const pointers to it (ie. they should only
243 modify their own ones).
246 +) Made a variety of little tweaks to the ENGINE code.
247 - "atalla" and "ubsec" string definitions were moved from header files
248 to C code. "nuron" string definitions were placed in variables
249 rather than hard-coded - allowing parameterisation of these values
250 later on via ctrl() commands.
251 - Removed unused "#if 0"'d code.
252 - Fixed engine list iteration code so it uses ENGINE_free() to release
253 structural references.
254 - Constified the RAND_METHOD element of ENGINE structures.
255 - Constified various get/set functions as appropriate and added
256 missing functions (including a catch-all ENGINE_cpy that duplicates
257 all ENGINE values onto a new ENGINE except reference counts/state).
258 - Removed NULL parameter checks in get/set functions. Setting a method
259 or function to NULL is a way of cancelling out a previously set
260 value. Passing a NULL ENGINE parameter is just plain stupid anyway
261 and doesn't justify the extra error symbols and code.
262 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
263 flags from engine_int.h to engine.h.
264 - Changed prototypes for ENGINE handler functions (init(), finish(),
265 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
268 *) Move 'if (!initialized) RAND_poll()' into regions protected by
269 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
270 having multiple threads call RAND_poll() concurrently.
273 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
274 combination of a flag and a thread ID variable.
275 Otherwise while one thread is in ssleay_rand_bytes (which sets the
276 flag), *other* threads can enter ssleay_add_bytes without obeying
277 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock
278 that they do not hold after the first thread unsets add_do_not_lock).
281 +) Implement binary inversion algorithm for BN_mod_inverse in addition
282 to the algorithm using long divison. The binary algorithm can be
283 used only if the modulus is odd. On 32-bit systems, it is faster
284 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
285 roughly 5-15% for 256-bit moduli), so we use it only for moduli
286 up to 450 bits. In 64-bit environments, the binary algorithm
287 appears to be advantageous for much longer moduli; here we use it
288 for moduli up to 2048 bits.
291 *) Change bctest again: '-x' expressions are not available in all
295 -) OpenSSL 0.9.6a released [5 Apr 2001]
297 *) Fix a couple of memory leaks in PKCS7_dataDecode()
298 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
300 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
301 the default extension for executables, if any. Also, make the perl
302 scripts that use symlink() to test if it really exists and use "cp"
303 if it doesn't. All this made OpenSSL compilable and installable in
307 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
308 could not support the combine flag in choice fields.
311 -) Fix for asn1_GetSequence() for indefinite length constructed data.
312 If SEQUENCE is length is indefinite just set c->slen to the total
313 amount of data available.
314 [Steve Henson, reported by shige@FreeBSD.org]
315 [This change does not apply to 0.9.7.]
317 *) Change bctest to avoid here-documents inside command substitution
318 (workaround for FreeBSD /bin/sh bug).
319 For compatibility with Ultrix, avoid shell functions (introduced
320 in the bctest version that searches along $PATH).
323 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
324 with des_encrypt() defined on some operating systems, like Solaris
328 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
329 On the Importance of Eliminating Errors in Cryptographic
330 Computations, J. Cryptology 14 (2001) 2, 101-119,
331 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
334 *) MIPS assembler BIGNUM division bug fix.
337 *) Disabled incorrect Alpha assembler code.
340 -) Fix PKCS#7 decode routines so they correctly update the length
341 after reading an EOC for the EXPLICIT tag.
343 [This change does not apply to 0.9.7.]
345 *) Fix bug in PKCS#12 key generation routines. This was triggered
346 if a 3DES key was generated with a 0 initial byte. Include
347 PKCS12_BROKEN_KEYGEN compilation option to retain the old
348 (but broken) behaviour.
351 *) Enhance bctest to search for a working bc along $PATH and print
353 [Tim Rice <tim@multitalents.net> via Richard Levitte]
355 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
356 extensions from a certificate request to the certificate.
359 +) Allow multiple 'certopt' and 'nameopt' options to be separated
360 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
361 file: this allows the display of the certificate about to be
362 signed to be customised, to allow certain fields to be included
363 or excluded and extension details. The old system didn't display
364 multicharacter strings properly, omitted fields not in the policy
365 and couldn't display additional details such as extensions.
368 *) Fix memory leaks in err.c: free err_data string if necessary;
369 don't write to the wrong index in ERR_set_error_data.
372 +) Function EC_POINTs_mul for simultaneous scalar multiplication
373 of an arbitrary number of elliptic curve points, optionally
374 including the generator defined for the EC_GROUP.
375 EC_POINT_mul is a simple wrapper function for the typical case
376 that the point list has just one item (besides the optional
380 +) First EC_METHODs for curves over GF(p):
382 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
383 operations and provides various method functions that can also
384 operate with faster implementations of modular arithmetic.
386 EC_GFp_mont_method() reuses most functions that are part of
387 EC_GFp_simple_method, but uses Montgomery arithmetic.
389 [Bodo Moeller; point addition and point doubling
390 implementation directly derived from source code provided by
391 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
393 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
396 Curves are EC_GROUP objects (with an optional group generator)
397 based on EC_METHODs that are built into the library.
399 Points are EC_POINT objects based on EC_GROUP objects.
401 Most of the framework would be able to handle curves over arbitrary
402 finite fields, but as there are no obvious types for fields other
403 than GF(p), some functions are limited to that for now.
406 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
407 that the file contains a complete HTTP response.
410 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
411 change the def and num file printf format specifier from "%-40sXXX"
412 to "%-39s XXX". The latter will always guarantee a space after the
413 field while the former will cause them to run together if the field
414 is 40 of more characters long.
417 +) Constify the cipher and digest 'method' functions and structures
418 and modify related functions to take constant EVP_MD and EVP_CIPHER
422 *) Implement ssl23_peek (analogous to ssl23_read), which previously
426 *) Replace rdtsc with _emit statements for VC++ version 5.
427 [Jeremy Cooper <jeremy@baymoo.org>]
429 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
430 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
433 +) Modify EVP_Digest*() routines so they now return values. Although the
434 internal software routines can never fail additional hardware versions
438 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
440 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
441 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
444 ERR_R_NESTED_ASN1_ERROR
446 ERR_R_MISSING_ASN1_EOS
447 were 4 .. 9, conflicting with
448 ERR_LIB_RSA (= ERR_R_RSA_LIB)
450 ERR_LIB_PEM (= ERR_R_PEM_LIB).
451 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
453 Add new error code 'ERR_R_INTERNAL_ERROR'.
456 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
460 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
461 sets the subject name for a new request or supersedes the
462 subject name in a given request. Formats that can be parsed are
463 'CN=Some Name, OU=myOU, C=IT'
465 'CN=Some Name/OU=myOU/C=IT'.
467 Add options '-batch' and '-verbose' to 'openssl req'.
468 [Massimiliano Pala <madwolf@hackmasters.net>]
470 *) Make it possible to reuse SSLv2 sessions.
473 +) Introduce the possibility to access global variables through
474 functions on platform were that's the best way to handle exporting
475 global variables in shared libraries. To enable this functionality,
476 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
477 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
478 is normally done by Configure or something similar).
480 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
481 in the source file (foo.c) like this:
483 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
484 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
486 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
487 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
489 OPENSSL_DECLARE_GLOBAL(int,foo);
490 #define foo OPENSSL_GLOBAL_REF(foo)
491 OPENSSL_DECLARE_GLOBAL(double,bar);
492 #define bar OPENSSL_GLOBAL_REF(bar)
494 The #defines are very important, and therefore so is including the
495 header file everywere where the defined globals are used.
497 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
498 of ASN.1 items, but that structure is a bt different.
500 The largest change is in util/mkdef.pl which has been enhanced with
501 better and easier to understand logic to choose which symbols should
502 go into the Windows .def files as well as a number of fixes and code
503 cleanup (among others, algorithm keywords are now sorted
504 lexicographically to avoid constant rewrites).
507 *) In copy_email() check for >= 0 as a return value for
508 X509_NAME_get_index_by_NID() since 0 is a valid index.
509 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
511 +) In BN_div() keep a copy of the sign of 'num' before writing the
512 result to 'rm' because if rm==num the value will be overwritten
513 and produce the wrong result if 'num' is negative: this caused
514 problems with BN_mod() and BN_nnmod().
517 +) Function OCSP_request_verify(). This checks the signature on an
518 OCSP request and verifies the signer certificate. The signer
519 certificate is just checked for a generic purpose and OCSP request
523 +) Add OCSP_check_validity() function to check the validity of OCSP
524 responses. OCSP responses are prepared in real time and may only
525 be a few seconds old. Simply checking that the current time lies
526 between thisUpdate and nextUpdate max reject otherwise valid responses
527 caused by either OCSP responder or client clock innacuracy. Instead
528 we allow thisUpdate and nextUpdate to fall within a certain period of
529 the current time. The age of the response can also optionally be
530 checked. Two new options -validity_period and -status_age added to
534 +) If signature or public key algorithm is unrecognized print out its
535 OID rather that just UNKOWN.
538 *) Avoid coredump with unsupported or invalid public keys by checking if
539 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
540 PKCS7_verify() fails with non detached data.
543 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
544 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
545 ID to be generated from the issuer certificate alone which can then be
546 passed to OCSP_id_issuer_cmp().
549 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
550 ASN1 modules to export functions returning ASN1_ITEM pointers
551 instead of the ASN1_ITEM structures themselves. This adds several
552 new macros which allow the underlying ASN1 function/structure to
553 be accessed transparently. As a result code should not use ASN1_ITEM
554 references directly (such as &X509_it) but instead use the relevant
555 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
556 use of the new ASN1 code on platforms where exporting structures
557 is problematical (for example in shared libraries) but exporting
558 functions returning pointers to structures is not.
561 +) Add support for overriding the generation of SSL/TLS session IDs.
562 These callbacks can be registered either in an SSL_CTX or per SSL.
563 The purpose of this is to allow applications to control, if they wish,
564 the arbitrary values chosen for use as session IDs, particularly as it
565 can be useful for session caching in multiple-server environments. A
566 command-line switch for testing this (and any client code that wishes
567 to use such a feature) has been added to "s_server".
568 [Geoff Thorpe, Lutz Jaenicke]
570 +) Modify mkdef.pl to recognise and parse prprocessor conditionals
571 of the form '#if defined(...) || defined(...) || ...' and
572 '#if !defined(...) && !defined(...) && ...'. This also avoids
573 the growing number of special cases it was previously handling.
576 *) Don't use getenv in library functions when run as setuid/setgid.
577 New function OPENSSL_issetugid().
580 +) Make all configuration macros available for application by making
581 sure they are available in opensslconf.h, by giving them names starting
582 with "OPENSSL_" to avoid conflicts with other packages and by making
583 sure e_os2.h will cover all platform-specific cases together with
585 Additionally, it is now possible to define configuration/platform-
586 specific names (called "system identities"). In the C code, these
587 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
588 macro with the name beginning with "OPENSSL_SYS_", which is determined
589 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
593 +) New option -set_serial to 'req' and 'x509' this allows the serial
594 number to use to be specified on the command line. Previously self
595 signed certificates were hard coded with serial number 0 and the
596 CA options of 'x509' had to use a serial number in a file which was
600 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
601 due to incorrect handling of multi-threading:
603 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
605 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
607 3. Count how many times MemCheck_off() has been called so that
608 nested use can be treated correctly. This also avoids
609 inband-signalling in the previous code (which relied on the
610 assumption that thread ID 0 is impossible).
613 +) New options to 'ca' utility to support V2 CRL entry extensions.
614 Currently CRL reason, invalidity date and hold instruction are
615 supported. Add new CRL extensions to V3 code and some new objects.
618 *) Add "-rand" option also to s_client and s_server.
621 +) New function EVP_CIPHER_CTX_set_padding() this is used to
622 disable standard block padding (aka PKCS#5 padding) in the EVP
623 API, which was previously mandatory. This means that the data is
624 not padded in any way and so the total length much be a multiple
625 of the block size, otherwise an error occurs.
628 +) Initial (incomplete) OCSP SSL support.
631 *) Fix CPU detection on Irix 6.x.
632 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
633 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
635 +) New function OCSP_parse_url(). This splits up a URL into its host,
636 port and path components: primarily to parse OCSP URLs. New -url
637 option to ocsp utility.
640 +) New nonce behavior. The return value of OCSP_check_nonce() now
641 reflects the various checks performed. Applications can decide
642 whether to tolerate certain situations such as an absent nonce
643 in a response when one was present in a request: the ocsp application
644 just prints out a warning. New function OCSP_add1_basic_nonce()
645 this is to allow responders to include a nonce in a response even if
646 the request is nonce-less.
649 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
651 [This change does not apply to 0.9.7.]
653 *) Use the cached encoding of an X509_NAME structure rather than
654 copying it. This is apparently the reason for the libsafe "errors"
655 but the code is actually correct.
658 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
659 skipped when using openssl x509 multiple times on a single input file,
660 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
663 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
664 set string type: to handle setting ASN1_TIME structures. Fix ca
665 utility to correctly initialize revocation date of CRLs.
668 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
669 the clients preferred ciphersuites and rather use its own preferences.
670 Should help to work around M$ SGC (Server Gated Cryptography) bug in
671 Internet Explorer by ensuring unchanged hash method during stepup.
674 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
675 to aes and add a new 'exist' option to print out symbols that don't
679 +) Additional options to ocsp utility to allow flags to be set and
680 additional certificates supplied.
683 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
684 OCSP client a number of certificate to only verify the response
688 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
689 Bleichenbacher's DSA attack.
690 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
691 to be set and top=0 forces the highest bit to be set; top=-1 is new
692 and leaves the highest bit random.
693 [Ulf Moeller, Bodo Moeller]
695 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
696 handle the new API. Currently only ECB, CBC modes supported. Add new
697 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
698 for TLS" draft-ietf-tls-ciphersuite-03.txt.
699 [Ben Laurie, Steve Henson]
701 *) In the NCONF_...-based implementations for CONF_... queries
702 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
703 a temporary CONF structure with the data component set to NULL
704 (which gives segmentation faults in lh_retrieve).
705 Instead, use NULL for the CONF pointer in CONF_get_string and
706 CONF_get_number (which may use environment variables) and directly
707 return NULL from CONF_get_section.
710 *) Fix potential buffer overrun for EBCDIC.
713 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
717 +) Functions for OCSP responders. OCSP_request_onereq_count(),
718 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
719 extract information from a certificate request. OCSP_response_create()
720 creates a response and optionally adds a basic response structure.
721 OCSP_basic_add1_status() adds a complete single response to a basic
722 reponse and returns the OCSP_SINGLERESP structure just added (to allow
723 extensions to be included for example). OCSP_basic_add1_cert() adds a
724 certificate to a basic response and OCSP_basic_sign() signs a basic
725 response with various flags. New helper functions ASN1_TIME_check()
726 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
727 (converts ASN1_TIME to GeneralizedTime).
730 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
731 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
732 structure from a certificate. X509_pubkey_digest() digests tha public_key
733 contents: this is used in various key identifiers.
736 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
737 keyUsage if basicConstraints absent for a CA.
740 *) Make SMIME_write_PKCS7() write mail header values with a format that
741 is more generally accepted (no spaces before the semicolon), since
742 some programs can't parse those values properly otherwise. Also make
743 sure BIO's that break lines after each write do not create invalid
747 +) Make sk_sort() tolerate a NULL argument.
748 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
750 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
751 passed by the function are trusted implicitly. If any of them signed the
752 reponse then it is assumed to be valid and is not verified.
755 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
756 macros previously used would not encode an empty SEQUENCE OF
757 and break the signature.
759 [This change does not apply to 0.9.7.]
761 *) Zero the premaster secret after deriving the master secret in
765 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
766 to data. This was previously part of the PKCS7 ASN1 code. This
767 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
768 [Steve Henson, reported by Kenneth R. Robinette
769 <support@securenetterm.com>]
771 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
772 routines: without these tracing memory leaks is very painful.
773 Fix leaks in PKCS12 and PKCS7 routines.
776 *) Add some EVP_add_digest_alias registrations (as found in
777 OpenSSL_add_all_digests()) to SSL_library_init()
778 aka OpenSSL_add_ssl_algorithms(). This provides improved
779 compatibility with peers using X.509 certificates
780 with unconventional AlgorithmIdentifier OIDs.
783 *) Fix for Irix with NO_ASM.
784 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
786 *) ./config script fixes.
787 [Ulf Moeller, Richard Levitte]
789 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
790 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
791 effectively meant GeneralizedTime would never be used. Now it
792 is initialised to -1 but X509_time_adj() now has to check the value
793 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
794 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
795 [Steve Henson, reported by Kenneth R. Robinette
796 <support@securenetterm.com>]
798 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
799 result in a zero length in the ASN1_INTEGER structure which was
800 not consistent with the structure when d2i_ASN1_INTEGER() was used
801 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
802 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
803 where it did not print out a minus for negative ASN1_INTEGER.
806 *) Fix 'openssl passwd -1'.
809 +) Add summary printout to ocsp utility. The various functions which
810 convert status values to strings have been renamed to:
811 OCSP_response_status_str(), OCSP_cert_status_str() and
812 OCSP_crl_reason_str() and are no longer static. New options
813 to verify nonce values and to disable verification. OCSP response
814 printout format cleaned up.
817 +) Add additional OCSP certificate checks. These are those specified
818 in RFC2560. This consists of two separate checks: the CA of the
819 certificate being checked must either be the OCSP signer certificate
820 or the issuer of the OCSP signer certificate. In the latter case the
821 OCSP signer certificate must contain the OCSP signing extended key
822 usage. This check is performed by attempting to match the OCSP
823 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
824 in the OCSP_CERTID structures of the response.
827 +) Initial OCSP certificate verification added to OCSP_basic_verify()
828 and related routines. This uses the standard OpenSSL certificate
829 verify routines to perform initial checks (just CA validity) and
830 to obtain the certificate chain. Then additional checks will be
831 performed on the chain. Currently the root CA is checked to see
832 if it is explicitly trusted for OCSP signing. This is used to set
833 a root CA as a global signing root: that is any certificate that
834 chains to that CA is an acceptable OCSP signing certificate.
837 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
838 extensions from a separate configuration file.
839 As when reading extensions from the main configuration file,
840 the '-extensions ...' option may be used for specifying the
842 [Massimiliano Pala <madwolf@comune.modena.it>]
844 *) Change PKCS12_key_gen_asc() so it can cope with non null
845 terminated strings whose length is passed in the passlen
846 parameter, for example from PEM callbacks. This was done
847 by adding an extra length parameter to asc2uni().
848 [Steve Henson, reported by <oddissey@samsung.co.kr>]
850 +) New OCSP utility. Allows OCSP requests to be generated or
851 read. The request can be sent to a responder and the output
852 parsed, outputed or printed in text form. Not complete yet:
853 still needs to check the OCSP response validity.
856 +) New subcommands for 'openssl ca':
857 'openssl ca -status <serial>' prints the status of the cert with
858 the given serial number (according to the index file).
859 'openssl ca -updatedb' updates the expiry status of certificates
861 [Massimiliano Pala <madwolf@comune.modena.it>]
863 +) New '-newreq-nodes' command option to CA.pl. This is like
864 '-newreq', but calls 'openssl req' with the '-nodes' option
865 so that the resulting key is not encrypted.
866 [Damien Miller <djm@mindrot.org>]
868 +) New configuration for the GNU Hurd.
869 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
871 +) Initial code to implement OCSP basic response verify. This
872 is currently incomplete. Currently just finds the signer's
873 certificate and verifies the signature on the response.
876 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
877 value of OPENSSLDIR. This is available via the new '-d' option
878 to 'openssl version', and is also included in 'openssl version -a'.
881 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
882 call failed, free the DSA structure.
885 +) Allowing defining memory allocation callbacks that will be given
886 file name and line number information in additional arguments
887 (a const char* and an int). The basic functionality remains, as
888 well as the original possibility to just replace malloc(),
889 realloc() and free() by functions that do not know about these
890 additional arguments. To register and find out the current
891 settings for extended allocation functions, the following
892 functions are provided:
894 CRYPTO_set_mem_ex_functions
895 CRYPTO_set_locked_mem_ex_functions
896 CRYPTO_get_mem_ex_functions
897 CRYPTO_get_locked_mem_ex_functions
899 These work the same way as CRYPTO_set_mem_functions and friends.
900 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
901 extended allocation function is enabled.
902 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
903 a conventional allocation function is enabled.
904 [Richard Levitte, Bodo Moeller]
906 *) Fix to uni2asc() to cope with zero length Unicode strings.
907 These are present in some PKCS#12 files.
910 +) Finish off removing the remaining LHASH function pointer casts.
911 There should no longer be any prototype-casting required when using
912 the LHASH abstraction, and any casts that remain are "bugs". See
913 the callback types and macros at the head of lhash.h for details
914 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
917 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
918 If /dev/[u]random devices are not available or do not return enough
919 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
921 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
922 /etc/entropy will be queried once each in this sequence, quering stops
923 when enough entropy was collected without querying more sockets.
926 +) Change the Unix RAND_poll() variant to be able to poll several
927 random devices, as specified by DEVRANDOM, until a sufficient amount
928 of data has been collected. We spend at most 10 ms on each file
929 (select timeout) and read in non-blocking mode. DEVRANDOM now
930 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
931 (previously it was just the string "/dev/urandom"), so on typical
932 platforms the 10 ms delay will never occur.
933 Also separate out the Unix variant to its own file, rand_unix.c.
934 For VMS, there's a currently-empty rand_vms.c.
937 +) Move OCSP client related routines to ocsp_cl.c. These
938 provide utility functions which an application needing
939 to issue a request to an OCSP responder and analyse the
940 response will typically need: as opposed to those which an
941 OCSP responder itself would need which will be added later.
943 OCSP_request_sign() signs an OCSP request with an API similar
944 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
945 response. OCSP_response_get1_basic() extracts basic response
946 from response. OCSP_resp_find_status(): finds and extracts status
947 information from an OCSP_CERTID structure (which will be created
948 when the request structure is built). These are built from lower
949 level functions which work on OCSP_SINGLERESP structures but
950 wont normally be used unless the application wishes to examine
951 extensions in the OCSP response for example.
953 Replace nonce routines with a pair of functions.
954 OCSP_request_add1_nonce() adds a nonce value and optionally
955 generates a random value. OCSP_check_nonce() checks the
956 validity of the nonce in an OCSP response.
959 +) Change function OCSP_request_add() to OCSP_request_add0_id().
960 This doesn't copy the supplied OCSP_CERTID and avoids the
961 need to free up the newly created id. Change return type
962 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
963 This can then be used to add extensions to the request.
964 Deleted OCSP_request_new(), since most of its functionality
965 is now in OCSP_REQUEST_new() (and the case insensitive name
966 clash) apart from the ability to set the request name which
967 will be added elsewhere.
970 +) Update OCSP API. Remove obsolete extensions argument from
971 various functions. Extensions are now handled using the new
972 OCSP extension code. New simple OCSP HTTP function which
973 can be used to send requests and parse the response.
976 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
977 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
978 uses the special reorder version of SET OF to sort the attributes
979 and reorder them to match the encoded order. This resolves a long
980 standing problem: a verify on a PKCS7 structure just after signing
981 it used to fail because the attribute order did not match the
982 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
983 it uses the received order. This is necessary to tolerate some broken
984 software that does not order SET OF. This is handled by encoding
985 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
986 to produce the required SET OF.
989 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
990 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
991 files to get correct declarations of the ASN.1 item variables.
994 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
995 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
996 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
997 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
998 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
999 ASN1_ITEM and no wrapper functions.
1002 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1003 replace the old function pointer based I/O routines. Change most of
1004 the *_d2i_bio() and *_d2i_fp() functions to use these.
1007 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1008 lines, recognice more "algorithms" that can be deselected, and make
1009 it complain about algorithm deselection that isn't recognised.
1012 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1013 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1014 to use new functions. Add NO_ASN1_OLD which can be set to remove
1015 some old style ASN1 functions: this can be used to determine if old
1016 code will still work when these eventually go away.
1019 +) New extension functions for OCSP structures, these follow the
1020 same conventions as certificates and CRLs.
1023 +) New function X509V3_add1_i2d(). This automatically encodes and
1024 adds an extension. Its behaviour can be customised with various
1025 flags to append, replace or delete. Various wrappers added for
1026 certifcates and CRLs.
1029 +) Fix to avoid calling the underlying ASN1 print routine when
1030 an extension cannot be parsed. Correct a typo in the
1031 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1034 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1035 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1036 when writing a 32767 byte record.
1037 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1039 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1040 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1042 (RSA objects have a reference count access to which is protected
1043 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1044 so they are meant to be shared between threads.)
1045 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1046 "Reddie, Steven" <Steven.Reddie@ca.com>]
1048 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1049 entries for variables.
1052 *) Fix a deadlock in CRYPTO_mem_leaks().
1055 +) Add functionality to apps/openssl.c for detecting locking
1056 problems: As the program is single-threaded, all we have
1057 to do is register a locking callback using an array for
1058 storing which locks are currently held by the program.
1061 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1062 SSL_get_ex_data_X509_STORE_idx(), which is used in
1063 ssl_verify_cert_chain() and thus can be called at any time
1064 during TLS/SSL handshakes so that thread-safety is essential.
1065 Unfortunately, the ex_data design is not at all suited
1066 for multi-threaded use, so it probably should be abolished.
1069 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1070 [Broadcom, tweaked and integrated by Geoff Thorpe]
1072 +) Move common extension printing code to new function
1073 X509V3_print_extensions(). Reorganise OCSP print routines and
1074 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1077 +) New function X509_signature_print() to remove duplication in some
1081 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1082 set (this was treated exactly the same as SET OF previously). This
1083 is used to reorder the STACK representing the structure to match the
1084 encoding. This will be used to get round a problem where a PKCS7
1085 structure which was signed could not be verified because the STACK
1086 order did not reflect the encoded order.
1089 +) Reimplement the OCSP ASN1 module using the new code.
1092 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1093 for its ASN1 operations. The old style function pointers still exist
1094 for now but they will eventually go away.
1097 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1098 completely replaces the old ASN1 functionality with a table driven
1099 encoder and decoder which interprets an ASN1_ITEM structure describing
1100 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1101 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1102 has also been converted to the new form.
1105 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1106 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1107 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1108 for negative moduli.
1111 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1112 of not touching the result's sign bit.
1115 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1119 +) Changed the LHASH code to use prototypes for callbacks, and created
1120 macros to declare and implement thin (optionally static) functions
1121 that provide type-safety and avoid function pointer casting for the
1122 type-specific callbacks.
1125 *) Use better test patterns in bntest.
1128 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1130 [Veers Staats <staatsvr@asc.hpc.mil>,
1131 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1133 *) rand_win.c fix for Borland C.
1136 *) BN_rshift bugfix for n == 0.
1139 +) Reformat the FAQ so the different questions and answers can be divided
1140 in sections depending on the subject.
1143 +) Have the zlib compression code load ZLIB.DLL dynamically under
1147 +) New function BN_mod_sqrt for computing square roots modulo a prime
1148 (using the probabilistic Tonelli-Shanks algorithm unless
1149 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1150 be handled deterministically).
1151 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1153 *) Add a 'bctest' script that checks for some known 'bc' bugs
1154 so that 'make test' does not abort just because 'bc' is broken.
1157 *) Store verify_result within SSL_SESSION also for client side to
1158 avoid potential security hole. (Re-used sessions on the client side
1159 always resulted in verify_result==X509_V_OK, not using the original
1160 result of the server certificate verification.)
1163 +) Make BN_mod_inverse faster by explicitly handling small quotients
1164 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1165 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1168 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1169 SSL3_RT_APPLICATION_DATA, return 0.
1170 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1174 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1175 releases, have been re-implemented by renaming the previous
1176 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1177 and ssl3_read_internal, respectively, and adding 'peek' parameters
1178 to them. The new ssl[23]_{read,peek} functions are calls to
1179 ssl[23]_read_internal with the 'peek' flag set appropriately.
1180 A 'peek' parameter has also been added to ssl3_read_bytes, which
1181 does the actual work for ssl3_read_internal.
1184 +) New function BN_kronecker.
1187 +) Fix BN_gcd so that it works on negative inputs; the result is
1188 positive unless both parameters are zero.
1189 Previously something reasonably close to an infinite loop was
1190 possible because numbers could be growing instead of shrinking
1191 in the implementation of Euclid's algorithm.
1194 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1195 sign of the number in question.
1197 Fix BN_is_word(a,w) to work correctly for w == 0.
1199 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1200 because its test if the absolute value of 'a' equals 'w'.
1201 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1202 it exists mostly for use in the implementations of BN_is_zero(),
1203 BN_is_one(), and BN_is_word().
1206 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1207 the method-specific "init()" handler. Also clean up ex_data after
1208 calling the method-specific "finish()" handler. Previously, this was
1209 happening the other way round.
1212 +) New function BN_swap.
1215 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1216 the exponentiation functions are more likely to produce reasonable
1217 results on negative inputs.
1220 +) Change BN_mod_mul so that the result is always non-negative.
1221 Previously, it could be negative if one of the factors was negative;
1222 I don't think anyone really wanted that behaviour.
1225 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1226 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1227 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1228 and add new functions:
1237 BN_mod_lshift1_quick
1241 These functions always generate non-negative results.
1243 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1244 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1246 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1247 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1248 be reduced modulo m.
1249 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1251 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1252 was actually never needed) and in BN_mul(). The removal in BN_mul()
1253 required a small change in bn_mul_part_recursive() and the addition
1254 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1255 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1256 bn_sub_words() and bn_add_words() except they take arrays with
1260 +) In 'openssl passwd', verify passwords read from the terminal
1261 unless the '-salt' option is used (which usually means that
1262 verification would just waste user's time since the resulting
1263 hash is going to be compared with some given password hash)
1264 or the new '-noverify' option is used.
1266 This is an incompatible change, but it does not affect
1267 non-interactive use of 'openssl passwd' (passwords on the command
1268 line, '-stdin' option, '-in ...' option) and thus should not
1272 +) Remove all references to RSAref, since there's no more need for it.
1275 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1276 The previous value, 12, was not always sufficient for BN_mod_exp().
1279 +) Make DSO load along a path given through an environment variable
1280 (SHLIB_PATH) with shl_load().
1283 +) Constify the ENGINE code as a result of BIGNUM constification.
1284 Also constify the RSA code and most things related to it. In a
1285 few places, most notable in the depth of the ASN.1 code, ugly
1286 casts back to non-const were required (to be solved at a later
1290 +) Make it so the openssl application has all engines loaded by default.
1293 +) Constify the BIGNUM routines a little more.
1296 *) Make sure that shared libraries get the internal name engine with
1297 the full version number and not just 0. This should mark the
1298 shared libraries as not backward compatible. Of course, this should
1299 be changed again when we can guarantee backward binary compatibility.
1302 +) Add the following functions:
1304 ENGINE_load_cswift()
1306 ENGINE_load_atalla()
1308 ENGINE_load_builtin_engines()
1310 That way, an application can itself choose if external engines that
1311 are built-in in OpenSSL shall ever be used or not. The benefit is
1312 that applications won't have to be linked with libdl or other dso
1313 libraries unless it's really needed.
1315 Changed 'openssl engine' to load all engines on demand.
1316 Changed the engine header files to avoid the duplication of some
1317 declarations (they differed!).
1320 +) 'openssl engine' can now list capabilities.
1323 +) Better error reporting in 'openssl engine'.
1326 +) Never call load_dh_param(NULL) in s_server.
1329 +) Add engine application. It can currently list engines by name and
1330 identity, and test if they are actually available.
1333 +) Add support for shared libraries under Irix.
1334 [Albert Chin-A-Young <china@thewrittenword.com>]
1336 +) Improve RPM specification file by forcing symbolic linking and making
1337 sure the installed documentation is also owned by root.root.
1338 [Damien Miller <djm@mindrot.org>]
1340 +) Add configuration option to build on Linux on both big-endian and
1342 [Ralf Baechle <ralf@uni-koblenz.de>]
1344 +) Give the OpenSSL applications more possibilities to make use of
1345 keys (public as well as private) handled by engines.
1348 +) Add OCSP code that comes from CertCo.
1351 +) Add VMS support for the Rijndael code.
1354 +) Added untested support for Nuron crypto accelerator.
1357 +) Add support for external cryptographic devices. This code was
1358 previously distributed separately as the "engine" branch.
1359 [Geoff Thorpe, Richard Levitte]
1361 +) Rework the filename-translation in the DSO code. It is now possible to
1362 have far greater control over how a "name" is turned into a filename
1363 depending on the operating environment and any oddities about the
1364 different shared library filenames on each system.
1367 +) Support threads on FreeBSD-elf in Configure.
1370 +) Add the possibility to create shared libraries on HP-UX
1373 +) Fix for SHA1 assembly problem with MASM: it produces
1374 warnings about corrupt line number information when assembling
1375 with debugging information. This is caused by the overlapping
1377 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1380 NCONF_get_number() has no error checking at all. As a replacement,
1381 NCONF_get_number_e() is defined (_e for "error checking") and is
1382 promoted strongly. The old NCONF_get_number is kept around for
1383 binary backward compatibility.
1384 Make it possible for methods to load from something other than a BIO,
1385 by providing a function pointer that is given a name instead of a BIO.
1386 For example, this could be used to load configuration data from an
1390 *) Fix typo in get_cert_by_subject() in by_dir.c
1391 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1393 *) Rework the system to generate shared libraries:
1395 - Make note of the expected extension for the shared libraries and
1396 if there is a need for symbolic links from for example libcrypto.so.0
1397 to libcrypto.so.0.9.7. There is extended info in Configure for
1400 - Make as few rebuilds of the shared libraries as possible.
1402 - Still avoid linking the OpenSSL programs with the shared libraries.
1404 - When installing, install the shared libraries separately from the
1408 +) Fix for non blocking accept BIOs. Added new I/O special reason
1409 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1410 with non blocking I/O was not possible because no retry code was
1411 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1415 +) Added the beginnings of Rijndael support.
1418 +) Fix for bug in DirectoryString mask setting. Add support for
1419 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1420 to allow certificate printing to more controllable, additional
1421 'certopt' option to 'x509' to allow new printing options to be
1425 +) Clean old EAY MD5 hack from e_os.h.
1428 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1430 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1431 and not in SSL_clear because the latter is also used by the
1432 accept/connect functions; previously, the settings made by
1433 SSL_set_read_ahead would be lost during the handshake.
1434 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1436 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1437 Previously, it would create entries for disableed algorithms no
1441 *) Added several new manual pages for SSL_* function.
1444 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1446 *) In ssl23_get_client_hello, generate an error message when faced
1447 with an initial SSL 3.0/TLS record that is too small to contain the
1448 first two bytes of the ClientHello message, i.e. client_version.
1449 (Note that this is a pathologic case that probably has never happened
1450 in real life.) The previous approach was to use the version number
1451 from the record header as a substitute; but our protocol choice
1452 should not depend on that one because it is not authenticated
1453 by the Finished messages.
1456 *) More robust randomness gathering functions for Windows.
1457 [Jeffrey Altman <jaltman@columbia.edu>]
1459 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1460 not set then we don't setup the error code for issuer check errors
1461 to avoid possibly overwriting other errors which the callback does
1462 handle. If an application does set the flag then we assume it knows
1463 what it is doing and can handle the new informational codes
1467 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1468 a general "ANY" type, as such it should be able to decode anything
1469 including tagged types. However it didn't check the class so it would
1470 wrongly interpret tagged types in the same way as their universal
1471 counterpart and unknown types were just rejected. Changed so that the
1472 tagged and unknown types are handled in the same way as a SEQUENCE:
1473 that is the encoding is stored intact. There is also a new type
1474 "V_ASN1_OTHER" which is used when the class is not universal, in this
1475 case we have no idea what the actual type is so we just lump them all
1479 *) On VMS, stdout may very well lead to a file that is written to
1480 in a record-oriented fashion. That means that every write() will
1481 write a separate record, which will be read separately by the
1482 programs trying to read from it. This can be very confusing.
1484 The solution is to put a BIO filter in the way that will buffer
1485 text until a linefeed is reached, and then write everything a
1486 line at a time, so every record written will be an actual line,
1487 not chunks of lines and not (usually doesn't happen, but I've
1488 seen it once) several lines in one record. BIO_f_linebuffer() is
1491 Currently, it's a VMS-only method, because that's where it has
1492 been tested well enough.
1495 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1496 it can return incorrect results.
1497 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1498 but it was in 0.9.6-beta[12].)
1501 *) Disable the check for content being present when verifying detached
1502 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1503 include zero length content when signing messages.
1506 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1507 BIO_ctrl (for BIO pairs).
1510 *) Add DSO method for VMS.
1513 *) Bug fix: Montgomery multiplication could produce results with the
1517 *) Add RPM specification openssl.spec and modify it to build three
1518 packages. The default package contains applications, application
1519 documentation and run-time libraries. The devel package contains
1520 include files, static libraries and function documentation. The
1521 doc package contains the contents of the doc directory. The original
1522 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1525 *) Add a large number of documentation files for many SSL routines.
1526 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1528 *) Add a configuration entry for Sony News 4.
1529 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1531 *) Don't set the two most significant bits to one when generating a
1532 random number < q in the DSA library.
1535 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1536 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1537 the underlying transport is blocking) if a handshake took place.
1538 (The default behaviour is needed by applications such as s_client
1539 and s_server that use select() to determine when to use SSL_read;
1540 but for applications that know in advance when to expect data, it
1541 just makes things more complicated.)
1544 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1548 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1549 work better on such systems.
1550 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1552 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1553 Update PKCS12_parse() so it copies the friendlyName and the
1554 keyid to the certificates aux info.
1557 *) Fix bug in PKCS7_verify() which caused an infinite loop
1558 if there was more than one signature.
1559 [Sven Uszpelkat <su@celocom.de>]
1561 *) Major change in util/mkdef.pl to include extra information
1562 about each symbol, as well as presentig variables as well
1563 as functions. This change means that there's n more need
1564 to rebuild the .num files when some algorithms are excluded.
1567 *) Allow the verify time to be set by an application,
1568 rather than always using the current time.
1571 *) Phase 2 verify code reorganisation. The certificate
1572 verify code now looks up an issuer certificate by a
1573 number of criteria: subject name, authority key id
1574 and key usage. It also verifies self signed certificates
1575 by the same criteria. The main comparison function is
1576 X509_check_issued() which performs these checks.
1578 Lot of changes were necessary in order to support this
1579 without completely rewriting the lookup code.
1581 Authority and subject key identifier are now cached.
1583 The LHASH 'certs' is X509_STORE has now been replaced
1584 by a STACK_OF(X509_OBJECT). This is mainly because an
1585 LHASH can't store or retrieve multiple objects with
1586 the same hash value.
1588 As a result various functions (which were all internal
1589 use only) have changed to handle the new X509_STORE
1590 structure. This will break anything that messed round
1591 with X509_STORE internally.
1593 The functions X509_STORE_add_cert() now checks for an
1594 exact match, rather than just subject name.
1596 The X509_STORE API doesn't directly support the retrieval
1597 of multiple certificates matching a given criteria, however
1598 this can be worked round by performing a lookup first
1599 (which will fill the cache with candidate certificates)
1600 and then examining the cache for matches. This is probably
1601 the best we can do without throwing out X509_LOOKUP
1602 entirely (maybe later...).
1604 The X509_VERIFY_CTX structure has been enhanced considerably.
1606 All certificate lookup operations now go via a get_issuer()
1607 callback. Although this currently uses an X509_STORE it
1608 can be replaced by custom lookups. This is a simple way
1609 to bypass the X509_STORE hackery necessary to make this
1610 work and makes it possible to use more efficient techniques
1611 in future. A very simple version which uses a simple
1612 STACK for its trusted certificate store is also provided
1613 using X509_STORE_CTX_trusted_stack().
1615 The verify_cb() and verify() callbacks now have equivalents
1616 in the X509_STORE_CTX structure.
1618 X509_STORE_CTX also has a 'flags' field which can be used
1619 to customise the verify behaviour.
1622 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1623 excludes S/MIME capabilities.
1626 *) When a certificate request is read in keep a copy of the
1627 original encoding of the signed data and use it when outputing
1628 again. Signatures then use the original encoding rather than
1629 a decoded, encoded version which may cause problems if the
1630 request is improperly encoded.
1633 *) For consistency with other BIO_puts implementations, call
1634 buffer_write(b, ...) directly in buffer_puts instead of calling
1637 In BIO_puts, increment b->num_write as in BIO_write.
1638 [Peter.Sylvester@EdelWeb.fr]
1640 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1641 BN_zero, we may not return a BIGNUM with an array consisting of
1645 *) Avoid calling abort() from within the library when problems are
1646 detected, except if preprocessor symbols have been defined
1647 (such as REF_CHECK, BN_DEBUG etc.).
1650 *) New openssl application 'rsautl'. This utility can be
1651 used for low level RSA operations. DER public key
1652 BIO/fp routines also added.
1655 *) New Configure entry and patches for compiling on QNX 4.
1656 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1658 *) A demo state-machine implementation was sponsored by
1659 Nuron (http://www.nuron.com/) and is now available in
1660 demos/state_machine.
1663 *) New options added to the 'dgst' utility for signature
1664 generation and verification.
1667 *) Unrecognized PKCS#7 content types are now handled via a
1668 catch all ASN1_TYPE structure. This allows unsupported
1669 types to be stored as a "blob" and an application can
1670 encode and decode it manually.
1673 *) Fix various signed/unsigned issues to make a_strex.c
1675 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1677 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1678 length if passed a buffer. ASN1_INTEGER_to_BN failed
1679 if passed a NULL BN and its argument was negative.
1680 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1682 *) Modification to PKCS#7 encoding routines to output definite
1683 length encoding. Since currently the whole structures are in
1684 memory there's not real point in using indefinite length
1685 constructed encoding. However if OpenSSL is compiled with
1686 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1689 *) Added BIO_vprintf() and BIO_vsnprintf().
1692 *) Added more prefixes to parse for in the the strings written
1693 through a logging bio, to cover all the levels that are available
1694 through syslog. The prefixes are now:
1696 PANIC, EMERG, EMR => LOG_EMERG
1697 ALERT, ALR => LOG_ALERT
1698 CRIT, CRI => LOG_CRIT
1699 ERROR, ERR => LOG_ERR
1700 WARNING, WARN, WAR => LOG_WARNING
1701 NOTICE, NOTE, NOT => LOG_NOTICE
1702 INFO, INF => LOG_INFO
1703 DEBUG, DBG => LOG_DEBUG
1705 and as before, if none of those prefixes are present at the
1706 beginning of the string, LOG_ERR is chosen.
1708 On Win32, the LOG_* levels are mapped according to this:
1710 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1711 LOG_WARNING => EVENTLOG_WARNING_TYPE
1712 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1716 *) Made it possible to reconfigure with just the configuration
1717 argument "reconf" or "reconfigure". The command line arguments
1718 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1719 and are retrieved from there when reconfiguring.
1723 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1725 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1728 *) The obj_dat.pl script was messing up the sorting of object
1729 names. The reason was that it compared the quoted version
1730 of strings as a result "OCSP" > "OCSP Signing" because
1731 " > SPACE. Changed script to store unquoted versions of
1732 names and add quotes on output. It was also omitting some
1733 names from the lookup table if they were given a default
1734 value (that is if SN is missing it is given the same
1735 value as LN and vice versa), these are now added on the
1736 grounds that if an object has a name we should be able to
1737 look it up. Finally added warning output when duplicate
1738 short or long names are found.
1741 *) Changes needed for Tandem NSK.
1742 [Scott Uroff <scott@xypro.com>]
1744 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1745 RSA_padding_check_SSLv23(), special padding was never detected
1746 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1747 version rollback attacks was not effective.
1749 In s23_clnt.c, don't use special rollback-attack detection padding
1750 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1751 client; similarly, in s23_srvr.c, don't do the rollback check if
1752 SSL 2.0 is the only protocol enabled in the server.
1755 *) Make it possible to get hexdumps of unprintable data with 'openssl
1756 asn1parse'. By implication, the functions ASN1_parse_dump() and
1757 BIO_dump_indent() are added.
1760 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1761 these print out strings and name structures based on various
1762 flags including RFC2253 support and proper handling of
1763 multibyte characters. Added options to the 'x509' utility
1764 to allow the various flags to be set.
1767 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1768 Also change the functions X509_cmp_current_time() and
1769 X509_gmtime_adj() work with an ASN1_TIME structure,
1770 this will enable certificates using GeneralizedTime in validity
1771 dates to be checked.
1774 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1775 negative public key encodings) on by default,
1776 NO_NEG_PUBKEY_BUG can be set to disable it.
1779 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1780 content octets. An i2c_ASN1_OBJECT is unnecessary because
1781 the encoding can be trivially obtained from the structure.
1784 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1785 not read locks (CRYPTO_r_[un]lock).
1788 *) A first attempt at creating official support for shared
1789 libraries through configuration. I've kept it so the
1790 default is static libraries only, and the OpenSSL programs
1791 are always statically linked for now, but there are
1792 preparations for dynamic linking in place.
1793 This has been tested on Linux and True64.
1796 *) Randomness polling function for Win9x, as described in:
1797 Peter Gutmann, Software Generation of Practically Strong
1801 *) Fix so PRNG is seeded in req if using an already existing
1805 *) New options to smime application. -inform and -outform
1806 allow alternative formats for the S/MIME message including
1807 PEM and DER. The -content option allows the content to be
1808 specified separately. This should allow things like Netscape
1809 form signing output easier to verify.
1812 *) Fix the ASN1 encoding of tags using the 'long form'.
1815 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1816 STRING types. These convert content octets to and from the
1817 underlying type. The actual tag and length octets are
1818 already assumed to have been read in and checked. These
1819 are needed because all other string types have virtually
1820 identical handling apart from the tag. By having versions
1821 of the ASN1 functions that just operate on content octets
1822 IMPLICIT tagging can be handled properly. It also allows
1823 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1824 and ASN1_INTEGER are identical apart from the tag.
1827 *) Change the handling of OID objects as follows:
1829 - New object identifiers are inserted in objects.txt, following
1830 the syntax given in objects.README.
1831 - objects.pl is used to process obj_mac.num and create a new
1833 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1836 This is currently kind of a hack, and the perl code in objects.pl
1837 isn't very elegant, but it works as I intended. The simplest way
1838 to check that it worked correctly is to look in obj_dat.h and
1839 check the array nid_objs and make sure the objects haven't moved
1840 around (this is important!). Additions are OK, as well as
1841 consistent name changes.
1844 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1847 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1848 The given file adds to whatever has already been seeded into the
1849 random pool through the RANDFILE configuration file option or
1850 environment variable, or the default random state file.
1853 *) mkstack.pl now sorts each macro group into lexical order.
1854 Previously the output order depended on the order the files
1855 appeared in the directory, resulting in needless rewriting
1859 *) Patches to make OpenSSL compile under Win32 again. Mostly
1860 work arounds for the VC++ problem that it treats func() as
1861 func(void). Also stripped out the parts of mkdef.pl that
1862 added extra typesafe functions: these no longer exist.
1865 *) Reorganisation of the stack code. The macros are now all
1866 collected in safestack.h . Each macro is defined in terms of
1867 a "stack macro" of the form SKM_<name>(type, a, b). The
1868 DEBUG_SAFESTACK is now handled in terms of function casts,
1869 this has the advantage of retaining type safety without the
1870 use of additional functions. If DEBUG_SAFESTACK is not defined
1871 then the non typesafe macros are used instead. Also modified the
1872 mkstack.pl script to handle the new form. Needs testing to see
1873 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1874 the default if no major problems. Similar behaviour for ASN1_SET_OF
1875 and PKCS12_STACK_OF.
1878 *) When some versions of IIS use the 'NET' form of private key the
1879 key derivation algorithm is different. Normally MD5(password) is
1880 used as a 128 bit RC4 key. In the modified case
1881 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1882 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1883 as the old Netscape_RSA functions except they have an additional
1884 'sgckey' parameter which uses the modified algorithm. Also added
1885 an -sgckey command line option to the rsa utility. Thanks to
1886 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1887 algorithm to openssl-dev.
1890 *) The evp_local.h macros were using 'c.##kname' which resulted in
1891 invalid expansion on some systems (SCO 5.0.5 for example).
1892 Corrected to 'c.kname'.
1893 [Phillip Porch <root@theporch.com>]
1895 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1896 a STACK of email addresses from a certificate or request, these look
1897 in the subject name and the subject alternative name extensions and
1898 omit any duplicate addresses.
1901 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1902 This makes DSA verification about 2 % faster.
1905 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1906 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1907 plus overhead for 1024 bit moduli).
1908 This makes exponentiations about 0.5 % faster for 1024 bit
1909 exponents (as measured by "openssl speed rsa2048").
1912 *) Rename memory handling macros to avoid conflicts with other
1914 Malloc => OPENSSL_malloc
1915 Malloc_locked => OPENSSL_malloc_locked
1916 Realloc => OPENSSL_realloc
1917 Free => OPENSSL_free
1920 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1921 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1924 *) CygWin32 support.
1925 [John Jarvie <jjarvie@newsguy.com>]
1927 *) The type-safe stack code has been rejigged. It is now only compiled
1928 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1929 by default all type-specific stack functions are "#define"d back to
1930 standard stack functions. This results in more streamlined output
1931 but retains the type-safety checking possibilities of the original
1935 *) The STACK code has been cleaned up, and certain type declarations
1936 that didn't make a lot of sense have been brought in line. This has
1937 also involved a cleanup of sorts in safestack.h to more correctly
1938 map type-safe stack functions onto their plain stack counterparts.
1939 This work has also resulted in a variety of "const"ifications of
1940 lots of the code, especially "_cmp" operations which should normally
1941 be prototyped with "const" parameters anyway.
1944 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1945 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1946 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1947 where all of 'md' is used each time the PRNG is used, but 'state'
1948 is used only indexed by a cyclic counter. As entropy may not be
1949 well distributed from the beginning, 'md' is important as a
1950 chaining variable. However, the output function chains only half
1951 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1952 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1953 in all of 'state' being rewritten, with the new values depending
1954 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1957 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1958 the handshake is continued after ssl_verify_cert_chain();
1959 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1960 can lead to 'unexplainable' connection aborts later.
1961 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1963 *) Major EVP API cipher revision.
1964 Add hooks for extra EVP features. This allows various cipher
1965 parameters to be set in the EVP interface. Support added for variable
1966 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1967 setting of RC2 and RC5 parameters.
1969 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1972 Remove lots of duplicated code from the EVP library. For example *every*
1973 cipher init() function handles the 'iv' in the same way according to the
1974 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1975 for CFB and OFB modes they zero ctx->num.
1977 New functionality allows removal of S/MIME code RC2 hack.
1979 Most of the routines have the same form and so can be declared in terms
1982 By shifting this to the top level EVP_CipherInit() it can be removed from
1983 all individual ciphers. If the cipher wants to handle IVs or keys
1984 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1987 Change lots of functions like EVP_EncryptUpdate() to now return a
1988 value: although software versions of the algorithms cannot fail
1989 any installed hardware versions can.
1992 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1993 this option is set, tolerate broken clients that send the negotiated
1994 protocol version number instead of the requested protocol version
1998 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1999 i.e. non-zero for export ciphersuites, zero otherwise.
2000 Previous versions had this flag inverted, inconsistent with
2001 rsa_tmp_cb (..._TMP_RSA_CB).
2002 [Bodo Moeller; problem reported by Amit Chopra]
2004 *) Add missing DSA library text string. Work around for some IIS
2005 key files with invalid SEQUENCE encoding.
2008 *) Add a document (doc/standards.txt) that list all kinds of standards
2009 and so on that are implemented in OpenSSL.
2012 *) Enhance c_rehash script. Old version would mishandle certificates
2013 with the same subject name hash and wouldn't handle CRLs at all.
2014 Added -fingerprint option to crl utility, to support new c_rehash
2018 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2021 *) Fix for SSL server purpose checking. Server checking was
2022 rejecting certificates which had extended key usage present
2023 but no ssl client purpose.
2024 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2026 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2027 is a little unclear about how a blank password is handled.
2028 Since the password in encoded as a BMPString with terminating
2029 double NULL a zero length password would end up as just the
2030 double NULL. However no password at all is different and is
2031 handled differently in the PKCS#12 key generation code. NS
2032 treats a blank password as zero length. MSIE treats it as no
2033 password on export: but it will try both on import. We now do
2034 the same: PKCS12_parse() tries zero length and no password if
2035 the password is set to "" or NULL (NULL is now a valid password:
2036 it wasn't before) as does the pkcs12 application.
2039 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2040 perror when PEM_read_bio_X509_REQ fails, the error message must
2041 be obtained from the error queue.
2044 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2045 it in ERR_remove_state if appropriate, and change ERR_get_state
2046 accordingly to avoid race conditions (this is necessary because
2047 thread_hash is no longer constant once set).
2050 *) Bugfix for linux-elf makefile.one.
2053 *) RSA_get_default_method() will now cause a default
2054 RSA_METHOD to be chosen if one doesn't exist already.
2055 Previously this was only set during a call to RSA_new()
2056 or RSA_new_method(NULL) meaning it was possible for
2057 RSA_get_default_method() to return NULL.
2060 *) Added native name translation to the existing DSO code
2061 that will convert (if the flag to do so is set) filenames
2062 that are sufficiently small and have no path information
2063 into a canonical native form. Eg. "blah" converted to
2064 "libblah.so" or "blah.dll" etc.
2067 *) New function ERR_error_string_n(e, buf, len) which is like
2068 ERR_error_string(e, buf), but writes at most 'len' bytes
2069 including the 0 terminator. For ERR_error_string_n, 'buf'
2071 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2073 *) CONF library reworked to become more general. A new CONF
2074 configuration file reader "class" is implemented as well as a
2075 new functions (NCONF_*, for "New CONF") to handle it. The now
2076 old CONF_* functions are still there, but are reimplemented to
2077 work in terms of the new functions. Also, a set of functions
2078 to handle the internal storage of the configuration data is
2079 provided to make it easier to write new configuration file
2080 reader "classes" (I can definitely see something reading a
2081 configuration file in XML format, for example), called _CONF_*,
2082 or "the configuration storage API"...
2084 The new configuration file reading functions are:
2086 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2087 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2089 NCONF_default, NCONF_WIN32
2091 NCONF_dump_fp, NCONF_dump_bio
2093 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2094 NCONF_new creates a new CONF object. This works in the same way
2095 as other interfaces in OpenSSL, like the BIO interface.
2096 NCONF_dump_* dump the internal storage of the configuration file,
2097 which is useful for debugging. All other functions take the same
2098 arguments as the old CONF_* functions wth the exception of the
2099 first that must be a `CONF *' instead of a `LHASH *'.
2101 To make it easer to use the new classes with the old CONF_* functions,
2102 the function CONF_set_default_method is provided.
2105 *) Add '-tls1' option to 'openssl ciphers', which was already
2106 mentioned in the documentation but had not been implemented.
2107 (This option is not yet really useful because even the additional
2108 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2111 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2112 OpenSSL-based applications) load shared libraries and bind to
2113 them in a portable way.
2114 [Geoff Thorpe, with contributions from Richard Levitte]
2116 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2118 *) Make sure _lrotl and _lrotr are only used with MSVC.
2120 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2121 (the default implementation of RAND_status).
2123 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2124 to '-clrext' (= clear extensions), as intended and documented.
2125 [Bodo Moeller; inconsistency pointed out by Michael Attili
2128 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2129 was larger than the MD block size.
2130 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2132 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2133 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2134 using the passed key: if the passed key was a private key the result
2135 of X509_print(), for example, would be to print out all the private key
2139 *) des_quad_cksum() byte order bug fix.
2140 [Ulf Möller, using the problem description in krb4-0.9.7, where
2141 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2143 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2145 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2147 *) For easily testing in shell scripts whether some command
2148 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2149 returns with exit code 0 iff no command of the given name is available.
2150 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2151 the output goes to stdout and nothing is printed to stderr.
2152 Additional arguments are always ignored.
2154 Since for each cipher there is a command of the same name,
2155 the 'no-cipher' compilation switches can be tested this way.
2157 ('openssl no-XXX' is not able to detect pseudo-commands such
2158 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2161 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2164 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2165 is set; it will be thrown away anyway because each handshake creates
2167 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2168 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2169 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2170 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2173 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2174 'Q' and 'R' lose their special meanings (quit/renegotiate).
2175 This is part of what -quiet does; unlike -quiet, -ign_eof
2176 does not suppress any output.
2179 *) Add compatibility options to the purpose and trust code. The
2180 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2181 accepts a certificate or CA, this was the previous behaviour,
2182 with all the associated security issues.
2184 X509_TRUST_COMPAT is the old trust behaviour: only and
2185 automatically trust self signed roots in certificate store. A
2186 new trust setting X509_TRUST_DEFAULT is used to specify that
2187 a purpose has no associated trust setting and it should instead
2188 use the value in the default purpose.
2191 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2192 and fix a memory leak.
2195 *) In util/mkerr.pl (which implements 'make errors'), preserve
2196 reason strings from the previous version of the .c file, as
2197 the default to have only downcase letters (and digits) in
2198 automatically generated reasons codes is not always appropriate.
2201 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2202 using strerror. Previously, ERR_reason_error_string() returned
2203 library names as reason strings for SYSerr; but SYSerr is a special
2204 case where small numbers are errno values, not library numbers.
2207 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2208 converts DSA parameters into DH parameters. (When creating parameters,
2209 DSA_generate_parameters is used.)
2212 *) Include 'length' (recommended exponent length) in C code generated
2213 by 'openssl dhparam -C'.
2216 *) The second argument to set_label in perlasm was already being used
2217 so couldn't be used as a "file scope" flag. Moved to third argument
2221 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2222 instead of RAND_bytes for encryption IVs and salts.
2225 *) Include RAND_status() into RAND_METHOD instead of implementing
2226 it only for md_rand.c Otherwise replacing the PRNG by calling
2227 RAND_set_rand_method would be impossible.
2230 *) Don't let DSA_generate_key() enter an infinite loop if the random
2231 number generation fails.
2234 *) New 'rand' application for creating pseudo-random output.
2237 *) Added configuration support for Linux/IA64
2238 [Rolf Haberrecker <rolf@suse.de>]
2240 *) Assembler module support for Mingw32.
2243 *) Shared library support for HPUX (in shlib/).
2244 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2246 *) Shared library support for Solaris gcc.
2247 [Lutz Behnke <behnke@trustcenter.de>]
2249 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2251 *) PKCS7_encrypt() was adding text MIME headers twice because they
2252 were added manually and by SMIME_crlf_copy().
2255 *) In bntest.c don't call BN_rand with zero bits argument.
2256 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2258 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2259 case was implemented. This caused BN_div_recp() to fail occasionally.
2262 *) Add an optional second argument to the set_label() in the perl
2263 assembly language builder. If this argument exists and is set
2264 to 1 it signals that the assembler should use a symbol whose
2265 scope is the entire file, not just the current function. This
2266 is needed with MASM which uses the format label:: for this scope.
2267 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2269 *) Change the ASN1 types so they are typedefs by default. Before
2270 almost all types were #define'd to ASN1_STRING which was causing
2271 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2275 *) Change names of new functions to the new get1/get0 naming
2276 convention: After 'get1', the caller owns a reference count
2277 and has to call ..._free; 'get0' returns a pointer to some
2278 data structure without incrementing reference counters.
2279 (Some of the existing 'get' functions increment a reference
2280 counter, some don't.)
2281 Similarly, 'set1' and 'add1' functions increase reference
2282 counters or duplicate objects.
2285 *) Allow for the possibility of temp RSA key generation failure:
2286 the code used to assume it always worked and crashed on failure.
2289 *) Fix potential buffer overrun problem in BIO_printf().
2290 [Ulf Möller, using public domain code by Patrick Powell; problem
2291 pointed out by David Sacerdote <das33@cornell.edu>]
2293 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2294 RAND_egd() and RAND_status(). In the command line application,
2295 the EGD socket can be specified like a seed file using RANDFILE
2299 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2300 Some CAs (e.g. Verisign) distribute certificates in this form.
2303 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2304 list to exclude them. This means that no special compilation option
2305 is needed to use anonymous DH: it just needs to be included in the
2309 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2310 EVP_MD_type. The old functionality is available in a new macro called
2311 EVP_MD_md(). Change code that uses it and update docs.
2314 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2315 where the 'void *' argument is replaced by a function pointer argument.
2316 Previously 'void *' was abused to point to functions, which works on
2317 many platforms, but is not correct. As these functions are usually
2318 called by macros defined in OpenSSL header files, most source code
2319 should work without changes.
2322 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2323 sections with information on -D... compiler switches used for
2324 compiling the library so that applications can see them. To enable
2325 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2326 must be defined. E.g.,
2327 #define OPENSSL_ALGORITHM_DEFINES
2328 #include <openssl/opensslconf.h>
2329 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2330 [Richard Levitte, Ulf and Bodo Möller]
2332 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2336 *) Change the 'other' type in certificate aux info to a STACK_OF
2337 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2338 the required ASN1 format: arbitrary types determined by an OID.
2341 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2342 argument to 'req'. This is not because the function is newer or
2343 better than others it just uses the work 'NEW' in the certificate
2344 request header lines. Some software needs this.
2347 *) Reorganise password command line arguments: now passwords can be
2348 obtained from various sources. Delete the PEM_cb function and make
2349 it the default behaviour: i.e. if the callback is NULL and the
2350 usrdata argument is not NULL interpret it as a null terminated pass
2351 phrase. If usrdata and the callback are NULL then the pass phrase
2352 is prompted for as usual.
2355 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2356 the support is automatically enabled. The resulting binaries will
2357 autodetect the card and use it if present.
2358 [Ben Laurie and Compaq Inc.]
2360 *) Work around for Netscape hang bug. This sends certificate request
2361 and server done in one record. Since this is perfectly legal in the
2362 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2363 the bugs/SSLv3 entry for more info.
2366 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2369 *) Add -rand argument to smime and pkcs12 applications and read/write
2373 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2376 *) Add command line password options to the remaining applications.
2379 *) Bug fix for BN_div_recp() for numerators with an even number of
2383 *) More tests in bntest.c, and changed test_bn output.
2386 *) ./config recognizes MacOS X now.
2389 *) Bug fix for BN_div() when the first words of num and divsor are
2390 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2393 *) Add support for various broken PKCS#8 formats, and command line
2394 options to produce them.
2397 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2398 get temporary BIGNUMs from a BN_CTX.
2401 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2405 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2406 include a #define from the old name to the new. The original intent
2407 was that statically linked binaries could for example just call
2408 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2409 link with digests. This never worked becayse SSLeay_add_all_digests()
2410 and SSLeay_add_all_ciphers() were in the same source file so calling
2411 one would link with the other. They are now in separate source files.
2414 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2417 *) Use a less unusual form of the Miller-Rabin primality test (it used
2418 a binary algorithm for exponentiation integrated into the Miller-Rabin
2419 loop, our standard modexp algorithms are faster).
2422 *) Support for the EBCDIC character set completed.
2423 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2425 *) Source code cleanups: use const where appropriate, eliminate casts,
2426 use void * instead of char * in lhash.
2429 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2430 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2431 this the server could overwrite ephemeral keys that the client
2435 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2436 using 50 iterations of the Rabin-Miller test.
2438 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2439 iterations of the Rabin-Miller test as required by the appendix
2440 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2441 As BN_is_prime_fasttest includes trial division, DSA parameter
2442 generation becomes much faster.
2444 This implies a change for the callback functions in DSA_is_prime
2445 and DSA_generate_parameters: The callback function is called once
2446 for each positive witness in the Rabin-Miller test, not just
2447 occasionally in the inner loop; and the parameters to the
2448 callback function now provide an iteration count for the outer
2449 loop rather than for the current invocation of the inner loop.
2450 DSA_generate_parameters additionally can call the callback
2451 function with an 'iteration count' of -1, meaning that a
2452 candidate has passed the trial division test (when q is generated
2453 from an application-provided seed, trial division is skipped).
2456 *) New function BN_is_prime_fasttest that optionally does trial
2457 division before starting the Rabin-Miller test and has
2458 an additional BN_CTX * argument (whereas BN_is_prime always
2459 has to allocate at least one BN_CTX).
2460 'callback(1, -1, cb_arg)' is called when a number has passed the
2461 trial division stage.
2464 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2468 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2471 *) New function BN_pseudo_rand().
2474 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2475 bignum version of BN_from_montgomery() with the working code from
2476 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2480 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2481 made it impossible to use the same SSL_SESSION data structure in
2482 SSL2 clients in multiple threads.
2485 *) The return value of RAND_load_file() no longer counts bytes obtained
2486 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2487 to seed the PRNG (previously an explicit byte count was required).
2488 [Ulf Möller, Bodo Möller]
2490 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2491 used (char *) instead of (void *) and had casts all over the place.
2494 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2497 *) Retain source code compatibility for BN_prime_checks macro:
2498 BN_is_prime(..., BN_prime_checks, ...) now uses
2499 BN_prime_checks_for_size to determine the appropriate number of
2500 Rabin-Miller iterations.
2503 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2504 DH_CHECK_P_NOT_SAFE_PRIME.
2505 (Check if this is true? OpenPGP calls them "strong".)
2508 *) Merge the functionality of "dh" and "gendh" programs into a new program
2509 "dhparam". The old programs are retained for now but will handle DH keys
2510 (instead of parameters) in future.
2513 *) Make the ciphers, s_server and s_client programs check the return values
2514 when a new cipher list is set.
2517 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2518 ciphers. Before when the 56bit ciphers were enabled the sorting was
2521 The syntax for the cipher sorting has been extended to support sorting by
2522 cipher-strength (using the strength_bits hard coded in the tables).
2523 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2525 Fix a bug in the cipher-command parser: when supplying a cipher command
2526 string with an "undefined" symbol (neither command nor alphanumeric
2527 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2528 an error is flagged.
2530 Due to the strength-sorting extension, the code of the
2531 ssl_create_cipher_list() function was completely rearranged. I hope that
2532 the readability was also increased :-)
2533 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2535 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2536 for the first serial number and places 2 in the serial number file. This
2537 avoids problems when the root CA is created with serial number zero and
2538 the first user certificate has the same issuer name and serial number
2542 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2543 the new code. Add documentation for this stuff.
2546 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2547 X509_*() to X509at_*() on the grounds that they don't handle X509
2548 structures and behave in an analagous way to the X509v3 functions:
2549 they shouldn't be called directly but wrapper functions should be used
2552 So we also now have some wrapper functions that call the X509at functions
2553 when passed certificate requests. (TO DO: similar things can be done with
2554 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2555 things. Some of these need some d2i or i2d and print functionality
2556 because they handle more complex structures.)
2559 *) Add missing #ifndefs that caused missing symbols when building libssl
2560 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2561 NO_RSA in ssl/s2*.c.
2562 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2564 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2565 has a return value which indicates the quality of the random data
2566 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2567 error queue. New function RAND_pseudo_bytes() generates output that is
2568 guaranteed to be unique but not unpredictable. RAND_add is like
2569 RAND_seed, but takes an extra argument for an entropy estimate
2570 (RAND_seed always assumes full entropy).
2573 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2574 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2575 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2576 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2577 false-positive rate of at most 2^-80 for random input.
2580 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2583 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2584 in the 0.9.5 release), this returns the chain
2585 from an X509_CTX structure with a dup of the stack and all
2586 the X509 reference counts upped: so the stack will exist
2587 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2590 Also make SSL_SESSION_print() print out the verify return
2594 *) Add manpage for the pkcs12 command. Also change the default
2595 behaviour so MAC iteration counts are used unless the new
2596 -nomaciter option is used. This improves file security and
2597 only older versions of MSIE (4.0 for example) need it.
2600 *) Honor the no-xxx Configure options when creating .DEF files.
2603 *) Add PKCS#10 attributes to field table: challengePassword,
2604 unstructuredName and unstructuredAddress. These are taken from
2605 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2606 international characters are used.
2608 More changes to X509_ATTRIBUTE code: allow the setting of types
2609 based on strings. Remove the 'loc' parameter when adding
2610 attributes because these will be a SET OF encoding which is sorted
2614 *) Initial changes to the 'req' utility to allow request generation
2615 automation. This will allow an application to just generate a template
2616 file containing all the field values and have req construct the
2619 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2620 used all over the place including certificate requests and PKCS#7
2621 structures. They are currently handled manually where necessary with
2622 some primitive wrappers for PKCS#7. The new functions behave in a
2623 manner analogous to the X509 extension functions: they allow
2624 attributes to be looked up by NID and added.
2626 Later something similar to the X509V3 code would be desirable to
2627 automatically handle the encoding, decoding and printing of the
2628 more complex types. The string types like challengePassword can
2629 be handled by the string table functions.
2631 Also modified the multi byte string table handling. Now there is
2632 a 'global mask' which masks out certain types. The table itself
2633 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2634 is useful when for example there is only one permissible type
2635 (as in countryName) and using the mask might result in no valid
2639 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2640 SSL_get_peer_finished to allow applications to obtain the latest
2641 Finished messages sent to the peer or expected from the peer,
2642 respectively. (SSL_get_peer_finished is usually the Finished message
2643 actually received from the peer, otherwise the protocol will be aborted.)
2645 As the Finished message are message digests of the complete handshake
2646 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2647 be used for external authentication procedures when the authentication
2648 provided by SSL/TLS is not desired or is not enough.
2651 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2652 the host supports BWX extension and if Compaq C is present on the
2653 $PATH. Just exploiting of the BWX extension results in 20-30%
2654 performance kick for some algorithms, e.g. DES and RC4 to mention
2655 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2659 *) Add support for MS "fast SGC". This is arguably a violation of the
2660 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2661 weak crypto and after checking the certificate is SGC a second one
2662 with strong crypto. MS SGC stops the first handshake after receiving
2663 the server certificate message and sends a second client hello. Since
2664 a server will typically do all the time consuming operations before
2665 expecting any further messages from the client (server key exchange
2666 is the most expensive) there is little difference between the two.
2668 To get OpenSSL to support MS SGC we have to permit a second client
2669 hello message after we have sent server done. In addition we have to
2670 reset the MAC if we do get this second client hello.
2673 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2674 if a DER encoded private key is RSA or DSA traditional format. Changed
2675 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2676 format DER encoded private key. Newer code should use PKCS#8 format which
2677 has the key type encoded in the ASN1 structure. Added DER private key
2678 support to pkcs8 application.
2681 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2682 ciphersuites has been selected (as required by the SSL 3/TLS 1
2683 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2684 is set, we interpret this as a request to violate the specification
2685 (the worst that can happen is a handshake failure, and 'correct'
2686 behaviour would result in a handshake failure anyway).
2689 *) In SSL_CTX_add_session, take into account that there might be multiple
2690 SSL_SESSION structures with the same session ID (e.g. when two threads
2691 concurrently obtain them from an external cache).
2692 The internal cache can handle only one SSL_SESSION with a given ID,
2693 so if there's a conflict, we now throw out the old one to achieve
2697 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2698 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2699 some routines that use cipher OIDs: some ciphers do not have OIDs
2700 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2704 *) Simplify the trust setting structure and code. Now we just have
2705 two sequences of OIDs for trusted and rejected settings. These will
2706 typically have values the same as the extended key usage extension
2707 and any application specific purposes.
2709 The trust checking code now has a default behaviour: it will just
2710 check for an object with the same NID as the passed id. Functions can
2711 be provided to override either the default behaviour or the behaviour
2712 for a given id. SSL client, server and email already have functions
2713 in place for compatibility: they check the NID and also return "trusted"
2714 if the certificate is self signed.
2717 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2718 traditional format into an EVP_PKEY structure.
2721 *) Add a password callback function PEM_cb() which either prompts for
2722 a password if usr_data is NULL or otherwise assumes it is a null
2723 terminated password. Allow passwords to be passed on command line
2724 environment or config files in a few more utilities.
2727 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2728 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2729 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2730 Update documentation.
2733 *) Support for ASN1 "NULL" type. This could be handled before by using
2734 ASN1_TYPE but there wasn't any function that would try to read a NULL
2735 and produce an error if it couldn't. For compatibility we also have
2736 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2737 don't allocate anything because they don't need to.
2740 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2742 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2744 *) Rebuild of the memory allocation routines used by OpenSSL code and
2745 possibly others as well. The purpose is to make an interface that
2746 provide hooks so anyone can build a separate set of allocation and
2747 deallocation routines to be used by OpenSSL, for example memory
2748 pool implementations, or something else, which was previously hard
2749 since Malloc(), Realloc() and Free() were defined as macros having
2750 the values malloc, realloc and free, respectively (except for Win32
2751 compilations). The same is provided for memory debugging code.
2752 OpenSSL already comes with functionality to find memory leaks, but
2753 this gives people a chance to debug other memory problems.
2755 With these changes, a new set of functions and macros have appeared:
2757 CRYPTO_set_mem_debug_functions() [F]
2758 CRYPTO_get_mem_debug_functions() [F]
2759 CRYPTO_dbg_set_options() [F]
2760 CRYPTO_dbg_get_options() [F]
2761 CRYPTO_malloc_debug_init() [M]
2763 The memory debug functions are NULL by default, unless the library
2764 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2765 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2766 gives the standard debugging functions that come with OpenSSL) or
2767 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2768 provided by the library user) must be used. When the standard
2769 debugging functions are used, CRYPTO_dbg_set_options can be used to
2770 request additional information:
2771 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2772 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2774 Also, things like CRYPTO_set_mem_functions will always give the
2775 expected result (the new set of functions is used for allocation
2776 and deallocation) at all times, regardless of platform and compiler
2779 To finish it up, some functions that were never use in any other
2780 way than through macros have a new API and new semantic:
2783 CRYPTO_dbg_realloc()
2786 All macros of value have retained their old syntax.
2787 [Richard Levitte and Bodo Moeller]
2789 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2790 ordering of SMIMECapabilities wasn't in "strength order" and there
2791 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2795 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2796 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2797 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2799 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2800 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2801 functionality to handle multipart/signed properly) and a utility
2802 called 'smime' to call all this stuff. This is based on code I
2803 originally wrote for Celo who have kindly allowed it to be
2804 included in OpenSSL.
2807 *) Add variants des_set_key_checked and des_set_key_unchecked of
2808 des_set_key (aka des_key_sched). Global variable des_check_key
2809 decides which of these is called by des_set_key; this way
2810 des_check_key behaves as it always did, but applications and
2811 the library itself, which was buggy for des_check_key == 1,
2812 have a cleaner way to pick the version they need.
2815 *) New function PKCS12_newpass() which changes the password of a
2819 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2820 dynamic mix. In both cases the ids can be used as an index into the
2821 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2822 functions so they accept a list of the field values and the
2823 application doesn't need to directly manipulate the X509_TRUST
2827 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2831 *) Modify the way the V3 extension code looks up extensions. This now
2832 works in a similar way to the object code: we have some "standard"
2833 extensions in a static table which is searched with OBJ_bsearch()
2834 and the application can add dynamic ones if needed. The file
2835 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2836 updated whenever a new extension is added to the core code and kept
2837 in ext_nid order. There is a simple program 'tabtest.c' which checks
2838 this. New extensions are not added too often so this file can readily
2839 be maintained manually.
2841 There are two big advantages in doing things this way. The extensions
2842 can be looked up immediately and no longer need to be "added" using
2843 X509V3_add_standard_extensions(): this function now does nothing.
2844 [Side note: I get *lots* of email saying the extension code doesn't
2845 work because people forget to call this function]
2846 Also no dynamic allocation is done unless new extensions are added:
2847 so if we don't add custom extensions there is no need to call
2848 X509V3_EXT_cleanup().
2851 *) Modify enc utility's salting as follows: make salting the default. Add a
2852 magic header, so unsalted files fail gracefully instead of just decrypting
2853 to garbage. This is because not salting is a big security hole, so people
2854 should be discouraged from doing it.
2857 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2858 digest to be passed on the command line but it only used this
2859 parameter when signing a certificate. Modified so all relevant
2860 operations are affected by the digest parameter including the
2861 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2862 DSA key was used because it didn't fix the digest.
2865 *) Initial certificate chain verify code. Currently tests the untrusted
2866 certificates for consistency with the verify purpose (which is set
2867 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2869 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2870 this is because it will reject chains with invalid extensions whereas
2871 every previous version of OpenSSL and SSLeay made no checks at all.
2873 Trust code: checks the root CA for the relevant trust settings. Trust
2874 settings have an initial value consistent with the verify purpose: e.g.
2875 if the verify purpose is for SSL client use it expects the CA to be
2876 trusted for SSL client use. However the default value can be changed to
2877 permit custom trust settings: one example of this would be to only trust
2878 certificates from a specific "secure" set of CAs.
2880 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2881 which should be used for version portability: especially since the
2882 verify structure is likely to change more often now.
2884 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2885 to set them. If not set then assume SSL clients will verify SSL servers
2888 Two new options to the verify program: -untrusted allows a set of
2889 untrusted certificates to be passed in and -purpose which sets the
2890 intended purpose of the certificate. If a purpose is set then the
2891 new chain verify code is used to check extension consistency.
2894 *) Support for the authority information access extension.
2897 *) Modify RSA and DSA PEM read routines to transparently handle
2898 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2899 public keys in a format compatible with certificate
2900 SubjectPublicKeyInfo structures. Unfortunately there were already
2901 functions called *_PublicKey_* which used various odd formats so
2902 these are retained for compatibility: however the DSA variants were
2903 never in a public release so they have been deleted. Changed dsa/rsa
2904 utilities to handle the new format: note no releases ever handled public
2905 keys so we should be OK.
2907 The primary motivation for this change is to avoid the same fiasco
2908 that dogs private keys: there are several incompatible private key
2909 formats some of which are standard and some OpenSSL specific and
2910 require various evil hacks to allow partial transparent handling and
2911 even then it doesn't work with DER formats. Given the option anything
2912 other than PKCS#8 should be dumped: but the other formats have to
2913 stay in the name of compatibility.
2915 With public keys and the benefit of hindsight one standard format
2916 is used which works with EVP_PKEY, RSA or DSA structures: though
2917 it clearly returns an error if you try to read the wrong kind of key.
2919 Added a -pubkey option to the 'x509' utility to output the public key.
2920 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2921 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2922 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2923 that do the same as the EVP_PKEY_assign_*() except they up the
2924 reference count of the added key (they don't "swallow" the
2928 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2929 CRLs would fail if the file contained no certificates or no CRLs:
2930 added a new function to read in both types and return the number
2931 read: this means that if none are read it will be an error. The
2932 DER versions of the certificate and CRL reader would always fail
2933 because it isn't possible to mix certificates and CRLs in DER format
2934 without choking one or the other routine. Changed this to just read
2935 a certificate: this is the best we can do. Also modified the code
2936 in apps/verify.c to take notice of return codes: it was previously
2937 attempting to read in certificates from NULL pointers and ignoring
2938 any errors: this is one reason why the cert and CRL reader seemed
2939 to work. It doesn't check return codes from the default certificate
2940 routines: these may well fail if the certificates aren't installed.
2943 *) Code to support otherName option in GeneralName.
2946 *) First update to verify code. Change the verify utility
2947 so it warns if it is passed a self signed certificate:
2948 for consistency with the normal behaviour. X509_verify
2949 has been modified to it will now verify a self signed
2950 certificate if *exactly* the same certificate appears
2951 in the store: it was previously impossible to trust a
2952 single self signed certificate. This means that:
2953 openssl verify ss.pem
2954 now gives a warning about a self signed certificate but
2955 openssl verify -CAfile ss.pem ss.pem
2959 *) For servers, store verify_result in SSL_SESSION data structure
2960 (and add it to external session representation).
2961 This is needed when client certificate verifications fails,
2962 but an application-provided verification callback (set by
2963 SSL_CTX_set_cert_verify_callback) allows accepting the session
2964 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2965 but returns 1): When the session is reused, we have to set
2966 ssl->verify_result to the appropriate error code to avoid
2968 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2970 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2971 case in PKCS7_dataInit() where the signed PKCS7 structure
2972 didn't contain any existing data because it was being created.
2973 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2975 *) Add a salt to the key derivation routines in enc.c. This
2976 forms the first 8 bytes of the encrypted file. Also add a
2977 -S option to allow a salt to be input on the command line.
2980 *) New function X509_cmp(). Oddly enough there wasn't a function
2981 to compare two certificates. We do this by working out the SHA1
2982 hash and comparing that. X509_cmp() will be needed by the trust
2986 *) SSL_get1_session() is like SSL_get_session(), but increments
2987 the reference count in the SSL_SESSION returned.
2988 [Geoff Thorpe <geoff@eu.c2.net>]
2990 *) Fix for 'req': it was adding a null to request attributes.
2991 Also change the X509_LOOKUP and X509_INFO code to handle
2992 certificate auxiliary information.
2995 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2999 *) Add the possibility to add extra information to the memory leak
3000 detecting output, to form tracebacks, showing from where each
3001 allocation was originated: CRYPTO_push_info("constant string") adds
3002 the string plus current file name and line number to a per-thread
3003 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3004 is like calling CYRPTO_pop_info() until the stack is empty.
3005 Also updated memory leak detection code to be multi-thread-safe.
3008 *) Add options -text and -noout to pkcs7 utility and delete the
3009 encryption options which never did anything. Update docs.
3012 *) Add options to some of the utilities to allow the pass phrase
3013 to be included on either the command line (not recommended on
3014 OSes like Unix) or read from the environment. Update the
3015 manpages and fix a few bugs.
3018 *) Add a few manpages for some of the openssl commands.
3021 *) Fix the -revoke option in ca. It was freeing up memory twice,
3022 leaking and not finding already revoked certificates.
3025 *) Extensive changes to support certificate auxiliary information.
3026 This involves the use of X509_CERT_AUX structure and X509_AUX
3027 functions. An X509_AUX function such as PEM_read_X509_AUX()
3028 can still read in a certificate file in the usual way but it
3029 will also read in any additional "auxiliary information". By
3030 doing things this way a fair degree of compatibility can be
3031 retained: existing certificates can have this information added
3032 using the new 'x509' options.
3034 Current auxiliary information includes an "alias" and some trust
3035 settings. The trust settings will ultimately be used in enhanced
3036 certificate chain verification routines: currently a certificate
3037 can only be trusted if it is self signed and then it is trusted
3041 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3042 The problem was that one of the replacement routines had not been working
3043 since SSLeay releases. For now the offending routine has been replaced
3044 with non-optimised assembler. Even so, this now gives around 95%
3045 performance improvement for 1024 bit RSA signs.
3048 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3049 handling. Most clients have the effective key size in bits equal to
3050 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3051 A few however don't do this and instead use the size of the decrypted key
3052 to determine the RC2 key length and the AlgorithmIdentifier to determine
3053 the effective key length. In this case the effective key length can still
3054 be 40 bits but the key length can be 168 bits for example. This is fixed
3055 by manually forcing an RC2 key into the EVP_PKEY structure because the
3056 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3057 the key length and effective key length are equal.
3060 *) Add a bunch of functions that should simplify the creation of
3061 X509_NAME structures. Now you should be able to do:
3062 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3063 and have it automatically work out the correct field type and fill in
3064 the structures. The more adventurous can try:
3065 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3066 and it will (hopefully) work out the correct multibyte encoding.
3069 *) Change the 'req' utility to use the new field handling and multibyte
3070 copy routines. Before the DN field creation was handled in an ad hoc
3071 way in req, ca, and x509 which was rather broken and didn't support
3072 BMPStrings or UTF8Strings. Since some software doesn't implement
3073 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3074 using the dirstring_type option. See the new comment in the default
3075 openssl.cnf for more info.
3078 *) Make crypto/rand/md_rand.c more robust:
3079 - Assure unique random numbers after fork().
3080 - Make sure that concurrent threads access the global counter and
3081 md serializably so that we never lose entropy in them
3082 or use exactly the same state in multiple threads.
3083 Access to the large state is not always serializable because
3084 the additional locking could be a performance killer, and
3085 md should be large enough anyway.
3088 *) New file apps/app_rand.c with commonly needed functionality
3089 for handling the random seed file.
3091 Use the random seed file in some applications that previously did not:
3093 dsaparam -genkey (which also ignored its '-rand' option),
3096 x509 (when signing).
3097 Except on systems with /dev/urandom, it is crucial to have a random
3098 seed file at least for key creation, DSA signing, and for DH exchanges;
3099 for RSA signatures we could do without one.
3101 gendh and gendsa (unlike genrsa) used to read only the first byte
3102 of each file listed in the '-rand' option. The function as previously
3103 found in genrsa is now in app_rand.c and is used by all programs
3104 that support '-rand'.
3107 *) In RAND_write_file, use mode 0600 for creating files;
3108 don't just chmod when it may be too late.
3111 *) Report an error from X509_STORE_load_locations
3112 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3115 *) New function ASN1_mbstring_copy() this copies a string in either
3116 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3117 into an ASN1_STRING type. A mask of permissible types is passed
3118 and it chooses the "minimal" type to use or an error if not type
3122 *) Add function equivalents to the various macros in asn1.h. The old
3123 macros are retained with an M_ prefix. Code inside the library can
3124 use the M_ macros. External code (including the openssl utility)
3125 should *NOT* in order to be "shared library friendly".
3128 *) Add various functions that can check a certificate's extensions
3129 to see if it usable for various purposes such as SSL client,
3130 server or S/MIME and CAs of these types. This is currently
3131 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3132 verification. Also added a -purpose flag to x509 utility to
3133 print out all the purposes.
3136 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3140 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3141 for, obtain and decode and extension and obtain its critical flag.
3142 This allows all the necessary extension code to be handled in a
3143 single function call.
3146 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3147 platforms. See crypto/rc4/rc4_enc.c for further details.
3150 *) New -noout option to asn1parse. This causes no output to be produced
3151 its main use is when combined with -strparse and -out to extract data
3152 from a file (which may not be in ASN.1 format).
3155 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3156 when producing the local key id.
3157 [Richard Levitte <levitte@stacken.kth.se>]
3159 *) New option -dhparam in s_server. This allows a DH parameter file to be
3160 stated explicitly. If it is not stated then it tries the first server
3161 certificate file. The previous behaviour hard coded the filename
3165 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3166 a public key to be input or output. For example:
3167 openssl rsa -in key.pem -pubout -out pubkey.pem
3168 Also added necessary DSA public key functions to handle this.
3171 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3172 in the message. This was handled by allowing
3173 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3174 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3176 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3177 to the end of the strings whereas this didn't. This would cause problems
3178 if strings read with d2i_ASN1_bytes() were later modified.
3179 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3181 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3182 data and it contains EOF it will end up returning an error. This is
3183 caused by input 46 bytes long. The cause is due to the way base64
3184 BIOs find the start of base64 encoded data. They do this by trying a
3185 trial decode on each line until they find one that works. When they
3186 do a flag is set and it starts again knowing it can pass all the
3187 data directly through the decoder. Unfortunately it doesn't reset
3188 the context it uses. This means that if EOF is reached an attempt
3189 is made to pass two EOFs through the context and this causes the
3190 resulting error. This can also cause other problems as well. As is
3191 usual with these problems it takes *ages* to find and the fix is
3192 trivial: move one line.
3193 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3195 *) Ugly workaround to get s_client and s_server working under Windows. The
3196 old code wouldn't work because it needed to select() on sockets and the
3197 tty (for keypresses and to see if data could be written). Win32 only
3198 supports select() on sockets so we select() with a 1s timeout on the
3199 sockets and then see if any characters are waiting to be read, if none
3200 are present then we retry, we also assume we can always write data to
3201 the tty. This isn't nice because the code then blocks until we've
3202 received a complete line of data and it is effectively polling the
3203 keyboard at 1s intervals: however it's quite a bit better than not
3204 working at all :-) A dedicated Windows application might handle this
3205 with an event loop for example.
3208 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3209 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3210 will be called when RSA_sign() and RSA_verify() are used. This is useful
3211 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3212 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3213 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3214 This necessitated the support of an extra signature type NID_md5_sha1
3215 for SSL signatures and modifications to the SSL library to use it instead
3216 of calling RSA_public_decrypt() and RSA_private_encrypt().
3219 *) Add new -verify -CAfile and -CApath options to the crl program, these
3220 will lookup a CRL issuers certificate and verify the signature in a
3221 similar way to the verify program. Tidy up the crl program so it
3222 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3223 less strict. It will now permit CRL extensions even if it is not
3224 a V2 CRL: this will allow it to tolerate some broken CRLs.
3227 *) Initialize all non-automatic variables each time one of the openssl
3228 sub-programs is started (this is necessary as they may be started
3229 multiple times from the "OpenSSL>" prompt).
3230 [Lennart Bang, Bodo Moeller]
3232 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3233 removing all other RSA functionality (this is what NO_RSA does). This
3234 is so (for example) those in the US can disable those operations covered
3235 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3239 *) Non-copying interface to BIO pairs.
3240 (still largely untested)
3243 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3244 ASCII string. This was handled independently in various places before.
3247 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3248 UTF8 strings a character at a time.
3251 *) Use client_version from client hello to select the protocol
3252 (s23_srvr.c) and for RSA client key exchange verification
3253 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3256 *) Add various utility functions to handle SPKACs, these were previously
3257 handled by poking round in the structure internals. Added new function
3258 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3259 print, verify and generate SPKACs. Based on an original idea from
3260 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3263 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3266 *) Allow the config file extension section to be overwritten on the
3267 command line. Based on an original idea from Massimiliano Pala
3268 <madwolf@comune.modena.it>. The new option is called -extensions
3269 and can be applied to ca, req and x509. Also -reqexts to override
3270 the request extensions in req and -crlexts to override the crl extensions
3274 *) Add new feature to the SPKAC handling in ca. Now you can include
3275 the same field multiple times by preceding it by "XXXX." for example:
3278 this is the same syntax as used in the req config file.
3281 *) Allow certificate extensions to be added to certificate requests. These
3282 are specified in a 'req_extensions' option of the req section of the
3283 config file. They can be printed out with the -text option to req but
3284 are otherwise ignored at present.
3287 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3288 data read consists of only the final block it would not decrypted because
3289 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3290 A misplaced 'break' also meant the decrypted final block might not be
3291 copied until the next read.
3294 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3295 a few extra parameters to the DH structure: these will be useful if
3296 for example we want the value of 'q' or implement X9.42 DH.
3299 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3300 provides hooks that allow the default DSA functions or functions on a
3301 "per key" basis to be replaced. This allows hardware acceleration and
3302 hardware key storage to be handled without major modification to the
3303 library. Also added low level modexp hooks and CRYPTO_EX structure and
3304 associated functions.
3307 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3308 as "read only": it can't be written to and the buffer it points to will
3309 not be freed. Reading from a read only BIO is much more efficient than
3310 a normal memory BIO. This was added because there are several times when
3311 an area of memory needs to be read from a BIO. The previous method was
3312 to create a memory BIO and write the data to it, this results in two
3313 copies of the data and an O(n^2) reading algorithm. There is a new
3314 function BIO_new_mem_buf() which creates a read only memory BIO from
3315 an area of memory. Also modified the PKCS#7 routines to use read only
3319 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3320 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3321 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3322 but a retry condition occured while trying to read the rest.
3325 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3326 NID_pkcs7_encrypted by default: this was wrong since this should almost
3327 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3328 the encrypted data type: this is a more sensible place to put it and it
3329 allows the PKCS#12 code to be tidied up that duplicated this
3333 *) Changed obj_dat.pl script so it takes its input and output files on
3334 the command line. This should avoid shell escape redirection problems
3338 *) Initial support for certificate extension requests, these are included
3339 in things like Xenroll certificate requests. Included functions to allow
3340 extensions to be obtained and added.
3343 *) -crlf option to s_client and s_server for sending newlines as
3344 CRLF (as required by many protocols).
3347 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3349 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3350 [Ralf S. Engelschall]
3352 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3353 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3355 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3359 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3360 DH parameters/keys (q is lost during that conversion, but the resulting
3361 DH parameters contain its length).
3363 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3364 much faster than DH_generate_parameters (which creates parameters
3365 where p = 2*q + 1), and also the smaller q makes DH computations
3366 much more efficient (160-bit exponentiation instead of 1024-bit
3367 exponentiation); so this provides a convenient way to support DHE
3368 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3369 utter importance to use
3370 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3372 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3373 when such DH parameters are used, because otherwise small subgroup
3374 attacks may become possible!
3377 *) Avoid memory leak in i2d_DHparams.
3380 *) Allow the -k option to be used more than once in the enc program:
3381 this allows the same encrypted message to be read by multiple recipients.
3384 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3385 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3386 it will always use the numerical form of the OID, even if it has a short
3390 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3391 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3392 otherwise bn_mod_exp was called. In the case of hardware keys for example
3393 no private key components need be present and it might store extra data
3394 in the RSA structure, which cannot be accessed from bn_mod_exp.
3395 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3396 private key operations.
3399 *) Added support for SPARC Linux.
3402 *) pem_password_cb function type incompatibly changed from
3403 typedef int pem_password_cb(char *buf, int size, int rwflag);
3405 ....(char *buf, int size, int rwflag, void *userdata);
3406 so that applications can pass data to their callbacks:
3407 The PEM[_ASN1]_{read,write}... functions and macros now take an
3408 additional void * argument, which is just handed through whenever
3409 the password callback is called.
3410 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3412 New function SSL_CTX_set_default_passwd_cb_userdata.
3414 Compatibility note: As many C implementations push function arguments
3415 onto the stack in reverse order, the new library version is likely to
3416 interoperate with programs that have been compiled with the old
3417 pem_password_cb definition (PEM_whatever takes some data that
3418 happens to be on the stack as its last argument, and the callback
3419 just ignores this garbage); but there is no guarantee whatsoever that
3422 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3423 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3424 problems not only on Windows, but also on some Unix platforms.
3425 To avoid problematic command lines, these definitions are now in an
3426 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3427 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3430 *) MIPS III/IV assembler module is reimplemented.
3433 *) More DES library cleanups: remove references to srand/rand and
3434 delete an unused file.
3437 *) Add support for the the free Netwide assembler (NASM) under Win32,
3438 since not many people have MASM (ml) and it can be hard to obtain.
3439 This is currently experimental but it seems to work OK and pass all
3440 the tests. Check out INSTALL.W32 for info.
3443 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3444 without temporary keys kept an extra copy of the server key,
3445 and connections with temporary keys did not free everything in case
3449 *) New function RSA_check_key and new openssl rsa option -check
3450 for verifying the consistency of RSA keys.
3451 [Ulf Moeller, Bodo Moeller]
3453 *) Various changes to make Win32 compile work:
3454 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3455 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3456 comparison" warnings.
3457 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3460 *) Add a debugging option to PKCS#5 v2 key generation function: when
3461 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3462 derived keys are printed to stderr.
3465 *) Copy the flags in ASN1_STRING_dup().
3466 [Roman E. Pavlov <pre@mo.msk.ru>]
3468 *) The x509 application mishandled signing requests containing DSA
3469 keys when the signing key was also DSA and the parameters didn't match.
3471 It was supposed to omit the parameters when they matched the signing key:
3472 the verifying software was then supposed to automatically use the CA's
3473 parameters if they were absent from the end user certificate.
3475 Omitting parameters is no longer recommended. The test was also
3476 the wrong way round! This was probably due to unusual behaviour in
3477 EVP_cmp_parameters() which returns 1 if the parameters match.
3478 This meant that parameters were omitted when they *didn't* match and
3479 the certificate was useless. Certificates signed with 'ca' didn't have
3481 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3483 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3484 The interface is as follows:
3485 Applications can use
3486 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3487 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3488 "off" is now the default.
3489 The library internally uses
3490 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3491 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3492 to disable memory-checking temporarily.
3494 Some inconsistent states that previously were possible (and were
3495 even the default) are now avoided.
3497 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3498 with each memory chunk allocated; this is occasionally more helpful
3499 than just having a counter.
3501 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3503 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3507 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3508 which largely parallels "options", but is for changing API behaviour,
3509 whereas "options" are about protocol behaviour.
3510 Initial "mode" flags are:
3512 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3513 a single record has been written.
3514 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3515 retries use the same buffer location.
3516 (But all of the contents must be
3520 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3523 *) Fix problems with no-hmac etc.
3524 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3526 *) New functions RSA_get_default_method(), RSA_set_method() and
3527 RSA_get_method(). These allows replacement of RSA_METHODs without having
3528 to mess around with the internals of an RSA structure.
3531 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3532 Also really enable memory leak checks in openssl.c and in some
3534 [Chad C. Mulligan, Bodo Moeller]
3536 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3537 up the length of negative integers. This has now been simplified to just
3538 store the length when it is first determined and use it later, rather
3539 than trying to keep track of where data is copied and updating it to
3541 [Steve Henson, reported by Brien Wheeler
3542 <bwheeler@authentica-security.com>]
3544 *) Add a new function PKCS7_signatureVerify. This allows the verification
3545 of a PKCS#7 signature but with the signing certificate passed to the
3546 function itself. This contrasts with PKCS7_dataVerify which assumes the
3547 certificate is present in the PKCS#7 structure. This isn't always the
3548 case: certificates can be omitted from a PKCS#7 structure and be
3549 distributed by "out of band" means (such as a certificate database).
3552 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3553 function prototypes in pem.h, also change util/mkdef.pl to add the
3554 necessary function names.
3557 *) mk1mf.pl (used by Windows builds) did not properly read the
3558 options set by Configure in the top level Makefile, and Configure
3559 was not even able to write more than one option correctly.
3560 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3563 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3564 file to be loaded from a BIO or FILE pointer. The BIO version will
3565 for example allow memory BIOs to contain config info.
3568 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3569 Whoever hopes to achieve shared-library compatibility across versions
3570 must use this, not the compile-time macro.
3571 (Exercise 0.9.4: Which is the minimum library version required by
3573 Note: All this applies only to multi-threaded programs, others don't
3577 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3578 through a BIO pair triggered the default case, i.e.
3579 SSLerr(...,SSL_R_UNKNOWN_STATE).
3582 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3583 can use the SSL library even if none of the specific BIOs is
3587 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3588 for the encoded length.
3589 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3591 *) Add initial documentation of the X509V3 functions.
3594 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3595 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3596 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3597 secure PKCS#8 private key format with a high iteration count.
3600 *) Fix determination of Perl interpreter: A perl or perl5
3601 _directory_ in $PATH was also accepted as the interpreter.
3602 [Ralf S. Engelschall]
3604 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3605 wrong with it but it was very old and did things like calling
3606 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3610 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3611 to use the new extension code.
3614 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3615 with macros. This should make it easier to change their form, add extra
3616 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3620 *) Add to configuration table a new entry that can specify an alternative
3621 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3622 according to Mark Crispin <MRC@Panda.COM>.
3626 *) DES CBC did not update the IV. Weird.
3629 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3630 Changing the behaviour of the former might break existing programs --
3631 where IV updating is needed, des_ncbc_encrypt can be used.
3634 *) When bntest is run from "make test" it drives bc to check its
3635 calculations, as well as internally checking them. If an internal check
3636 fails, it needs to cause bc to give a non-zero result or make test carries
3637 on without noticing the failure. Fixed.
3640 *) DES library cleanups.
3643 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3644 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3645 ciphers. NOTE: although the key derivation function has been verified
3646 against some published test vectors it has not been extensively tested
3647 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3651 *) Instead of "mkdir -p", which is not fully portable, use new
3652 Perl script "util/mkdir-p.pl".
3655 *) Rewrite the way password based encryption (PBE) is handled. It used to
3656 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3657 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3658 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3659 the 'parameter' field of the AlgorithmIdentifier is passed to the
3660 underlying key generation function so it must do its own ASN1 parsing.
3661 This has also changed the EVP_PBE_CipherInit() function which now has a
3662 'parameter' argument instead of literal salt and iteration count values
3663 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3666 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3667 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3668 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3669 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3670 value was just used as a "magic string" and not used directly its
3671 value doesn't matter.
3674 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3678 *) "linux-sparc64" configuration (ultrapenguin).
3679 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3680 "linux-sparc" configuration.
3681 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3683 *) config now generates no-xxx options for missing ciphers.
3686 *) Support the EBCDIC character set (work in progress).
3687 File ebcdic.c not yet included because it has a different license.
3688 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3690 *) Support BS2000/OSD-POSIX.
3691 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3693 *) Make callbacks for key generation use void * instead of char *.
3696 *) Make S/MIME samples compile (not yet tested).
3699 *) Additional typesafe stacks.
3702 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3706 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3708 *) New configuration variant "sco5-gcc".
3710 *) Updated some demos.
3711 [Sean O Riordain, Wade Scholine]
3713 *) Add missing BIO_free at exit of pkcs12 application.
3716 *) Fix memory leak in conf.c.
3719 *) Updates for Win32 to assembler version of MD5.
3722 *) Set #! path to perl in apps/der_chop to where we found it
3723 instead of using a fixed path.
3726 *) SHA library changes for irix64-mips4-cc.
3729 *) Improvements for VMS support.
3733 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3735 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3736 This also avoids the problems with SC4.2 and unpatched SC5.
3737 [Andy Polyakov <appro@fy.chalmers.se>]
3739 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3740 These are required because of the typesafe stack would otherwise break
3741 existing code. If old code used a structure member which used to be STACK
3742 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3743 sk_num or sk_value it would produce an error because the num, data members
3744 are not present in STACK_OF. Now it just produces a warning. sk_set
3745 replaces the old method of assigning a value to sk_value
3746 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3747 that does this will no longer work (and should use sk_set instead) but
3748 this could be regarded as a "questionable" behaviour anyway.
3751 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3752 correctly handle encrypted S/MIME data.
3755 *) Change type of various DES function arguments from des_cblock
3756 (which means, in function argument declarations, pointer to char)
3757 to des_cblock * (meaning pointer to array with 8 char elements),
3758 which allows the compiler to do more typechecking; it was like
3759 that back in SSLeay, but with lots of ugly casts.
3761 Introduce new type const_des_cblock.
3764 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3765 problems: find RecipientInfo structure that matches recipient certificate
3766 and initialise the ASN1 structures properly based on passed cipher.
3769 *) Belatedly make the BN tests actually check the results.
3772 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3773 to and from BNs: it was completely broken. New compilation option
3774 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3775 key elements as negative integers.
3778 *) Reorganize and speed up MD5.
3779 [Andy Polyakov <appro@fy.chalmers.se>]
3782 [Richard Levitte <richard@levitte.org>]
3784 *) New option -out to asn1parse to allow the parsed structure to be
3785 output to a file. This is most useful when combined with the -strparse
3786 option to examine the output of things like OCTET STRINGS.
3789 *) Make SSL library a little more fool-proof by not requiring any longer
3790 that SSL_set_{accept,connect}_state be called before
3791 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3792 in many applications because usually everything *appeared* to work as
3793 intended anyway -- now it really works as intended).
3796 *) Move openssl.cnf out of lib/.
3799 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3800 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3801 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3802 [Ralf S. Engelschall]
3804 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3805 handle PKCS#7 enveloped data properly.
3806 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3808 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3809 copying pointers. The cert_st handling is changed by this in
3810 various ways (and thus what used to be known as ctx->default_cert
3811 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3812 any longer when s->cert does not give us what we need).
3813 ssl_cert_instantiate becomes obsolete by this change.
3814 As soon as we've got the new code right (possibly it already is?),
3815 we have solved a couple of bugs of the earlier code where s->cert
3816 was used as if it could not have been shared with other SSL structures.
3818 Note that using the SSL API in certain dirty ways now will result
3819 in different behaviour than observed with earlier library versions:
3820 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3821 does not influence s as it used to.
3823 In order to clean up things more thoroughly, inside SSL_SESSION
3824 we don't use CERT any longer, but a new structure SESS_CERT
3825 that holds per-session data (if available); currently, this is
3826 the peer's certificate chain and, for clients, the server's certificate
3827 and temporary key. CERT holds only those values that can have
3828 meaningful defaults in an SSL_CTX.
3831 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3832 from the internal representation. Various PKCS#7 fixes: remove some
3833 evil casts and set the enc_dig_alg field properly based on the signing
3837 *) Allow PKCS#12 password to be set from the command line or the
3838 environment. Let 'ca' get its config file name from the environment
3839 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3843 *) Allow certificate policies extension to use an IA5STRING for the
3844 organization field. This is contrary to the PKIX definition but
3845 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3849 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3850 without disallowing inline assembler and the like for non-pedantic builds.
3853 *) Support Borland C++ builder.
3854 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3859 *) SHA-1 cleanups and performance enhancements.
3860 [Andy Polyakov <appro@fy.chalmers.se>]
3862 *) Sparc v8plus assembler for the bignum library.
3863 [Andy Polyakov <appro@fy.chalmers.se>]
3865 *) Accept any -xxx and +xxx compiler options in Configure.
3868 *) Update HPUX configuration.
3871 *) Add missing sk_<type>_unshift() function to safestack.h
3872 [Ralf S. Engelschall]
3874 *) New function SSL_CTX_use_certificate_chain_file that sets the
3875 "extra_cert"s in addition to the certificate. (This makes sense
3876 only for "PEM" format files, as chains as a whole are not
3880 *) Support verify_depth from the SSL API.
3881 x509_vfy.c had what can be considered an off-by-one-error:
3882 Its depth (which was not part of the external interface)
3883 was actually counting the number of certificates in a chain;
3884 now it really counts the depth.
3887 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3888 instead of X509err, which often resulted in confusing error
3889 messages since the error codes are not globally unique
3890 (e.g. an alleged error in ssl3_accept when a certificate
3891 didn't match the private key).
3893 *) New function SSL_CTX_set_session_id_context that allows to set a default
3894 value (so that you don't need SSL_set_session_id_context for each
3895 connection using the SSL_CTX).
3898 *) OAEP decoding bug fix.
3901 *) Support INSTALL_PREFIX for package builders, as proposed by
3905 *) New Configure options "threads" and "no-threads". For systems
3906 where the proper compiler options are known (currently Solaris
3907 and Linux), "threads" is the default.
3910 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3913 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3914 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3915 such as /usr/local/bin.
3918 *) "make linux-shared" to build shared libraries.
3919 [Niels Poppe <niels@netbox.org>]
3921 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3924 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3925 extension adding in x509 utility.
3928 *) Remove NOPROTO sections and error code comments.
3931 *) Partial rewrite of the DEF file generator to now parse the ANSI
3935 *) New Configure options --prefix=DIR and --openssldir=DIR.
3938 *) Complete rewrite of the error code script(s). It is all now handled
3939 by one script at the top level which handles error code gathering,
3940 header rewriting and C source file generation. It should be much better
3941 than the old method: it now uses a modified version of Ulf's parser to
3942 read the ANSI prototypes in all header files (thus the old K&R definitions
3943 aren't needed for error creation any more) and do a better job of
3944 translating function codes into names. The old 'ASN1 error code imbedded
3945 in a comment' is no longer necessary and it doesn't use .err files which
3946 have now been deleted. Also the error code call doesn't have to appear all
3947 on one line (which resulted in some large lines...).
3950 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3953 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3954 0 (which usually indicates a closed connection), but continue reading.
3957 *) Fix some race conditions.
3960 *) Add support for CRL distribution points extension. Add Certificate
3961 Policies and CRL distribution points documentation.
3964 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3967 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3968 8 of keying material. Merlin has also confirmed interop with this fix
3969 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3970 [Merlin Hughes <merlin@baltimore.ie>]
3972 *) Fix lots of warnings.
3973 [Richard Levitte <levitte@stacken.kth.se>]
3975 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3976 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3977 [Richard Levitte <levitte@stacken.kth.se>]
3979 *) Fix problems with sizeof(long) == 8.
3980 [Andy Polyakov <appro@fy.chalmers.se>]
3982 *) Change functions to ANSI C.
3985 *) Fix typos in error codes.
3986 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3988 *) Remove defunct assembler files from Configure.
3991 *) SPARC v8 assembler BIGNUM implementation.
3992 [Andy Polyakov <appro@fy.chalmers.se>]
3994 *) Support for Certificate Policies extension: both print and set.
3995 Various additions to support the r2i method this uses.
3998 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3999 return a const string when you are expecting an allocated buffer.
4002 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4003 types DirectoryString and DisplayText.
4006 *) Add code to allow r2i extensions to access the configuration database,
4007 add an LHASH database driver and add several ctx helper functions.
4010 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4011 fail when they extended the size of a BIGNUM.
4014 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4015 support typesafe stack.
4018 *) Fix typo in SSL_[gs]et_options().
4019 [Nils Frostberg <nils@medcom.se>]
4021 *) Delete various functions and files that belonged to the (now obsolete)
4022 old X509V3 handling code.
4025 *) New Configure option "rsaref".
4028 *) Don't auto-generate pem.h.
4031 *) Introduce type-safe ASN.1 SETs.
4034 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4035 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4037 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4038 that links with OpenSSL (well at least cause lots of warnings), but fear
4039 not: the conversion is trivial, and it eliminates loads of evil casts. A
4040 few STACKed things have been converted already. Feel free to convert more.
4041 In the fullness of time, I'll do away with the STACK type altogether.
4044 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4045 specified in <certfile> by updating the entry in the index.txt file.
4046 This way one no longer has to edit the index.txt file manually for
4047 revoking a certificate. The -revoke option does the gory details now.
4048 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4050 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4051 `-text' option at all and this way the `-noout -text' combination was
4052 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4053 [Ralf S. Engelschall]
4055 *) Make sure a corresponding plain text error message exists for the
4056 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4057 verify callback function determined that a certificate was revoked.
4058 [Ralf S. Engelschall]
4060 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4061 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4062 all available cipers including rc5, which was forgotten until now.
4063 In order to let the testing shell script know which algorithms
4064 are available, a new (up to now undocumented) command
4065 "openssl list-cipher-commands" is used.
4068 *) Bugfix: s_client occasionally would sleep in select() when
4069 it should have checked SSL_pending() first.
4072 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4073 the raw DSA values prior to ASN.1 encoding.
4076 *) Tweaks to Configure
4077 [Niels Poppe <niels@netbox.org>]
4079 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4083 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4086 *) New config option to avoid instructions that are illegal on the 80386.
4087 The default code is faster, but requires at least a 486.
4090 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4091 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4092 same as SSL2_VERSION anyway.
4095 *) New "-showcerts" option for s_client.
4098 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4099 application. Various cleanups and fixes.
4102 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4103 modify error routines to work internally. Add error codes and PBE init
4104 to library startup routines.
4107 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4108 packing functions to asn1 and evp. Changed function names and error
4109 codes along the way.
4112 *) PKCS12 integration: and so it begins... First of several patches to
4113 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4114 objects to objects.h
4117 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4118 and display support for Thawte strong extranet extension.
4121 *) Add LinuxPPC support.
4122 [Jeff Dubrule <igor@pobox.org>]
4124 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4125 bn_div_words in alpha.s.
4126 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4128 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4129 OAEP isn't supported when OpenSSL is built with RSAref.
4130 [Ulf Moeller <ulf@fitug.de>]
4132 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4133 so they no longer are missing under -DNOPROTO.
4134 [Soren S. Jorvang <soren@t.dk>]
4137 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4139 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4140 doesn't work when the session is reused. Coming soon!
4143 *) Fix a security hole, that allows sessions to be reused in the wrong
4144 context thus bypassing client cert protection! All software that uses
4145 client certs and session caches in multiple contexts NEEDS PATCHING to
4146 allow session reuse! A fuller solution is in the works.
4147 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4149 *) Some more source tree cleanups (removed obsolete files
4150 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4151 permission on "config" script to be executable) and a fix for the INSTALL
4153 [Ulf Moeller <ulf@fitug.de>]
4155 *) Remove some legacy and erroneous uses of malloc, free instead of
4157 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4159 *) Make rsa_oaep_test return non-zero on error.
4160 [Ulf Moeller <ulf@fitug.de>]
4162 *) Add support for native Solaris shared libraries. Configure
4163 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4164 if someone would make that last step automatic.
4165 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4167 *) ctx_size was not built with the right compiler during "make links". Fixed.
4170 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4171 except NULL ciphers". This means the default cipher list will no longer
4172 enable NULL ciphers. They need to be specifically enabled e.g. with
4173 the string "DEFAULT:eNULL".
4176 *) Fix to RSA private encryption routines: if p < q then it would
4177 occasionally produce an invalid result. This will only happen with
4178 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4181 *) Be less restrictive and allow also `perl util/perlpath.pl
4182 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4183 because this way one can also use an interpreter named `perl5' (which is
4184 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4185 installed as `perl').
4186 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4188 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4189 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4191 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4192 advapi32.lib to Win32 build and change the pem test comparision
4193 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4194 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4195 and crypto/des/ede_cbcm_enc.c.
4198 *) DES quad checksum was broken on big-endian architectures. Fixed.
4201 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4202 Win32 test batch file so it (might) work again. The Win32 test batch file
4203 is horrible: I feel ill....
4206 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4207 in e_os.h. Audit of header files to check ANSI and non ANSI
4208 sections: 10 functions were absent from non ANSI section and not exported
4209 from Windows DLLs. Fixed up libeay.num for new functions.
4212 *) Make `openssl version' output lines consistent.
4213 [Ralf S. Engelschall]
4215 *) Fix Win32 symbol export lists for BIO functions: Added
4216 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4217 to ms/libeay{16,32}.def.
4218 [Ralf S. Engelschall]
4220 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4221 fine under Unix and passes some trivial tests I've now added. But the
4222 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4223 added to make sure no one expects that this stuff really works in the
4224 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4225 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4227 [Ralf S. Engelschall]
4229 *) Fix the generation of two part addresses in perl.
4230 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4232 *) Add config entry for Linux on MIPS.
4233 [John Tobey <jtobey@channel1.com>]
4235 *) Make links whenever Configure is run, unless we are on Windoze.
4238 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4239 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4243 *) Add a useful kludge to allow package maintainers to specify compiler and
4244 other platforms details on the command line without having to patch the
4245 Configure script everytime: One now can use ``perl Configure
4246 <id>:<details>'', i.e. platform ids are allowed to have details appended
4247 to them (seperated by colons). This is treated as there would be a static
4248 pre-configured entry in Configure's %table under key <id> with value
4249 <details> and ``perl Configure <id>'' is called. So, when you want to
4250 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4251 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4252 now, which overrides the FreeBSD-elf entry on-the-fly.
4253 [Ralf S. Engelschall]
4255 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4258 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4259 on the `perl Configure ...' command line. This way one can compile
4260 OpenSSL libraries with Position Independent Code (PIC) which is needed
4261 for linking it into DSOs.
4262 [Ralf S. Engelschall]
4264 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4268 *) Cleaned up the LICENSE document: The official contact for any license
4269 questions now is the OpenSSL core team under openssl-core@openssl.org.
4270 And add a paragraph about the dual-license situation to make sure people
4271 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4272 to the OpenSSL toolkit.
4273 [Ralf S. Engelschall]
4275 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4276 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4277 Additonally cleaned up the `make links' target: Remove unnecessary
4278 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4279 to speed processing and no longer clutter the display with confusing
4280 stuff. Instead only the actually done links are displayed.
4281 [Ralf S. Engelschall]
4283 *) Permit null encryption ciphersuites, used for authentication only. It used
4284 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4285 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4289 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4290 signed attributes when verifying signatures (this would break them),
4291 the detached data encoding was wrong and public keys obtained using
4292 X509_get_pubkey() weren't freed.
4295 *) Add text documentation for the BUFFER functions. Also added a work around
4296 to a Win95 console bug. This was triggered by the password read stuff: the
4297 last character typed gets carried over to the next fread(). If you were
4298 generating a new cert request using 'req' for example then the last
4299 character of the passphrase would be CR which would then enter the first
4303 *) Added the new `Includes OpenSSL Cryptography Software' button as
4304 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4305 button and can be used by applications based on OpenSSL to show the
4306 relationship to the OpenSSL project.
4307 [Ralf S. Engelschall]
4309 *) Remove confusing variables in function signatures in files
4310 ssl/ssl_lib.c and ssl/ssl.h.
4311 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4313 *) Don't install bss_file.c under PREFIX/include/
4314 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4316 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4317 functions that return function pointers and has support for NT specific
4318 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4319 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4320 unsigned to signed types: this was killing the Win32 compile.
4323 *) Add new certificate file to stack functions,
4324 SSL_add_dir_cert_subjects_to_stack() and
4325 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4326 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4327 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4328 This means that Apache-SSL and similar packages don't have to mess around
4329 to add as many CAs as they want to the preferred list.
4332 *) Experiment with doxygen documentation. Currently only partially applied to
4334 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4335 openssl.doxy as the configuration file.
4338 *) Get rid of remaining C++-style comments which strict C compilers hate.
4339 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4341 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4342 compiled in by default: it has problems with large keys.
4345 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4346 DH private keys and/or callback functions which directly correspond to
4347 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4348 is needed for applications which have to configure certificates on a
4349 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4351 For the RSA certificate situation is makes no difference, but
4352 for the DSA certificate situation this fixes the "no shared cipher"
4353 problem where the OpenSSL cipher selection procedure failed because the
4354 temporary keys were not overtaken from the context and the API provided
4355 no way to reconfigure them.
4356 The new functions now let applications reconfigure the stuff and they
4357 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4358 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4359 non-public-API function ssl_cert_instantiate() is used as a helper
4360 function and also to reduce code redundancy inside ssl_rsa.c.
4361 [Ralf S. Engelschall]
4363 *) Move s_server -dcert and -dkey options out of the undocumented feature
4364 area because they are useful for the DSA situation and should be
4365 recognized by the users.
4366 [Ralf S. Engelschall]
4368 *) Fix the cipher decision scheme for export ciphers: the export bits are
4369 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4370 SSL_EXP_MASK. So, the original variable has to be used instead of the
4371 already masked variable.
4372 [Richard Levitte <levitte@stacken.kth.se>]
4374 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4375 [Richard Levitte <levitte@stacken.kth.se>]
4377 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4378 from `int' to `unsigned int' because it's a length and initialized by
4379 EVP_DigestFinal() which expects an `unsigned int *'.
4380 [Richard Levitte <levitte@stacken.kth.se>]
4382 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4383 script. Instead use the usual Shell->Perl transition trick.
4384 [Ralf S. Engelschall]
4386 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4387 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4388 -noout -modulus' as it's already the case for `openssl rsa -noout
4389 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4390 currently the public key is printed (a decision which was already done by
4391 `openssl dsa -modulus' in the past) which serves a similar purpose.
4392 Additionally the NO_RSA no longer completely removes the whole -modulus
4393 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4395 [Ralf S. Engelschall]
4397 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4398 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4399 [Arne Ansper <arne@ats.cyber.ee>]
4401 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4402 to be added. Now both 'req' and 'ca' can use new objects defined in the
4406 *) Add cool BIO that does syslog (or event log on NT).
4407 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4409 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4410 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4411 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4412 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4415 *) Add preliminary config info for new extension code.
4418 *) Make RSA_NO_PADDING really use no padding.
4419 [Ulf Moeller <ulf@fitug.de>]
4421 *) Generate errors when private/public key check is done.
4424 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4425 for some CRL extensions and new objects added.
4428 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4429 key usage extension and fuller support for authority key id.
4432 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4433 padding method for RSA, which is recommended for new applications in PKCS
4434 #1 v2.0 (RFC 2437, October 1998).
4435 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4436 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4437 against Bleichbacher's attack on RSA.
4438 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4441 *) Updates to the new SSL compression code
4442 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4444 *) Fix so that the version number in the master secret, when passed
4445 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4446 (because the server will not accept higher), that the version number
4447 is 0x03,0x01, not 0x03,0x00
4448 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4450 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4451 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4452 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4455 *) Support for RAW extensions where an arbitrary extension can be
4456 created by including its DER encoding. See apps/openssl.cnf for
4460 *) Make sure latest Perl versions don't interpret some generated C array
4461 code as Perl array code in the crypto/err/err_genc.pl script.
4462 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4464 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4465 not many people have the assembler. Various Win32 compilation fixes and
4466 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4470 *) Modify configure script 'Configure' to automatically create crypto/date.h
4471 file under Win32 and also build pem.h from pem.org. New script
4472 util/mkfiles.pl to create the MINFO file on environments that can't do a
4473 'make files': perl util/mkfiles.pl >MINFO should work.
4476 *) Major rework of DES function declarations, in the pursuit of correctness
4477 and purity. As a result, many evil casts evaporated, and some weirdness,
4478 too. You may find this causes warnings in your code. Zapping your evil
4479 casts will probably fix them. Mostly.
4482 *) Fix for a typo in asn1.h. Bug fix to object creation script
4483 obj_dat.pl. It considered a zero in an object definition to mean
4484 "end of object": none of the objects in objects.h have any zeros
4485 so it wasn't spotted.
4486 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
4488 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
4489 Masking (CBCM). In the absence of test vectors, the best I have been able
4490 to do is check that the decrypt undoes the encrypt, so far. Send me test
4491 vectors if you have them.
4494 *) Correct calculation of key length for export ciphers (too much space was
4495 allocated for null ciphers). This has not been tested!
4498 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
4499 message is now correct (it understands "crypto" and "ssl" on its
4500 command line). There is also now an "update" option. This will update
4501 the util/ssleay.num and util/libeay.num files with any new functions.
4503 perl util/mkdef.pl crypto ssl update
4504 it will update them.
4507 *) Overhauled the Perl interface (perl/*):
4508 - ported BN stuff to OpenSSL's different BN library
4509 - made the perl/ source tree CVS-aware
4510 - renamed the package from SSLeay to OpenSSL (the files still contain
4511 their history because I've copied them in the repository)
4512 - removed obsolete files (the test scripts will be replaced
4513 by better Test::Harness variants in the future)
4514 [Ralf S. Engelschall]
4516 *) First cut for a very conservative source tree cleanup:
4517 1. merge various obsolete readme texts into doc/ssleay.txt
4518 where we collect the old documents and readme texts.
4519 2. remove the first part of files where I'm already sure that we no
4520 longer need them because of three reasons: either they are just temporary
4521 files which were left by Eric or they are preserved original files where
4522 I've verified that the diff is also available in the CVS via "cvs diff
4523 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
4524 the crypto/md/ stuff).
4525 [Ralf S. Engelschall]
4527 *) More extension code. Incomplete support for subject and issuer alt
4528 name, issuer and authority key id. Change the i2v function parameters
4529 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
4530 what that's for :-) Fix to ASN1 macro which messed up
4531 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
4534 *) Preliminary support for ENUMERATED type. This is largely copied from the
4538 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
4539 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4541 *) Make sure `make rehash' target really finds the `openssl' program.
4542 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4544 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
4545 like to hear about it if this slows down other processors.
4548 *) Add CygWin32 platform information to Configure script.
4549 [Alan Batie <batie@aahz.jf.intel.com>]
4551 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
4552 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
4554 *) New program nseq to manipulate netscape certificate sequences
4557 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
4561 *) Fixes to BN code. Previously the default was to define BN_RECURSION
4562 but the BN code had some problems that would cause failures when
4563 doing certificate verification and some other functions.
4564 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4566 *) Add ASN1 and PEM code to support netscape certificate sequences.
4569 *) Add ASN1 and PEM code to support netscape certificate sequences.
4572 *) Add several PKIX and private extended key usage OIDs.
4575 *) Modify the 'ca' program to handle the new extension code. Modify
4576 openssl.cnf for new extension format, add comments.
4579 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
4580 and add a sample to openssl.cnf so req -x509 now adds appropriate
4584 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
4585 error code, add initial support to X509_print() and x509 application.
4588 *) Takes a deep breath and start addding X509 V3 extension support code. Add
4589 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
4590 stuff is currently isolated and isn't even compiled yet.
4593 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
4594 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
4595 Removed the versions check from X509 routines when loading extensions:
4596 this allows certain broken certificates that don't set the version
4597 properly to be processed.
4600 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
4601 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
4602 can still be regenerated with "make depend".
4605 *) Spelling mistake in C version of CAST-128.
4606 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
4608 *) Changes to the error generation code. The perl script err-code.pl
4609 now reads in the old error codes and retains the old numbers, only
4610 adding new ones if necessary. It also only changes the .err files if new
4611 codes are added. The makefiles have been modified to only insert errors
4612 when needed (to avoid needlessly modifying header files). This is done
4613 by only inserting errors if the .err file is newer than the auto generated
4614 C file. To rebuild all the error codes from scratch (the old behaviour)
4615 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
4616 or delete all the .err files.
4619 *) CAST-128 was incorrectly implemented for short keys. The C version has
4620 been fixed, but is untested. The assembler versions are also fixed, but
4621 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
4622 to regenerate it if needed.
4623 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4624 Hagino <itojun@kame.net>]
4626 *) File was opened incorrectly in randfile.c.
4627 [Ulf Möller <ulf@fitug.de>]
4629 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4630 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4631 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4632 al: it's just almost always a UTCTime. Note this patch adds new error
4633 codes so do a "make errors" if there are problems.
4636 *) Correct Linux 1 recognition in config.
4637 [Ulf Möller <ulf@fitug.de>]
4639 *) Remove pointless MD5 hash when using DSA keys in ca.
4640 [Anonymous <nobody@replay.com>]
4642 *) Generate an error if given an empty string as a cert directory. Also
4643 generate an error if handed NULL (previously returned 0 to indicate an
4644 error, but didn't set one).
4645 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4647 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4650 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4651 parameters. This was causing a warning which killed off the Win32 compile.
4654 *) Remove C++ style comments from crypto/bn/bn_local.h.
4655 [Neil Costigan <neil.costigan@celocom.com>]
4657 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4658 based on a text string, looking up short and long names and finally
4659 "dot" format. The "dot" format stuff didn't work. Added new function
4660 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4661 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4662 OID is not part of the table.
4665 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4666 X509_LOOKUP_by_alias().
4669 *) Sort openssl functions by name.
4672 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4673 encryption from sample DSA keys (in case anyone is interested the password
4677 *) Make _all_ *_free functions accept a NULL pointer.
4678 [Frans Heymans <fheymans@isaserver.be>]
4680 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4682 [Anonymous <nobody@replay.com>]
4684 *) s_server should send the CAfile as acceptable CAs, not its own cert.
4685 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4687 *) Don't blow it for numeric -newkey arguments to apps/req.
4688 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4690 *) Temp key "for export" tests were wrong in s3_srvr.c.
4691 [Anonymous <nobody@replay.com>]
4693 *) Add prototype for temp key callback functions
4694 SSL_CTX_set_tmp_{rsa,dh}_callback().
4697 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4698 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4701 *) X509_name_add_entry() freed the wrong thing after an error.
4702 [Arne Ansper <arne@ats.cyber.ee>]
4704 *) rsa_eay.c would attempt to free a NULL context.
4705 [Arne Ansper <arne@ats.cyber.ee>]
4707 *) BIO_s_socket() had a broken should_retry() on Windoze.
4708 [Arne Ansper <arne@ats.cyber.ee>]
4710 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4711 [Arne Ansper <arne@ats.cyber.ee>]
4713 *) Make sure the already existing X509_STORE->depth variable is initialized
4714 in X509_STORE_new(), but document the fact that this variable is still
4715 unused in the certificate verification process.
4716 [Ralf S. Engelschall]
4718 *) Fix the various library and apps files to free up pkeys obtained from
4719 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4722 *) Fix reference counting in X509_PUBKEY_get(). This makes
4723 demos/maurice/example2.c work, amongst others, probably.
4724 [Steve Henson and Ben Laurie]
4726 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4727 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4728 are no longer created. This way we have a single and consistent command
4729 line interface `openssl <command>', similar to `cvs <command>'.
4730 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4732 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4733 BIT STRING wrapper always have zero unused bits.
4736 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4739 *) Make the top-level INSTALL documentation easier to understand.
4742 *) Makefiles updated to exit if an error occurs in a sub-directory
4743 make (including if user presses ^C) [Paul Sutton]
4745 *) Make Montgomery context stuff explicit in RSA data structure.
4748 *) Fix build order of pem and err to allow for generated pem.h.
4751 *) Fix renumbering bug in X509_NAME_delete_entry().
4754 *) Enhanced the err-ins.pl script so it makes the error library number
4755 global and can add a library name. This is needed for external ASN1 and
4756 other error libraries.
4759 *) Fixed sk_insert which never worked properly.
4762 *) Fix ASN1 macros so they can handle indefinite length construted
4763 EXPLICIT tags. Some non standard certificates use these: they can now
4767 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4768 into a single doc/ssleay.txt bundle. This way the information is still
4769 preserved but no longer messes up this directory. Now it's new room for
4770 the new set of documenation files.
4771 [Ralf S. Engelschall]
4773 *) SETs were incorrectly DER encoded. This was a major pain, because they
4774 shared code with SEQUENCEs, which aren't coded the same. This means that
4775 almost everything to do with SETs or SEQUENCEs has either changed name or
4776 number of arguments.
4777 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4779 *) Fix test data to work with the above.
4782 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4783 was already fixed by Eric for 0.9.1 it seems.
4784 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4786 *) Autodetect FreeBSD3.
4789 *) Fix various bugs in Configure. This affects the following platforms:
4793 unixware-2.0-pentium
4797 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4798 before they are needed.
4801 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4805 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4807 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4808 changed SSLeay to OpenSSL in version strings.
4809 [Ralf S. Engelschall]
4811 *) Some fixups to the top-level documents.
4814 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4815 because the symlink to include/ was missing.
4816 [Ralf S. Engelschall]
4818 *) Incorporated the popular no-RSA/DSA-only patches
4819 which allow to compile a RSA-free SSLeay.
4820 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4822 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4823 when "ssleay" is still not found.
4824 [Ralf S. Engelschall]
4826 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4827 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4829 *) Updated the README file.
4830 [Ralf S. Engelschall]
4832 *) Added various .cvsignore files in the CVS repository subdirs
4833 to make a "cvs update" really silent.
4834 [Ralf S. Engelschall]
4836 *) Recompiled the error-definition header files and added
4837 missing symbols to the Win32 linker tables.
4838 [Ralf S. Engelschall]
4840 *) Cleaned up the top-level documents;
4841 o new files: CHANGES and LICENSE
4842 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4843 o merged COPYRIGHT into LICENSE
4844 o removed obsolete TODO file
4845 o renamed MICROSOFT to INSTALL.W32
4846 [Ralf S. Engelschall]
4848 *) Removed dummy files from the 0.9.1b source tree:
4849 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4850 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4851 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4852 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4853 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4854 [Ralf S. Engelschall]
4856 *) Added various platform portability fixes.
4859 *) The Genesis of the OpenSSL rpject:
4860 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4861 Young and Tim J. Hudson created while they were working for C2Net until
4863 [The OpenSSL Project]
4866 Changes between 0.9.0b and 0.9.1b [not released]
4868 *) Updated a few CA certificates under certs/
4871 *) Changed some BIGNUM api stuff.
4874 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4875 DGUX x86, Linux Alpha, etc.
4878 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4879 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4883 *) Add -strparse option to asn1pars program which parses nested
4885 [Dr Stephen Henson <shenson@bigfoot.com>]
4887 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4890 *) DSA fix for "ca" program.
4893 *) Added "-genkey" option to "dsaparam" program.
4896 *) Added RIPE MD160 (rmd160) message digest.
4899 *) Added -a (all) option to "ssleay version" command.
4902 *) Added PLATFORM define which is the id given to Configure.
4905 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4908 *) Extended the ASN.1 parser routines.
4911 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4914 *) Added a BN_CTX to the BN library.
4917 *) Fixed the weak key values in DES library
4920 *) Changed API in EVP library for cipher aliases.
4923 *) Added support for RC2/64bit cipher.
4926 *) Converted the lhash library to the crypto/mem.c functions.
4929 *) Added more recognized ASN.1 object ids.
4932 *) Added more RSA padding checks for SSL/TLS.
4935 *) Added BIO proxy/filter functionality.
4938 *) Added extra_certs to SSL_CTX which can be used
4939 send extra CA certificates to the client in the CA cert chain sending
4940 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4943 *) Now Fortezza is denied in the authentication phase because
4944 this is key exchange mechanism is not supported by SSLeay at all.
4947 *) Additional PKCS1 checks.
4950 *) Support the string "TLSv1" for all TLS v1 ciphers.
4953 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4954 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4957 *) Fixed a few memory leaks.
4960 *) Fixed various code and comment typos.
4963 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4964 bytes sent in the client random.
4965 [Edward Bishop <ebishop@spyglass.com>]