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 *) Enhance the general user interface with mechanisms for inner control
15 and with pssibilities to have yes/no kind of prompts.
18 *) Change all calls to low level digest routines in the library and
19 applications to use EVP. Add missing calls to HMAC_cleanup() and
20 don't assume HMAC_CTX can be copied using memcpy().
21 [Verdon Walker <VWalker@novell.com>, Steve Henson]
23 +) Add the possibility to control engines through control names but with
24 arbitrary arguments instead of just a string.
25 Change the key loaders to take a UI_METHOD instead of a callback
26 function pointer. NOTE: this breaks binary compatibility with earlier
27 versions of OpenSSL [engine].
28 Addapt the nCipher code for these new conditions and add a card insertion
32 +) Enhance the general user interface with mechanisms to better support
33 dialog box interfaces, application-defined prompts, the possibility
34 to use defaults (for example default passwords from somewhere else)
35 and interrupts/cancelations.
38 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
39 used: it isn't thread safe and the add_lock_callback should handle
41 [Paul Rose <Paul.Rose@bridge.com>]
43 *) Verify that incoming data obeys the block size in
44 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
47 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
48 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
52 [Ulf Möller, Bodo Möller]
54 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
55 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5
56 when fixing the server behaviour for backwards-compatible 'client
57 hello' messages. (Note that the attack is impractical against
58 SSL 3.0 and TLS 1.0 anyway because length and version checking
59 means that the probability of guessing a valid ciphertext is
60 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
63 Before 0.9.5, the countermeasure (hide the error by generating a
64 random 'decryption result') did not work properly because
65 ERR_clear_error() was missing, meaning that SSL_get_error() would
66 detect the supposedly ignored error.
68 Both problems are now fixed.
71 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
72 (previously it was 1024).
75 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
76 tidy up some unecessarily weird code in 'sk_new()').
77 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
79 +) Change the key loading routines for ENGINEs to use the same kind
80 callback (pem_password_cb) as all other routines that need this
84 *) Fix for compatibility mode trust settings: ignore trust settings
85 unless some valid trust or reject settings are present.
88 *) Fix for blowfish EVP: its a variable length cipher.
91 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
92 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
93 than this minimum value is recommended.
96 +) New random seeder for OpenVMS, using the system process statistics
97 that are easily reachable.
100 +) Windows apparently can't transparently handle global
101 variables defined in DLLs. Initialisations such as:
103 const ASN1_ITEM *it = &ASN1_INTEGER_it;
105 wont compile. This is used by the any applications that need to
106 delcare their own ASN1 modules. This was fixed by adding the option
107 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
108 needed for static libraries under Win32.
111 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
112 setting of purpose and trust fields. New X509_STORE trust and
113 purpose functions and tidy up setting in other SSL functions.
116 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
117 structure. These are inherited by X509_STORE_CTX when it is
118 initialised. This allows various defaults to be set in the
119 X509_STORE structure (such as flags for CRL checking and custom
120 purpose or trust settings) for functions which only use X509_STORE_CTX
121 internally such as S/MIME.
123 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
124 trust settings if they are not set in X509_STORE. This allows X509_STORE
125 purposes and trust (in S/MIME for example) to override any set by default.
127 Add command line options for CRL checking to smime, s_client and s_server
131 +) Initial CRL based revocation checking. If the CRL checking flag(s)
132 are set then the CRL is looked up in the X509_STORE structure and
133 its validity and signature checked, then if the certificate is found
134 in the CRL the verify fails with a revoked error.
136 Various new CRL related callbacks added to X509_STORE_CTX structure.
138 Command line options added to 'verify' application to support this.
140 This needs some additional work, such as being able to handle multiple
141 CRLs with different times, extension based lookup (rather than just
142 by subject name) and ultimately more complete V2 CRL extension
146 +) Add a general user interface API (crypto/ui/). This is designed
147 to replace things like des_read_password and friends (backward
148 compatibility functions using this new API are provided).
149 The purpose is to remove prompting functions from the DES code
150 section as well as provide for prompting through dialog boxes in
151 a window system and the like.
154 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
155 in the current directory if neither $RANDFILE nor $HOME was set.
156 RAND_file_name() in 0.9.6a returned NULL in this case. This has
157 caused some confusion to Windows users who haven't defined $HOME.
158 Thus RAND_file_name() is changed again: e_os.h can define a
159 DEFAULT_HOME, which will be used if $HOME is not set.
160 For Windows, we use "C:"; on other platforms, we still require
161 environment variables.
163 +) Add "ex_data" support to ENGINE so implementations can add state at a
164 per-structure level rather than having to store it globally.
167 +) Make it possible for ENGINE structures to be copied when retrieved by
168 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
169 This causes the "original" ENGINE structure to act like a template,
170 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
171 operational state can be localised to each ENGINE structure, despite the
172 fact they all share the same "methods". New ENGINE structures returned in
173 this case have no functional references and the return value is the single
174 structural reference. This matches the single structural reference returned
175 by ENGINE_by_id() normally, when it is incremented on the pre-existing
179 *) Fix various bugs related to DSA S/MIME verification. Handle missing
180 parameters in DSA public key structures and return an error in the
181 DSA routines if parameters are absent.
184 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
185 needs to match any other type at all we need to manually clear the
189 +) Changes to the "openssl engine" utility to include;
190 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
191 about an ENGINE's available control commands.
192 - executing control commands from command line arguments using the
193 '-pre' and '-post' switches. '-post' is only used if '-t' is
194 specified and the ENGINE is successfully initialised. The syntax for
195 the individual commands are colon-separated, for example;
196 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
199 +) New dynamic control command support for ENGINEs. ENGINEs can now
200 declare their own commands (numbers), names (strings), descriptions,
201 and input types for run-time discovery by calling applications. A
202 subset of these commands are implicitly classed as "executable"
203 depending on their input type, and only these can be invoked through
204 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
205 can be based on user input, config files, etc). The distinction is
206 that "executable" commands cannot return anything other than a boolean
207 result and can only support numeric or string input, whereas some
208 discoverable commands may only be for direct use through
209 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
210 pointers, or other custom uses. The "executable" commands are to
211 support parameterisations of ENGINE behaviour that can be
212 unambiguously defined by ENGINEs and used consistently across any
213 OpenSSL-based application. Commands have been added to all the
214 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
215 control over shared-library paths without source code alterations.
218 +) Changed all ENGINE implementations to dynamically allocate their
219 ENGINEs rather than declaring them statically. Apart from this being
220 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
221 this also allows the implementations to compile without using the
222 internal engine_int.h header.
225 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
226 'const' value. Any code that should be able to modify a RAND_METHOD
227 should already have non-const pointers to it (ie. they should only
228 modify their own ones).
231 +) Made a variety of little tweaks to the ENGINE code.
232 - "atalla" and "ubsec" string definitions were moved from header files
233 to C code. "nuron" string definitions were placed in variables
234 rather than hard-coded - allowing parameterisation of these values
235 later on via ctrl() commands.
236 - Removed unused "#if 0"'d code.
237 - Fixed engine list iteration code so it uses ENGINE_free() to release
238 structural references.
239 - Constified the RAND_METHOD element of ENGINE structures.
240 - Constified various get/set functions as appropriate and added
241 missing functions (including a catch-all ENGINE_cpy that duplicates
242 all ENGINE values onto a new ENGINE except reference counts/state).
243 - Removed NULL parameter checks in get/set functions. Setting a method
244 or function to NULL is a way of cancelling out a previously set
245 value. Passing a NULL ENGINE parameter is just plain stupid anyway
246 and doesn't justify the extra error symbols and code.
247 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
248 flags from engine_int.h to engine.h.
249 - Changed prototypes for ENGINE handler functions (init(), finish(),
250 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
253 *) Move 'if (!initialized) RAND_poll()' into regions protected by
254 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
255 having multiple threads call RAND_poll() concurrently.
258 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
259 combination of a flag and a thread ID variable.
260 Otherwise while one thread is in ssleay_rand_bytes (which sets the
261 flag), *other* threads can enter ssleay_add_bytes without obeying
262 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock
263 that they do not hold after the first thread unsets add_do_not_lock).
266 +) Implement binary inversion algorithm for BN_mod_inverse in addition
267 to the algorithm using long divison. The binary algorithm can be
268 used only if the modulus is odd. On 32-bit systems, it is faster
269 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
270 roughly 5-15% for 256-bit moduli), so we use it only for moduli
271 up to 450 bits. In 64-bit environments, the binary algorithm
272 appears to be advantageous for much longer moduli; here we use it
273 for moduli up to 2048 bits.
276 *) Change bctest again: '-x' expressions are not available in all
280 -) OpenSSL 0.9.6a released [5 Apr 2001]
282 *) Fix a couple of memory leaks in PKCS7_dataDecode()
283 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
285 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
286 the default extension for executables, if any. Also, make the perl
287 scripts that use symlink() to test if it really exists and use "cp"
288 if it doesn't. All this made OpenSSL compilable and installable in
292 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
293 could not support the combine flag in choice fields.
296 -) Fix for asn1_GetSequence() for indefinite length constructed data.
297 If SEQUENCE is length is indefinite just set c->slen to the total
298 amount of data available.
299 [Steve Henson, reported by shige@FreeBSD.org]
300 [This change does not apply to 0.9.7.]
302 *) Change bctest to avoid here-documents inside command substitution
303 (workaround for FreeBSD /bin/sh bug).
304 For compatibility with Ultrix, avoid shell functions (introduced
305 in the bctest version that searches along $PATH).
308 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
309 with des_encrypt() defined on some operating systems, like Solaris
313 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
314 On the Importance of Eliminating Errors in Cryptographic
315 Computations, J. Cryptology 14 (2001) 2, 101-119,
316 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
319 *) MIPS assembler BIGNUM division bug fix.
322 *) Disabled incorrect Alpha assembler code.
325 -) Fix PKCS#7 decode routines so they correctly update the length
326 after reading an EOC for the EXPLICIT tag.
328 [This change does not apply to 0.9.7.]
330 *) Fix bug in PKCS#12 key generation routines. This was triggered
331 if a 3DES key was generated with a 0 initial byte. Include
332 PKCS12_BROKEN_KEYGEN compilation option to retain the old
333 (but broken) behaviour.
336 *) Enhance bctest to search for a working bc along $PATH and print
338 [Tim Rice <tim@multitalents.net> via Richard Levitte]
340 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
341 extensions from a certificate request to the certificate.
344 +) Allow multiple 'certopt' and 'nameopt' options to be separated
345 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
346 file: this allows the display of the certificate about to be
347 signed to be customised, to allow certain fields to be included
348 or excluded and extension details. The old system didn't display
349 multicharacter strings properly, omitted fields not in the policy
350 and couldn't display additional details such as extensions.
353 *) Fix memory leaks in err.c: free err_data string if necessary;
354 don't write to the wrong index in ERR_set_error_data.
357 +) Function EC_POINTs_mul for simultaneous scalar multiplication
358 of an arbitrary number of elliptic curve points, optionally
359 including the generator defined for the EC_GROUP.
360 EC_POINT_mul is a simple wrapper function for the typical case
361 that the point list has just one item (besides the optional
365 +) First EC_METHODs for curves over GF(p):
367 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
368 operations and provides various method functions that can also
369 operate with faster implementations of modular arithmetic.
371 EC_GFp_mont_method() reuses most functions that are part of
372 EC_GFp_simple_method, but uses Montgomery arithmetic.
374 [Bodo Moeller; point addition and point doubling
375 implementation directly derived from source code provided by
376 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
378 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
381 Curves are EC_GROUP objects (with an optional group generator)
382 based on EC_METHODs that are built into the library.
384 Points are EC_POINT objects based on EC_GROUP objects.
386 Most of the framework would be able to handle curves over arbitrary
387 finite fields, but as there are no obvious types for fields other
388 than GF(p), some functions are limited to that for now.
391 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
392 that the file contains a complete HTTP response.
395 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
396 change the def and num file printf format specifier from "%-40sXXX"
397 to "%-39s XXX". The latter will always guarantee a space after the
398 field while the former will cause them to run together if the field
399 is 40 of more characters long.
402 +) Constify the cipher and digest 'method' functions and structures
403 and modify related functions to take constant EVP_MD and EVP_CIPHER
407 *) Implement ssl23_peek (analogous to ssl23_read), which previously
411 *) Replace rdtsc with _emit statements for VC++ version 5.
412 [Jeremy Cooper <jeremy@baymoo.org>]
414 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
415 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
418 +) Modify EVP_Digest*() routines so they now return values. Although the
419 internal software routines can never fail additional hardware versions
423 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
425 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
426 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
429 ERR_R_NESTED_ASN1_ERROR
431 ERR_R_MISSING_ASN1_EOS
432 were 4 .. 9, conflicting with
433 ERR_LIB_RSA (= ERR_R_RSA_LIB)
435 ERR_LIB_PEM (= ERR_R_PEM_LIB).
436 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
438 Add new error code 'ERR_R_INTERNAL_ERROR'.
441 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
445 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
446 sets the subject name for a new request or supersedes the
447 subject name in a given request. Formats that can be parsed are
448 'CN=Some Name, OU=myOU, C=IT'
450 'CN=Some Name/OU=myOU/C=IT'.
452 Add options '-batch' and '-verbose' to 'openssl req'.
453 [Massimiliano Pala <madwolf@hackmasters.net>]
455 *) Make it possible to reuse SSLv2 sessions.
458 +) Introduce the possibility to access global variables through
459 functions on platform were that's the best way to handle exporting
460 global variables in shared libraries. To enable this functionality,
461 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
462 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
463 is normally done by Configure or something similar).
465 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
466 in the source file (foo.c) like this:
468 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
469 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
471 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
472 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
474 OPENSSL_DECLARE_GLOBAL(int,foo);
475 #define foo OPENSSL_GLOBAL_REF(foo)
476 OPENSSL_DECLARE_GLOBAL(double,bar);
477 #define bar OPENSSL_GLOBAL_REF(bar)
479 The #defines are very important, and therefore so is including the
480 header file everywere where the defined globals are used.
482 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
483 of ASN.1 items, but that structure is a bt different.
485 The largest change is in util/mkdef.pl which has been enhanced with
486 better and easier to understand logic to choose which symbols should
487 go into the Windows .def files as well as a number of fixes and code
488 cleanup (among others, algorithm keywords are now sorted
489 lexicographically to avoid constant rewrites).
492 *) In copy_email() check for >= 0 as a return value for
493 X509_NAME_get_index_by_NID() since 0 is a valid index.
494 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
496 +) In BN_div() keep a copy of the sign of 'num' before writing the
497 result to 'rm' because if rm==num the value will be overwritten
498 and produce the wrong result if 'num' is negative: this caused
499 problems with BN_mod() and BN_nnmod().
502 +) Function OCSP_request_verify(). This checks the signature on an
503 OCSP request and verifies the signer certificate. The signer
504 certificate is just checked for a generic purpose and OCSP request
508 +) Add OCSP_check_validity() function to check the validity of OCSP
509 responses. OCSP responses are prepared in real time and may only
510 be a few seconds old. Simply checking that the current time lies
511 between thisUpdate and nextUpdate max reject otherwise valid responses
512 caused by either OCSP responder or client clock innacuracy. Instead
513 we allow thisUpdate and nextUpdate to fall within a certain period of
514 the current time. The age of the response can also optionally be
515 checked. Two new options -validity_period and -status_age added to
519 +) If signature or public key algorithm is unrecognized print out its
520 OID rather that just UNKOWN.
523 *) Avoid coredump with unsupported or invalid public keys by checking if
524 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
525 PKCS7_verify() fails with non detached data.
528 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
529 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
530 ID to be generated from the issuer certificate alone which can then be
531 passed to OCSP_id_issuer_cmp().
534 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
535 ASN1 modules to export functions returning ASN1_ITEM pointers
536 instead of the ASN1_ITEM structures themselves. This adds several
537 new macros which allow the underlying ASN1 function/structure to
538 be accessed transparently. As a result code should not use ASN1_ITEM
539 references directly (such as &X509_it) but instead use the relevant
540 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
541 use of the new ASN1 code on platforms where exporting structures
542 is problematical (for example in shared libraries) but exporting
543 functions returning pointers to structures is not.
546 +) Add support for overriding the generation of SSL/TLS session IDs.
547 These callbacks can be registered either in an SSL_CTX or per SSL.
548 The purpose of this is to allow applications to control, if they wish,
549 the arbitrary values chosen for use as session IDs, particularly as it
550 can be useful for session caching in multiple-server environments. A
551 command-line switch for testing this (and any client code that wishes
552 to use such a feature) has been added to "s_server".
553 [Geoff Thorpe, Lutz Jaenicke]
555 +) Modify mkdef.pl to recognise and parse prprocessor conditionals
556 of the form '#if defined(...) || defined(...) || ...' and
557 '#if !defined(...) && !defined(...) && ...'. This also avoids
558 the growing number of special cases it was previously handling.
561 *) Don't use getenv in library functions when run as setuid/setgid.
562 New function OPENSSL_issetugid().
565 +) Make all configuration macros available for application by making
566 sure they are available in opensslconf.h, by giving them names starting
567 with "OPENSSL_" to avoid conflicts with other packages and by making
568 sure e_os2.h will cover all platform-specific cases together with
570 Additionally, it is now possible to define configuration/platform-
571 specific names (called "system identities"). In the C code, these
572 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
573 macro with the name beginning with "OPENSSL_SYS_", which is determined
574 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
578 +) New option -set_serial to 'req' and 'x509' this allows the serial
579 number to use to be specified on the command line. Previously self
580 signed certificates were hard coded with serial number 0 and the
581 CA options of 'x509' had to use a serial number in a file which was
585 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
586 due to incorrect handling of multi-threading:
588 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
590 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
592 3. Count how many times MemCheck_off() has been called so that
593 nested use can be treated correctly. This also avoids
594 inband-signalling in the previous code (which relied on the
595 assumption that thread ID 0 is impossible).
598 +) New options to 'ca' utility to support V2 CRL entry extensions.
599 Currently CRL reason, invalidity date and hold instruction are
600 supported. Add new CRL extensions to V3 code and some new objects.
603 *) Add "-rand" option also to s_client and s_server.
606 +) New function EVP_CIPHER_CTX_set_padding() this is used to
607 disable standard block padding (aka PKCS#5 padding) in the EVP
608 API, which was previously mandatory. This means that the data is
609 not padded in any way and so the total length much be a multiple
610 of the block size, otherwise an error occurs.
613 +) Initial (incomplete) OCSP SSL support.
616 *) Fix CPU detection on Irix 6.x.
617 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
618 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
620 +) New function OCSP_parse_url(). This splits up a URL into its host,
621 port and path components: primarily to parse OCSP URLs. New -url
622 option to ocsp utility.
625 +) New nonce behavior. The return value of OCSP_check_nonce() now
626 reflects the various checks performed. Applications can decide
627 whether to tolerate certain situations such as an absent nonce
628 in a response when one was present in a request: the ocsp application
629 just prints out a warning. New function OCSP_add1_basic_nonce()
630 this is to allow responders to include a nonce in a response even if
631 the request is nonce-less.
634 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
636 [This change does not apply to 0.9.7.]
638 *) Use the cached encoding of an X509_NAME structure rather than
639 copying it. This is apparently the reason for the libsafe "errors"
640 but the code is actually correct.
643 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
644 skipped when using openssl x509 multiple times on a single input file,
645 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
648 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
649 set string type: to handle setting ASN1_TIME structures. Fix ca
650 utility to correctly initialize revocation date of CRLs.
653 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
654 the clients preferred ciphersuites and rather use its own preferences.
655 Should help to work around M$ SGC (Server Gated Cryptography) bug in
656 Internet Explorer by ensuring unchanged hash method during stepup.
659 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
660 to aes and add a new 'exist' option to print out symbols that don't
664 +) Additional options to ocsp utility to allow flags to be set and
665 additional certificates supplied.
668 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
669 OCSP client a number of certificate to only verify the response
673 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
674 Bleichenbacher's DSA attack.
675 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
676 to be set and top=0 forces the highest bit to be set; top=-1 is new
677 and leaves the highest bit random.
678 [Ulf Moeller, Bodo Moeller]
680 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
681 handle the new API. Currently only ECB, CBC modes supported. Add new
682 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
683 for TLS" draft-ietf-tls-ciphersuite-03.txt.
684 [Ben Laurie, Steve Henson]
686 *) In the NCONF_...-based implementations for CONF_... queries
687 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
688 a temporary CONF structure with the data component set to NULL
689 (which gives segmentation faults in lh_retrieve).
690 Instead, use NULL for the CONF pointer in CONF_get_string and
691 CONF_get_number (which may use environment variables) and directly
692 return NULL from CONF_get_section.
695 *) Fix potential buffer overrun for EBCDIC.
698 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
702 +) Functions for OCSP responders. OCSP_request_onereq_count(),
703 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
704 extract information from a certificate request. OCSP_response_create()
705 creates a response and optionally adds a basic response structure.
706 OCSP_basic_add1_status() adds a complete single response to a basic
707 reponse and returns the OCSP_SINGLERESP structure just added (to allow
708 extensions to be included for example). OCSP_basic_add1_cert() adds a
709 certificate to a basic response and OCSP_basic_sign() signs a basic
710 response with various flags. New helper functions ASN1_TIME_check()
711 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
712 (converts ASN1_TIME to GeneralizedTime).
715 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
716 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
717 structure from a certificate. X509_pubkey_digest() digests tha public_key
718 contents: this is used in various key identifiers.
721 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
722 keyUsage if basicConstraints absent for a CA.
725 *) Make SMIME_write_PKCS7() write mail header values with a format that
726 is more generally accepted (no spaces before the semicolon), since
727 some programs can't parse those values properly otherwise. Also make
728 sure BIO's that break lines after each write do not create invalid
732 +) Make sk_sort() tolerate a NULL argument.
733 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
735 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
736 passed by the function are trusted implicitly. If any of them signed the
737 reponse then it is assumed to be valid and is not verified.
740 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
741 macros previously used would not encode an empty SEQUENCE OF
742 and break the signature.
744 [This change does not apply to 0.9.7.]
746 *) Zero the premaster secret after deriving the master secret in
750 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
751 to data. This was previously part of the PKCS7 ASN1 code. This
752 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
753 [Steve Henson, reported by Kenneth R. Robinette
754 <support@securenetterm.com>]
756 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
757 routines: without these tracing memory leaks is very painful.
758 Fix leaks in PKCS12 and PKCS7 routines.
761 *) Add some EVP_add_digest_alias registrations (as found in
762 OpenSSL_add_all_digests()) to SSL_library_init()
763 aka OpenSSL_add_ssl_algorithms(). This provides improved
764 compatibility with peers using X.509 certificates
765 with unconventional AlgorithmIdentifier OIDs.
768 *) Fix for Irix with NO_ASM.
769 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
771 *) ./config script fixes.
772 [Ulf Moeller, Richard Levitte]
774 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
775 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
776 effectively meant GeneralizedTime would never be used. Now it
777 is initialised to -1 but X509_time_adj() now has to check the value
778 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
779 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
780 [Steve Henson, reported by Kenneth R. Robinette
781 <support@securenetterm.com>]
783 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
784 result in a zero length in the ASN1_INTEGER structure which was
785 not consistent with the structure when d2i_ASN1_INTEGER() was used
786 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
787 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
788 where it did not print out a minus for negative ASN1_INTEGER.
791 *) Fix 'openssl passwd -1'.
794 +) Add summary printout to ocsp utility. The various functions which
795 convert status values to strings have been renamed to:
796 OCSP_response_status_str(), OCSP_cert_status_str() and
797 OCSP_crl_reason_str() and are no longer static. New options
798 to verify nonce values and to disable verification. OCSP response
799 printout format cleaned up.
802 +) Add additional OCSP certificate checks. These are those specified
803 in RFC2560. This consists of two separate checks: the CA of the
804 certificate being checked must either be the OCSP signer certificate
805 or the issuer of the OCSP signer certificate. In the latter case the
806 OCSP signer certificate must contain the OCSP signing extended key
807 usage. This check is performed by attempting to match the OCSP
808 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
809 in the OCSP_CERTID structures of the response.
812 +) Initial OCSP certificate verification added to OCSP_basic_verify()
813 and related routines. This uses the standard OpenSSL certificate
814 verify routines to perform initial checks (just CA validity) and
815 to obtain the certificate chain. Then additional checks will be
816 performed on the chain. Currently the root CA is checked to see
817 if it is explicitly trusted for OCSP signing. This is used to set
818 a root CA as a global signing root: that is any certificate that
819 chains to that CA is an acceptable OCSP signing certificate.
822 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
823 extensions from a separate configuration file.
824 As when reading extensions from the main configuration file,
825 the '-extensions ...' option may be used for specifying the
827 [Massimiliano Pala <madwolf@comune.modena.it>]
829 *) Change PKCS12_key_gen_asc() so it can cope with non null
830 terminated strings whose length is passed in the passlen
831 parameter, for example from PEM callbacks. This was done
832 by adding an extra length parameter to asc2uni().
833 [Steve Henson, reported by <oddissey@samsung.co.kr>]
835 +) New OCSP utility. Allows OCSP requests to be generated or
836 read. The request can be sent to a responder and the output
837 parsed, outputed or printed in text form. Not complete yet:
838 still needs to check the OCSP response validity.
841 +) New subcommands for 'openssl ca':
842 'openssl ca -status <serial>' prints the status of the cert with
843 the given serial number (according to the index file).
844 'openssl ca -updatedb' updates the expiry status of certificates
846 [Massimiliano Pala <madwolf@comune.modena.it>]
848 +) New '-newreq-nodes' command option to CA.pl. This is like
849 '-newreq', but calls 'openssl req' with the '-nodes' option
850 so that the resulting key is not encrypted.
851 [Damien Miller <djm@mindrot.org>]
853 +) New configuration for the GNU Hurd.
854 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
856 +) Initial code to implement OCSP basic response verify. This
857 is currently incomplete. Currently just finds the signer's
858 certificate and verifies the signature on the response.
861 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
862 value of OPENSSLDIR. This is available via the new '-d' option
863 to 'openssl version', and is also included in 'openssl version -a'.
866 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
867 call failed, free the DSA structure.
870 +) Allowing defining memory allocation callbacks that will be given
871 file name and line number information in additional arguments
872 (a const char* and an int). The basic functionality remains, as
873 well as the original possibility to just replace malloc(),
874 realloc() and free() by functions that do not know about these
875 additional arguments. To register and find out the current
876 settings for extended allocation functions, the following
877 functions are provided:
879 CRYPTO_set_mem_ex_functions
880 CRYPTO_set_locked_mem_ex_functions
881 CRYPTO_get_mem_ex_functions
882 CRYPTO_get_locked_mem_ex_functions
884 These work the same way as CRYPTO_set_mem_functions and friends.
885 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
886 extended allocation function is enabled.
887 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
888 a conventional allocation function is enabled.
889 [Richard Levitte, Bodo Moeller]
891 *) Fix to uni2asc() to cope with zero length Unicode strings.
892 These are present in some PKCS#12 files.
895 +) Finish off removing the remaining LHASH function pointer casts.
896 There should no longer be any prototype-casting required when using
897 the LHASH abstraction, and any casts that remain are "bugs". See
898 the callback types and macros at the head of lhash.h for details
899 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
902 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
903 If /dev/[u]random devices are not available or do not return enough
904 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
906 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
907 /etc/entropy will be queried once each in this sequence, quering stops
908 when enough entropy was collected without querying more sockets.
911 +) Change the Unix RAND_poll() variant to be able to poll several
912 random devices, as specified by DEVRANDOM, until a sufficient amount
913 of data has been collected. We spend at most 10 ms on each file
914 (select timeout) and read in non-blocking mode. DEVRANDOM now
915 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
916 (previously it was just the string "/dev/urandom"), so on typical
917 platforms the 10 ms delay will never occur.
918 Also separate out the Unix variant to its own file, rand_unix.c.
919 For VMS, there's a currently-empty rand_vms.c.
922 +) Move OCSP client related routines to ocsp_cl.c. These
923 provide utility functions which an application needing
924 to issue a request to an OCSP responder and analyse the
925 response will typically need: as opposed to those which an
926 OCSP responder itself would need which will be added later.
928 OCSP_request_sign() signs an OCSP request with an API similar
929 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
930 response. OCSP_response_get1_basic() extracts basic response
931 from response. OCSP_resp_find_status(): finds and extracts status
932 information from an OCSP_CERTID structure (which will be created
933 when the request structure is built). These are built from lower
934 level functions which work on OCSP_SINGLERESP structures but
935 wont normally be used unless the application wishes to examine
936 extensions in the OCSP response for example.
938 Replace nonce routines with a pair of functions.
939 OCSP_request_add1_nonce() adds a nonce value and optionally
940 generates a random value. OCSP_check_nonce() checks the
941 validity of the nonce in an OCSP response.
944 +) Change function OCSP_request_add() to OCSP_request_add0_id().
945 This doesn't copy the supplied OCSP_CERTID and avoids the
946 need to free up the newly created id. Change return type
947 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
948 This can then be used to add extensions to the request.
949 Deleted OCSP_request_new(), since most of its functionality
950 is now in OCSP_REQUEST_new() (and the case insensitive name
951 clash) apart from the ability to set the request name which
952 will be added elsewhere.
955 +) Update OCSP API. Remove obsolete extensions argument from
956 various functions. Extensions are now handled using the new
957 OCSP extension code. New simple OCSP HTTP function which
958 can be used to send requests and parse the response.
961 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
962 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
963 uses the special reorder version of SET OF to sort the attributes
964 and reorder them to match the encoded order. This resolves a long
965 standing problem: a verify on a PKCS7 structure just after signing
966 it used to fail because the attribute order did not match the
967 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
968 it uses the received order. This is necessary to tolerate some broken
969 software that does not order SET OF. This is handled by encoding
970 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
971 to produce the required SET OF.
974 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
975 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
976 files to get correct declarations of the ASN.1 item variables.
979 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
980 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
981 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
982 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
983 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
984 ASN1_ITEM and no wrapper functions.
987 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
988 replace the old function pointer based I/O routines. Change most of
989 the *_d2i_bio() and *_d2i_fp() functions to use these.
992 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
993 lines, recognice more "algorithms" that can be deselected, and make
994 it complain about algorithm deselection that isn't recognised.
997 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
998 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
999 to use new functions. Add NO_ASN1_OLD which can be set to remove
1000 some old style ASN1 functions: this can be used to determine if old
1001 code will still work when these eventually go away.
1004 +) New extension functions for OCSP structures, these follow the
1005 same conventions as certificates and CRLs.
1008 +) New function X509V3_add1_i2d(). This automatically encodes and
1009 adds an extension. Its behaviour can be customised with various
1010 flags to append, replace or delete. Various wrappers added for
1011 certifcates and CRLs.
1014 +) Fix to avoid calling the underlying ASN1 print routine when
1015 an extension cannot be parsed. Correct a typo in the
1016 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1019 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1020 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1021 when writing a 32767 byte record.
1022 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1024 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1025 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1027 (RSA objects have a reference count access to which is protected
1028 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1029 so they are meant to be shared between threads.)
1030 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1031 "Reddie, Steven" <Steven.Reddie@ca.com>]
1033 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1034 entries for variables.
1037 *) Fix a deadlock in CRYPTO_mem_leaks().
1040 +) Add functionality to apps/openssl.c for detecting locking
1041 problems: As the program is single-threaded, all we have
1042 to do is register a locking callback using an array for
1043 storing which locks are currently held by the program.
1046 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1047 SSL_get_ex_data_X509_STORE_idx(), which is used in
1048 ssl_verify_cert_chain() and thus can be called at any time
1049 during TLS/SSL handshakes so that thread-safety is essential.
1050 Unfortunately, the ex_data design is not at all suited
1051 for multi-threaded use, so it probably should be abolished.
1054 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1055 [Broadcom, tweaked and integrated by Geoff Thorpe]
1057 +) Move common extension printing code to new function
1058 X509V3_print_extensions(). Reorganise OCSP print routines and
1059 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1062 +) New function X509_signature_print() to remove duplication in some
1066 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1067 set (this was treated exactly the same as SET OF previously). This
1068 is used to reorder the STACK representing the structure to match the
1069 encoding. This will be used to get round a problem where a PKCS7
1070 structure which was signed could not be verified because the STACK
1071 order did not reflect the encoded order.
1074 +) Reimplement the OCSP ASN1 module using the new code.
1077 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1078 for its ASN1 operations. The old style function pointers still exist
1079 for now but they will eventually go away.
1082 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1083 completely replaces the old ASN1 functionality with a table driven
1084 encoder and decoder which interprets an ASN1_ITEM structure describing
1085 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1086 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1087 has also been converted to the new form.
1090 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1091 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1092 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1093 for negative moduli.
1096 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1097 of not touching the result's sign bit.
1100 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1104 +) Changed the LHASH code to use prototypes for callbacks, and created
1105 macros to declare and implement thin (optionally static) functions
1106 that provide type-safety and avoid function pointer casting for the
1107 type-specific callbacks.
1110 *) Use better test patterns in bntest.
1113 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1115 [Veers Staats <staatsvr@asc.hpc.mil>,
1116 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1118 *) rand_win.c fix for Borland C.
1121 *) BN_rshift bugfix for n == 0.
1124 +) Reformat the FAQ so the different questions and answers can be divided
1125 in sections depending on the subject.
1128 +) Have the zlib compression code load ZLIB.DLL dynamically under
1132 +) New function BN_mod_sqrt for computing square roots modulo a prime
1133 (using the probabilistic Tonelli-Shanks algorithm unless
1134 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1135 be handled deterministically).
1136 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1138 *) Add a 'bctest' script that checks for some known 'bc' bugs
1139 so that 'make test' does not abort just because 'bc' is broken.
1142 *) Store verify_result within SSL_SESSION also for client side to
1143 avoid potential security hole. (Re-used sessions on the client side
1144 always resulted in verify_result==X509_V_OK, not using the original
1145 result of the server certificate verification.)
1148 +) Make BN_mod_inverse faster by explicitly handling small quotients
1149 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1150 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1153 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1154 SSL3_RT_APPLICATION_DATA, return 0.
1155 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1159 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1160 releases, have been re-implemented by renaming the previous
1161 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1162 and ssl3_read_internal, respectively, and adding 'peek' parameters
1163 to them. The new ssl[23]_{read,peek} functions are calls to
1164 ssl[23]_read_internal with the 'peek' flag set appropriately.
1165 A 'peek' parameter has also been added to ssl3_read_bytes, which
1166 does the actual work for ssl3_read_internal.
1169 +) New function BN_kronecker.
1172 +) Fix BN_gcd so that it works on negative inputs; the result is
1173 positive unless both parameters are zero.
1174 Previously something reasonably close to an infinite loop was
1175 possible because numbers could be growing instead of shrinking
1176 in the implementation of Euclid's algorithm.
1179 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1180 sign of the number in question.
1182 Fix BN_is_word(a,w) to work correctly for w == 0.
1184 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1185 because its test if the absolute value of 'a' equals 'w'.
1186 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1187 it exists mostly for use in the implementations of BN_is_zero(),
1188 BN_is_one(), and BN_is_word().
1191 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1192 the method-specific "init()" handler. Also clean up ex_data after
1193 calling the method-specific "finish()" handler. Previously, this was
1194 happening the other way round.
1197 +) New function BN_swap.
1200 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1201 the exponentiation functions are more likely to produce reasonable
1202 results on negative inputs.
1205 +) Change BN_mod_mul so that the result is always non-negative.
1206 Previously, it could be negative if one of the factors was negative;
1207 I don't think anyone really wanted that behaviour.
1210 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1211 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1212 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1213 and add new functions:
1222 BN_mod_lshift1_quick
1226 These functions always generate non-negative results.
1228 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1229 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1231 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1232 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1233 be reduced modulo m.
1234 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1236 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1237 was actually never needed) and in BN_mul(). The removal in BN_mul()
1238 required a small change in bn_mul_part_recursive() and the addition
1239 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1240 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1241 bn_sub_words() and bn_add_words() except they take arrays with
1245 +) In 'openssl passwd', verify passwords read from the terminal
1246 unless the '-salt' option is used (which usually means that
1247 verification would just waste user's time since the resulting
1248 hash is going to be compared with some given password hash)
1249 or the new '-noverify' option is used.
1251 This is an incompatible change, but it does not affect
1252 non-interactive use of 'openssl passwd' (passwords on the command
1253 line, '-stdin' option, '-in ...' option) and thus should not
1257 +) Remove all references to RSAref, since there's no more need for it.
1260 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1261 The previous value, 12, was not always sufficient for BN_mod_exp().
1264 +) Make DSO load along a path given through an environment variable
1265 (SHLIB_PATH) with shl_load().
1268 +) Constify the ENGINE code as a result of BIGNUM constification.
1269 Also constify the RSA code and most things related to it. In a
1270 few places, most notable in the depth of the ASN.1 code, ugly
1271 casts back to non-const were required (to be solved at a later
1275 +) Make it so the openssl application has all engines loaded by default.
1278 +) Constify the BIGNUM routines a little more.
1281 *) Make sure that shared libraries get the internal name engine with
1282 the full version number and not just 0. This should mark the
1283 shared libraries as not backward compatible. Of course, this should
1284 be changed again when we can guarantee backward binary compatibility.
1287 +) Add the following functions:
1289 ENGINE_load_cswift()
1291 ENGINE_load_atalla()
1293 ENGINE_load_builtin_engines()
1295 That way, an application can itself choose if external engines that
1296 are built-in in OpenSSL shall ever be used or not. The benefit is
1297 that applications won't have to be linked with libdl or other dso
1298 libraries unless it's really needed.
1300 Changed 'openssl engine' to load all engines on demand.
1301 Changed the engine header files to avoid the duplication of some
1302 declarations (they differed!).
1305 +) 'openssl engine' can now list capabilities.
1308 +) Better error reporting in 'openssl engine'.
1311 +) Never call load_dh_param(NULL) in s_server.
1314 +) Add engine application. It can currently list engines by name and
1315 identity, and test if they are actually available.
1318 +) Add support for shared libraries under Irix.
1319 [Albert Chin-A-Young <china@thewrittenword.com>]
1321 +) Improve RPM specification file by forcing symbolic linking and making
1322 sure the installed documentation is also owned by root.root.
1323 [Damien Miller <djm@mindrot.org>]
1325 +) Add configuration option to build on Linux on both big-endian and
1327 [Ralf Baechle <ralf@uni-koblenz.de>]
1329 +) Give the OpenSSL applications more possibilities to make use of
1330 keys (public as well as private) handled by engines.
1333 +) Add OCSP code that comes from CertCo.
1336 +) Add VMS support for the Rijndael code.
1339 +) Added untested support for Nuron crypto accelerator.
1342 +) Add support for external cryptographic devices. This code was
1343 previously distributed separately as the "engine" branch.
1344 [Geoff Thorpe, Richard Levitte]
1346 +) Rework the filename-translation in the DSO code. It is now possible to
1347 have far greater control over how a "name" is turned into a filename
1348 depending on the operating environment and any oddities about the
1349 different shared library filenames on each system.
1352 +) Support threads on FreeBSD-elf in Configure.
1355 +) Add the possibility to create shared libraries on HP-UX
1358 +) Fix for SHA1 assembly problem with MASM: it produces
1359 warnings about corrupt line number information when assembling
1360 with debugging information. This is caused by the overlapping
1362 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1365 NCONF_get_number() has no error checking at all. As a replacement,
1366 NCONF_get_number_e() is defined (_e for "error checking") and is
1367 promoted strongly. The old NCONF_get_number is kept around for
1368 binary backward compatibility.
1369 Make it possible for methods to load from something other than a BIO,
1370 by providing a function pointer that is given a name instead of a BIO.
1371 For example, this could be used to load configuration data from an
1375 *) Fix typo in get_cert_by_subject() in by_dir.c
1376 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1378 *) Rework the system to generate shared libraries:
1380 - Make note of the expected extension for the shared libraries and
1381 if there is a need for symbolic links from for example libcrypto.so.0
1382 to libcrypto.so.0.9.7. There is extended info in Configure for
1385 - Make as few rebuilds of the shared libraries as possible.
1387 - Still avoid linking the OpenSSL programs with the shared libraries.
1389 - When installing, install the shared libraries separately from the
1393 +) Fix for non blocking accept BIOs. Added new I/O special reason
1394 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1395 with non blocking I/O was not possible because no retry code was
1396 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1400 +) Added the beginnings of Rijndael support.
1403 +) Fix for bug in DirectoryString mask setting. Add support for
1404 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1405 to allow certificate printing to more controllable, additional
1406 'certopt' option to 'x509' to allow new printing options to be
1410 +) Clean old EAY MD5 hack from e_os.h.
1413 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1415 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1416 and not in SSL_clear because the latter is also used by the
1417 accept/connect functions; previously, the settings made by
1418 SSL_set_read_ahead would be lost during the handshake.
1419 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1421 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1422 Previously, it would create entries for disableed algorithms no
1426 *) Added several new manual pages for SSL_* function.
1429 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1431 *) In ssl23_get_client_hello, generate an error message when faced
1432 with an initial SSL 3.0/TLS record that is too small to contain the
1433 first two bytes of the ClientHello message, i.e. client_version.
1434 (Note that this is a pathologic case that probably has never happened
1435 in real life.) The previous approach was to use the version number
1436 from the record header as a substitute; but our protocol choice
1437 should not depend on that one because it is not authenticated
1438 by the Finished messages.
1441 *) More robust randomness gathering functions for Windows.
1442 [Jeffrey Altman <jaltman@columbia.edu>]
1444 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1445 not set then we don't setup the error code for issuer check errors
1446 to avoid possibly overwriting other errors which the callback does
1447 handle. If an application does set the flag then we assume it knows
1448 what it is doing and can handle the new informational codes
1452 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1453 a general "ANY" type, as such it should be able to decode anything
1454 including tagged types. However it didn't check the class so it would
1455 wrongly interpret tagged types in the same way as their universal
1456 counterpart and unknown types were just rejected. Changed so that the
1457 tagged and unknown types are handled in the same way as a SEQUENCE:
1458 that is the encoding is stored intact. There is also a new type
1459 "V_ASN1_OTHER" which is used when the class is not universal, in this
1460 case we have no idea what the actual type is so we just lump them all
1464 *) On VMS, stdout may very well lead to a file that is written to
1465 in a record-oriented fashion. That means that every write() will
1466 write a separate record, which will be read separately by the
1467 programs trying to read from it. This can be very confusing.
1469 The solution is to put a BIO filter in the way that will buffer
1470 text until a linefeed is reached, and then write everything a
1471 line at a time, so every record written will be an actual line,
1472 not chunks of lines and not (usually doesn't happen, but I've
1473 seen it once) several lines in one record. BIO_f_linebuffer() is
1476 Currently, it's a VMS-only method, because that's where it has
1477 been tested well enough.
1480 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
1481 it can return incorrect results.
1482 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
1483 but it was in 0.9.6-beta[12].)
1486 *) Disable the check for content being present when verifying detached
1487 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
1488 include zero length content when signing messages.
1491 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
1492 BIO_ctrl (for BIO pairs).
1495 *) Add DSO method for VMS.
1498 *) Bug fix: Montgomery multiplication could produce results with the
1502 *) Add RPM specification openssl.spec and modify it to build three
1503 packages. The default package contains applications, application
1504 documentation and run-time libraries. The devel package contains
1505 include files, static libraries and function documentation. The
1506 doc package contains the contents of the doc directory. The original
1507 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1510 *) Add a large number of documentation files for many SSL routines.
1511 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1513 *) Add a configuration entry for Sony News 4.
1514 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1516 *) Don't set the two most significant bits to one when generating a
1517 random number < q in the DSA library.
1520 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1521 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1522 the underlying transport is blocking) if a handshake took place.
1523 (The default behaviour is needed by applications such as s_client
1524 and s_server that use select() to determine when to use SSL_read;
1525 but for applications that know in advance when to expect data, it
1526 just makes things more complicated.)
1529 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1533 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1534 work better on such systems.
1535 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1537 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1538 Update PKCS12_parse() so it copies the friendlyName and the
1539 keyid to the certificates aux info.
1542 *) Fix bug in PKCS7_verify() which caused an infinite loop
1543 if there was more than one signature.
1544 [Sven Uszpelkat <su@celocom.de>]
1546 *) Major change in util/mkdef.pl to include extra information
1547 about each symbol, as well as presentig variables as well
1548 as functions. This change means that there's n more need
1549 to rebuild the .num files when some algorithms are excluded.
1552 *) Allow the verify time to be set by an application,
1553 rather than always using the current time.
1556 *) Phase 2 verify code reorganisation. The certificate
1557 verify code now looks up an issuer certificate by a
1558 number of criteria: subject name, authority key id
1559 and key usage. It also verifies self signed certificates
1560 by the same criteria. The main comparison function is
1561 X509_check_issued() which performs these checks.
1563 Lot of changes were necessary in order to support this
1564 without completely rewriting the lookup code.
1566 Authority and subject key identifier are now cached.
1568 The LHASH 'certs' is X509_STORE has now been replaced
1569 by a STACK_OF(X509_OBJECT). This is mainly because an
1570 LHASH can't store or retrieve multiple objects with
1571 the same hash value.
1573 As a result various functions (which were all internal
1574 use only) have changed to handle the new X509_STORE
1575 structure. This will break anything that messed round
1576 with X509_STORE internally.
1578 The functions X509_STORE_add_cert() now checks for an
1579 exact match, rather than just subject name.
1581 The X509_STORE API doesn't directly support the retrieval
1582 of multiple certificates matching a given criteria, however
1583 this can be worked round by performing a lookup first
1584 (which will fill the cache with candidate certificates)
1585 and then examining the cache for matches. This is probably
1586 the best we can do without throwing out X509_LOOKUP
1587 entirely (maybe later...).
1589 The X509_VERIFY_CTX structure has been enhanced considerably.
1591 All certificate lookup operations now go via a get_issuer()
1592 callback. Although this currently uses an X509_STORE it
1593 can be replaced by custom lookups. This is a simple way
1594 to bypass the X509_STORE hackery necessary to make this
1595 work and makes it possible to use more efficient techniques
1596 in future. A very simple version which uses a simple
1597 STACK for its trusted certificate store is also provided
1598 using X509_STORE_CTX_trusted_stack().
1600 The verify_cb() and verify() callbacks now have equivalents
1601 in the X509_STORE_CTX structure.
1603 X509_STORE_CTX also has a 'flags' field which can be used
1604 to customise the verify behaviour.
1607 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1608 excludes S/MIME capabilities.
1611 *) When a certificate request is read in keep a copy of the
1612 original encoding of the signed data and use it when outputing
1613 again. Signatures then use the original encoding rather than
1614 a decoded, encoded version which may cause problems if the
1615 request is improperly encoded.
1618 *) For consistency with other BIO_puts implementations, call
1619 buffer_write(b, ...) directly in buffer_puts instead of calling
1622 In BIO_puts, increment b->num_write as in BIO_write.
1623 [Peter.Sylvester@EdelWeb.fr]
1625 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1626 BN_zero, we may not return a BIGNUM with an array consisting of
1630 *) Avoid calling abort() from within the library when problems are
1631 detected, except if preprocessor symbols have been defined
1632 (such as REF_CHECK, BN_DEBUG etc.).
1635 *) New openssl application 'rsautl'. This utility can be
1636 used for low level RSA operations. DER public key
1637 BIO/fp routines also added.
1640 *) New Configure entry and patches for compiling on QNX 4.
1641 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1643 *) A demo state-machine implementation was sponsored by
1644 Nuron (http://www.nuron.com/) and is now available in
1645 demos/state_machine.
1648 *) New options added to the 'dgst' utility for signature
1649 generation and verification.
1652 *) Unrecognized PKCS#7 content types are now handled via a
1653 catch all ASN1_TYPE structure. This allows unsupported
1654 types to be stored as a "blob" and an application can
1655 encode and decode it manually.
1658 *) Fix various signed/unsigned issues to make a_strex.c
1660 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1662 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1663 length if passed a buffer. ASN1_INTEGER_to_BN failed
1664 if passed a NULL BN and its argument was negative.
1665 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1667 *) Modification to PKCS#7 encoding routines to output definite
1668 length encoding. Since currently the whole structures are in
1669 memory there's not real point in using indefinite length
1670 constructed encoding. However if OpenSSL is compiled with
1671 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1674 *) Added BIO_vprintf() and BIO_vsnprintf().
1677 *) Added more prefixes to parse for in the the strings written
1678 through a logging bio, to cover all the levels that are available
1679 through syslog. The prefixes are now:
1681 PANIC, EMERG, EMR => LOG_EMERG
1682 ALERT, ALR => LOG_ALERT
1683 CRIT, CRI => LOG_CRIT
1684 ERROR, ERR => LOG_ERR
1685 WARNING, WARN, WAR => LOG_WARNING
1686 NOTICE, NOTE, NOT => LOG_NOTICE
1687 INFO, INF => LOG_INFO
1688 DEBUG, DBG => LOG_DEBUG
1690 and as before, if none of those prefixes are present at the
1691 beginning of the string, LOG_ERR is chosen.
1693 On Win32, the LOG_* levels are mapped according to this:
1695 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1696 LOG_WARNING => EVENTLOG_WARNING_TYPE
1697 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1701 *) Made it possible to reconfigure with just the configuration
1702 argument "reconf" or "reconfigure". The command line arguments
1703 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1704 and are retrieved from there when reconfiguring.
1708 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1710 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1713 *) The obj_dat.pl script was messing up the sorting of object
1714 names. The reason was that it compared the quoted version
1715 of strings as a result "OCSP" > "OCSP Signing" because
1716 " > SPACE. Changed script to store unquoted versions of
1717 names and add quotes on output. It was also omitting some
1718 names from the lookup table if they were given a default
1719 value (that is if SN is missing it is given the same
1720 value as LN and vice versa), these are now added on the
1721 grounds that if an object has a name we should be able to
1722 look it up. Finally added warning output when duplicate
1723 short or long names are found.
1726 *) Changes needed for Tandem NSK.
1727 [Scott Uroff <scott@xypro.com>]
1729 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1730 RSA_padding_check_SSLv23(), special padding was never detected
1731 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1732 version rollback attacks was not effective.
1734 In s23_clnt.c, don't use special rollback-attack detection padding
1735 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1736 client; similarly, in s23_srvr.c, don't do the rollback check if
1737 SSL 2.0 is the only protocol enabled in the server.
1740 *) Make it possible to get hexdumps of unprintable data with 'openssl
1741 asn1parse'. By implication, the functions ASN1_parse_dump() and
1742 BIO_dump_indent() are added.
1745 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1746 these print out strings and name structures based on various
1747 flags including RFC2253 support and proper handling of
1748 multibyte characters. Added options to the 'x509' utility
1749 to allow the various flags to be set.
1752 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1753 Also change the functions X509_cmp_current_time() and
1754 X509_gmtime_adj() work with an ASN1_TIME structure,
1755 this will enable certificates using GeneralizedTime in validity
1756 dates to be checked.
1759 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1760 negative public key encodings) on by default,
1761 NO_NEG_PUBKEY_BUG can be set to disable it.
1764 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1765 content octets. An i2c_ASN1_OBJECT is unnecessary because
1766 the encoding can be trivially obtained from the structure.
1769 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1770 not read locks (CRYPTO_r_[un]lock).
1773 *) A first attempt at creating official support for shared
1774 libraries through configuration. I've kept it so the
1775 default is static libraries only, and the OpenSSL programs
1776 are always statically linked for now, but there are
1777 preparations for dynamic linking in place.
1778 This has been tested on Linux and True64.
1781 *) Randomness polling function for Win9x, as described in:
1782 Peter Gutmann, Software Generation of Practically Strong
1786 *) Fix so PRNG is seeded in req if using an already existing
1790 *) New options to smime application. -inform and -outform
1791 allow alternative formats for the S/MIME message including
1792 PEM and DER. The -content option allows the content to be
1793 specified separately. This should allow things like Netscape
1794 form signing output easier to verify.
1797 *) Fix the ASN1 encoding of tags using the 'long form'.
1800 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1801 STRING types. These convert content octets to and from the
1802 underlying type. The actual tag and length octets are
1803 already assumed to have been read in and checked. These
1804 are needed because all other string types have virtually
1805 identical handling apart from the tag. By having versions
1806 of the ASN1 functions that just operate on content octets
1807 IMPLICIT tagging can be handled properly. It also allows
1808 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1809 and ASN1_INTEGER are identical apart from the tag.
1812 *) Change the handling of OID objects as follows:
1814 - New object identifiers are inserted in objects.txt, following
1815 the syntax given in objects.README.
1816 - objects.pl is used to process obj_mac.num and create a new
1818 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1821 This is currently kind of a hack, and the perl code in objects.pl
1822 isn't very elegant, but it works as I intended. The simplest way
1823 to check that it worked correctly is to look in obj_dat.h and
1824 check the array nid_objs and make sure the objects haven't moved
1825 around (this is important!). Additions are OK, as well as
1826 consistent name changes.
1829 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1832 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1833 The given file adds to whatever has already been seeded into the
1834 random pool through the RANDFILE configuration file option or
1835 environment variable, or the default random state file.
1838 *) mkstack.pl now sorts each macro group into lexical order.
1839 Previously the output order depended on the order the files
1840 appeared in the directory, resulting in needless rewriting
1844 *) Patches to make OpenSSL compile under Win32 again. Mostly
1845 work arounds for the VC++ problem that it treats func() as
1846 func(void). Also stripped out the parts of mkdef.pl that
1847 added extra typesafe functions: these no longer exist.
1850 *) Reorganisation of the stack code. The macros are now all
1851 collected in safestack.h . Each macro is defined in terms of
1852 a "stack macro" of the form SKM_<name>(type, a, b). The
1853 DEBUG_SAFESTACK is now handled in terms of function casts,
1854 this has the advantage of retaining type safety without the
1855 use of additional functions. If DEBUG_SAFESTACK is not defined
1856 then the non typesafe macros are used instead. Also modified the
1857 mkstack.pl script to handle the new form. Needs testing to see
1858 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1859 the default if no major problems. Similar behaviour for ASN1_SET_OF
1860 and PKCS12_STACK_OF.
1863 *) When some versions of IIS use the 'NET' form of private key the
1864 key derivation algorithm is different. Normally MD5(password) is
1865 used as a 128 bit RC4 key. In the modified case
1866 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1867 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1868 as the old Netscape_RSA functions except they have an additional
1869 'sgckey' parameter which uses the modified algorithm. Also added
1870 an -sgckey command line option to the rsa utility. Thanks to
1871 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1872 algorithm to openssl-dev.
1875 *) The evp_local.h macros were using 'c.##kname' which resulted in
1876 invalid expansion on some systems (SCO 5.0.5 for example).
1877 Corrected to 'c.kname'.
1878 [Phillip Porch <root@theporch.com>]
1880 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1881 a STACK of email addresses from a certificate or request, these look
1882 in the subject name and the subject alternative name extensions and
1883 omit any duplicate addresses.
1886 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1887 This makes DSA verification about 2 % faster.
1890 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1891 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1892 plus overhead for 1024 bit moduli).
1893 This makes exponentiations about 0.5 % faster for 1024 bit
1894 exponents (as measured by "openssl speed rsa2048").
1897 *) Rename memory handling macros to avoid conflicts with other
1899 Malloc => OPENSSL_malloc
1900 Malloc_locked => OPENSSL_malloc_locked
1901 Realloc => OPENSSL_realloc
1902 Free => OPENSSL_free
1905 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1906 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1909 *) CygWin32 support.
1910 [John Jarvie <jjarvie@newsguy.com>]
1912 *) The type-safe stack code has been rejigged. It is now only compiled
1913 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1914 by default all type-specific stack functions are "#define"d back to
1915 standard stack functions. This results in more streamlined output
1916 but retains the type-safety checking possibilities of the original
1920 *) The STACK code has been cleaned up, and certain type declarations
1921 that didn't make a lot of sense have been brought in line. This has
1922 also involved a cleanup of sorts in safestack.h to more correctly
1923 map type-safe stack functions onto their plain stack counterparts.
1924 This work has also resulted in a variety of "const"ifications of
1925 lots of the code, especially "_cmp" operations which should normally
1926 be prototyped with "const" parameters anyway.
1929 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1930 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1931 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1932 where all of 'md' is used each time the PRNG is used, but 'state'
1933 is used only indexed by a cyclic counter. As entropy may not be
1934 well distributed from the beginning, 'md' is important as a
1935 chaining variable. However, the output function chains only half
1936 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1937 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1938 in all of 'state' being rewritten, with the new values depending
1939 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1942 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1943 the handshake is continued after ssl_verify_cert_chain();
1944 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1945 can lead to 'unexplainable' connection aborts later.
1946 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1948 *) Major EVP API cipher revision.
1949 Add hooks for extra EVP features. This allows various cipher
1950 parameters to be set in the EVP interface. Support added for variable
1951 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1952 setting of RC2 and RC5 parameters.
1954 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1957 Remove lots of duplicated code from the EVP library. For example *every*
1958 cipher init() function handles the 'iv' in the same way according to the
1959 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1960 for CFB and OFB modes they zero ctx->num.
1962 New functionality allows removal of S/MIME code RC2 hack.
1964 Most of the routines have the same form and so can be declared in terms
1967 By shifting this to the top level EVP_CipherInit() it can be removed from
1968 all individual ciphers. If the cipher wants to handle IVs or keys
1969 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1972 Change lots of functions like EVP_EncryptUpdate() to now return a
1973 value: although software versions of the algorithms cannot fail
1974 any installed hardware versions can.
1977 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1978 this option is set, tolerate broken clients that send the negotiated
1979 protocol version number instead of the requested protocol version
1983 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1984 i.e. non-zero for export ciphersuites, zero otherwise.
1985 Previous versions had this flag inverted, inconsistent with
1986 rsa_tmp_cb (..._TMP_RSA_CB).
1987 [Bodo Moeller; problem reported by Amit Chopra]
1989 *) Add missing DSA library text string. Work around for some IIS
1990 key files with invalid SEQUENCE encoding.
1993 *) Add a document (doc/standards.txt) that list all kinds of standards
1994 and so on that are implemented in OpenSSL.
1997 *) Enhance c_rehash script. Old version would mishandle certificates
1998 with the same subject name hash and wouldn't handle CRLs at all.
1999 Added -fingerprint option to crl utility, to support new c_rehash
2003 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2006 *) Fix for SSL server purpose checking. Server checking was
2007 rejecting certificates which had extended key usage present
2008 but no ssl client purpose.
2009 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2011 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2012 is a little unclear about how a blank password is handled.
2013 Since the password in encoded as a BMPString with terminating
2014 double NULL a zero length password would end up as just the
2015 double NULL. However no password at all is different and is
2016 handled differently in the PKCS#12 key generation code. NS
2017 treats a blank password as zero length. MSIE treats it as no
2018 password on export: but it will try both on import. We now do
2019 the same: PKCS12_parse() tries zero length and no password if
2020 the password is set to "" or NULL (NULL is now a valid password:
2021 it wasn't before) as does the pkcs12 application.
2024 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2025 perror when PEM_read_bio_X509_REQ fails, the error message must
2026 be obtained from the error queue.
2029 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2030 it in ERR_remove_state if appropriate, and change ERR_get_state
2031 accordingly to avoid race conditions (this is necessary because
2032 thread_hash is no longer constant once set).
2035 *) Bugfix for linux-elf makefile.one.
2038 *) RSA_get_default_method() will now cause a default
2039 RSA_METHOD to be chosen if one doesn't exist already.
2040 Previously this was only set during a call to RSA_new()
2041 or RSA_new_method(NULL) meaning it was possible for
2042 RSA_get_default_method() to return NULL.
2045 *) Added native name translation to the existing DSO code
2046 that will convert (if the flag to do so is set) filenames
2047 that are sufficiently small and have no path information
2048 into a canonical native form. Eg. "blah" converted to
2049 "libblah.so" or "blah.dll" etc.
2052 *) New function ERR_error_string_n(e, buf, len) which is like
2053 ERR_error_string(e, buf), but writes at most 'len' bytes
2054 including the 0 terminator. For ERR_error_string_n, 'buf'
2056 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2058 *) CONF library reworked to become more general. A new CONF
2059 configuration file reader "class" is implemented as well as a
2060 new functions (NCONF_*, for "New CONF") to handle it. The now
2061 old CONF_* functions are still there, but are reimplemented to
2062 work in terms of the new functions. Also, a set of functions
2063 to handle the internal storage of the configuration data is
2064 provided to make it easier to write new configuration file
2065 reader "classes" (I can definitely see something reading a
2066 configuration file in XML format, for example), called _CONF_*,
2067 or "the configuration storage API"...
2069 The new configuration file reading functions are:
2071 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2072 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2074 NCONF_default, NCONF_WIN32
2076 NCONF_dump_fp, NCONF_dump_bio
2078 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2079 NCONF_new creates a new CONF object. This works in the same way
2080 as other interfaces in OpenSSL, like the BIO interface.
2081 NCONF_dump_* dump the internal storage of the configuration file,
2082 which is useful for debugging. All other functions take the same
2083 arguments as the old CONF_* functions wth the exception of the
2084 first that must be a `CONF *' instead of a `LHASH *'.
2086 To make it easer to use the new classes with the old CONF_* functions,
2087 the function CONF_set_default_method is provided.
2090 *) Add '-tls1' option to 'openssl ciphers', which was already
2091 mentioned in the documentation but had not been implemented.
2092 (This option is not yet really useful because even the additional
2093 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2096 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2097 OpenSSL-based applications) load shared libraries and bind to
2098 them in a portable way.
2099 [Geoff Thorpe, with contributions from Richard Levitte]
2101 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2103 *) Make sure _lrotl and _lrotr are only used with MSVC.
2105 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2106 (the default implementation of RAND_status).
2108 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2109 to '-clrext' (= clear extensions), as intended and documented.
2110 [Bodo Moeller; inconsistency pointed out by Michael Attili
2113 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2114 was larger than the MD block size.
2115 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2117 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2118 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2119 using the passed key: if the passed key was a private key the result
2120 of X509_print(), for example, would be to print out all the private key
2124 *) des_quad_cksum() byte order bug fix.
2125 [Ulf Möller, using the problem description in krb4-0.9.7, where
2126 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2128 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2130 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2132 *) For easily testing in shell scripts whether some command
2133 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2134 returns with exit code 0 iff no command of the given name is available.
2135 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2136 the output goes to stdout and nothing is printed to stderr.
2137 Additional arguments are always ignored.
2139 Since for each cipher there is a command of the same name,
2140 the 'no-cipher' compilation switches can be tested this way.
2142 ('openssl no-XXX' is not able to detect pseudo-commands such
2143 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2146 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2149 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2150 is set; it will be thrown away anyway because each handshake creates
2152 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2153 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2154 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2155 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2158 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2159 'Q' and 'R' lose their special meanings (quit/renegotiate).
2160 This is part of what -quiet does; unlike -quiet, -ign_eof
2161 does not suppress any output.
2164 *) Add compatibility options to the purpose and trust code. The
2165 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2166 accepts a certificate or CA, this was the previous behaviour,
2167 with all the associated security issues.
2169 X509_TRUST_COMPAT is the old trust behaviour: only and
2170 automatically trust self signed roots in certificate store. A
2171 new trust setting X509_TRUST_DEFAULT is used to specify that
2172 a purpose has no associated trust setting and it should instead
2173 use the value in the default purpose.
2176 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2177 and fix a memory leak.
2180 *) In util/mkerr.pl (which implements 'make errors'), preserve
2181 reason strings from the previous version of the .c file, as
2182 the default to have only downcase letters (and digits) in
2183 automatically generated reasons codes is not always appropriate.
2186 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2187 using strerror. Previously, ERR_reason_error_string() returned
2188 library names as reason strings for SYSerr; but SYSerr is a special
2189 case where small numbers are errno values, not library numbers.
2192 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2193 converts DSA parameters into DH parameters. (When creating parameters,
2194 DSA_generate_parameters is used.)
2197 *) Include 'length' (recommended exponent length) in C code generated
2198 by 'openssl dhparam -C'.
2201 *) The second argument to set_label in perlasm was already being used
2202 so couldn't be used as a "file scope" flag. Moved to third argument
2206 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2207 instead of RAND_bytes for encryption IVs and salts.
2210 *) Include RAND_status() into RAND_METHOD instead of implementing
2211 it only for md_rand.c Otherwise replacing the PRNG by calling
2212 RAND_set_rand_method would be impossible.
2215 *) Don't let DSA_generate_key() enter an infinite loop if the random
2216 number generation fails.
2219 *) New 'rand' application for creating pseudo-random output.
2222 *) Added configuration support for Linux/IA64
2223 [Rolf Haberrecker <rolf@suse.de>]
2225 *) Assembler module support for Mingw32.
2228 *) Shared library support for HPUX (in shlib/).
2229 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2231 *) Shared library support for Solaris gcc.
2232 [Lutz Behnke <behnke@trustcenter.de>]
2234 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2236 *) PKCS7_encrypt() was adding text MIME headers twice because they
2237 were added manually and by SMIME_crlf_copy().
2240 *) In bntest.c don't call BN_rand with zero bits argument.
2241 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2243 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2244 case was implemented. This caused BN_div_recp() to fail occasionally.
2247 *) Add an optional second argument to the set_label() in the perl
2248 assembly language builder. If this argument exists and is set
2249 to 1 it signals that the assembler should use a symbol whose
2250 scope is the entire file, not just the current function. This
2251 is needed with MASM which uses the format label:: for this scope.
2252 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2254 *) Change the ASN1 types so they are typedefs by default. Before
2255 almost all types were #define'd to ASN1_STRING which was causing
2256 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2260 *) Change names of new functions to the new get1/get0 naming
2261 convention: After 'get1', the caller owns a reference count
2262 and has to call ..._free; 'get0' returns a pointer to some
2263 data structure without incrementing reference counters.
2264 (Some of the existing 'get' functions increment a reference
2265 counter, some don't.)
2266 Similarly, 'set1' and 'add1' functions increase reference
2267 counters or duplicate objects.
2270 *) Allow for the possibility of temp RSA key generation failure:
2271 the code used to assume it always worked and crashed on failure.
2274 *) Fix potential buffer overrun problem in BIO_printf().
2275 [Ulf Möller, using public domain code by Patrick Powell; problem
2276 pointed out by David Sacerdote <das33@cornell.edu>]
2278 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2279 RAND_egd() and RAND_status(). In the command line application,
2280 the EGD socket can be specified like a seed file using RANDFILE
2284 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2285 Some CAs (e.g. Verisign) distribute certificates in this form.
2288 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2289 list to exclude them. This means that no special compilation option
2290 is needed to use anonymous DH: it just needs to be included in the
2294 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2295 EVP_MD_type. The old functionality is available in a new macro called
2296 EVP_MD_md(). Change code that uses it and update docs.
2299 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2300 where the 'void *' argument is replaced by a function pointer argument.
2301 Previously 'void *' was abused to point to functions, which works on
2302 many platforms, but is not correct. As these functions are usually
2303 called by macros defined in OpenSSL header files, most source code
2304 should work without changes.
2307 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2308 sections with information on -D... compiler switches used for
2309 compiling the library so that applications can see them. To enable
2310 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2311 must be defined. E.g.,
2312 #define OPENSSL_ALGORITHM_DEFINES
2313 #include <openssl/opensslconf.h>
2314 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2315 [Richard Levitte, Ulf and Bodo Möller]
2317 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2321 *) Change the 'other' type in certificate aux info to a STACK_OF
2322 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2323 the required ASN1 format: arbitrary types determined by an OID.
2326 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2327 argument to 'req'. This is not because the function is newer or
2328 better than others it just uses the work 'NEW' in the certificate
2329 request header lines. Some software needs this.
2332 *) Reorganise password command line arguments: now passwords can be
2333 obtained from various sources. Delete the PEM_cb function and make
2334 it the default behaviour: i.e. if the callback is NULL and the
2335 usrdata argument is not NULL interpret it as a null terminated pass
2336 phrase. If usrdata and the callback are NULL then the pass phrase
2337 is prompted for as usual.
2340 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2341 the support is automatically enabled. The resulting binaries will
2342 autodetect the card and use it if present.
2343 [Ben Laurie and Compaq Inc.]
2345 *) Work around for Netscape hang bug. This sends certificate request
2346 and server done in one record. Since this is perfectly legal in the
2347 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2348 the bugs/SSLv3 entry for more info.
2351 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2354 *) Add -rand argument to smime and pkcs12 applications and read/write
2358 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2361 *) Add command line password options to the remaining applications.
2364 *) Bug fix for BN_div_recp() for numerators with an even number of
2368 *) More tests in bntest.c, and changed test_bn output.
2371 *) ./config recognizes MacOS X now.
2374 *) Bug fix for BN_div() when the first words of num and divsor are
2375 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2378 *) Add support for various broken PKCS#8 formats, and command line
2379 options to produce them.
2382 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2383 get temporary BIGNUMs from a BN_CTX.
2386 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2390 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2391 include a #define from the old name to the new. The original intent
2392 was that statically linked binaries could for example just call
2393 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2394 link with digests. This never worked becayse SSLeay_add_all_digests()
2395 and SSLeay_add_all_ciphers() were in the same source file so calling
2396 one would link with the other. They are now in separate source files.
2399 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2402 *) Use a less unusual form of the Miller-Rabin primality test (it used
2403 a binary algorithm for exponentiation integrated into the Miller-Rabin
2404 loop, our standard modexp algorithms are faster).
2407 *) Support for the EBCDIC character set completed.
2408 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2410 *) Source code cleanups: use const where appropriate, eliminate casts,
2411 use void * instead of char * in lhash.
2414 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2415 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2416 this the server could overwrite ephemeral keys that the client
2420 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2421 using 50 iterations of the Rabin-Miller test.
2423 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2424 iterations of the Rabin-Miller test as required by the appendix
2425 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2426 As BN_is_prime_fasttest includes trial division, DSA parameter
2427 generation becomes much faster.
2429 This implies a change for the callback functions in DSA_is_prime
2430 and DSA_generate_parameters: The callback function is called once
2431 for each positive witness in the Rabin-Miller test, not just
2432 occasionally in the inner loop; and the parameters to the
2433 callback function now provide an iteration count for the outer
2434 loop rather than for the current invocation of the inner loop.
2435 DSA_generate_parameters additionally can call the callback
2436 function with an 'iteration count' of -1, meaning that a
2437 candidate has passed the trial division test (when q is generated
2438 from an application-provided seed, trial division is skipped).
2441 *) New function BN_is_prime_fasttest that optionally does trial
2442 division before starting the Rabin-Miller test and has
2443 an additional BN_CTX * argument (whereas BN_is_prime always
2444 has to allocate at least one BN_CTX).
2445 'callback(1, -1, cb_arg)' is called when a number has passed the
2446 trial division stage.
2449 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2453 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2456 *) New function BN_pseudo_rand().
2459 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2460 bignum version of BN_from_montgomery() with the working code from
2461 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2465 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2466 made it impossible to use the same SSL_SESSION data structure in
2467 SSL2 clients in multiple threads.
2470 *) The return value of RAND_load_file() no longer counts bytes obtained
2471 by stat(). RAND_load_file(..., -1) is new and uses the complete file
2472 to seed the PRNG (previously an explicit byte count was required).
2473 [Ulf Möller, Bodo Möller]
2475 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
2476 used (char *) instead of (void *) and had casts all over the place.
2479 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
2482 *) Retain source code compatibility for BN_prime_checks macro:
2483 BN_is_prime(..., BN_prime_checks, ...) now uses
2484 BN_prime_checks_for_size to determine the appropriate number of
2485 Rabin-Miller iterations.
2488 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
2489 DH_CHECK_P_NOT_SAFE_PRIME.
2490 (Check if this is true? OpenPGP calls them "strong".)
2493 *) Merge the functionality of "dh" and "gendh" programs into a new program
2494 "dhparam". The old programs are retained for now but will handle DH keys
2495 (instead of parameters) in future.
2498 *) Make the ciphers, s_server and s_client programs check the return values
2499 when a new cipher list is set.
2502 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
2503 ciphers. Before when the 56bit ciphers were enabled the sorting was
2506 The syntax for the cipher sorting has been extended to support sorting by
2507 cipher-strength (using the strength_bits hard coded in the tables).
2508 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2510 Fix a bug in the cipher-command parser: when supplying a cipher command
2511 string with an "undefined" symbol (neither command nor alphanumeric
2512 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2513 an error is flagged.
2515 Due to the strength-sorting extension, the code of the
2516 ssl_create_cipher_list() function was completely rearranged. I hope that
2517 the readability was also increased :-)
2518 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2520 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2521 for the first serial number and places 2 in the serial number file. This
2522 avoids problems when the root CA is created with serial number zero and
2523 the first user certificate has the same issuer name and serial number
2527 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2528 the new code. Add documentation for this stuff.
2531 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2532 X509_*() to X509at_*() on the grounds that they don't handle X509
2533 structures and behave in an analagous way to the X509v3 functions:
2534 they shouldn't be called directly but wrapper functions should be used
2537 So we also now have some wrapper functions that call the X509at functions
2538 when passed certificate requests. (TO DO: similar things can be done with
2539 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2540 things. Some of these need some d2i or i2d and print functionality
2541 because they handle more complex structures.)
2544 *) Add missing #ifndefs that caused missing symbols when building libssl
2545 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2546 NO_RSA in ssl/s2*.c.
2547 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2549 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2550 has a return value which indicates the quality of the random data
2551 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2552 error queue. New function RAND_pseudo_bytes() generates output that is
2553 guaranteed to be unique but not unpredictable. RAND_add is like
2554 RAND_seed, but takes an extra argument for an entropy estimate
2555 (RAND_seed always assumes full entropy).
2558 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2559 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2560 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2561 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2562 false-positive rate of at most 2^-80 for random input.
2565 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2568 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2569 in the 0.9.5 release), this returns the chain
2570 from an X509_CTX structure with a dup of the stack and all
2571 the X509 reference counts upped: so the stack will exist
2572 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2575 Also make SSL_SESSION_print() print out the verify return
2579 *) Add manpage for the pkcs12 command. Also change the default
2580 behaviour so MAC iteration counts are used unless the new
2581 -nomaciter option is used. This improves file security and
2582 only older versions of MSIE (4.0 for example) need it.
2585 *) Honor the no-xxx Configure options when creating .DEF files.
2588 *) Add PKCS#10 attributes to field table: challengePassword,
2589 unstructuredName and unstructuredAddress. These are taken from
2590 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2591 international characters are used.
2593 More changes to X509_ATTRIBUTE code: allow the setting of types
2594 based on strings. Remove the 'loc' parameter when adding
2595 attributes because these will be a SET OF encoding which is sorted
2599 *) Initial changes to the 'req' utility to allow request generation
2600 automation. This will allow an application to just generate a template
2601 file containing all the field values and have req construct the
2604 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2605 used all over the place including certificate requests and PKCS#7
2606 structures. They are currently handled manually where necessary with
2607 some primitive wrappers for PKCS#7. The new functions behave in a
2608 manner analogous to the X509 extension functions: they allow
2609 attributes to be looked up by NID and added.
2611 Later something similar to the X509V3 code would be desirable to
2612 automatically handle the encoding, decoding and printing of the
2613 more complex types. The string types like challengePassword can
2614 be handled by the string table functions.
2616 Also modified the multi byte string table handling. Now there is
2617 a 'global mask' which masks out certain types. The table itself
2618 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2619 is useful when for example there is only one permissible type
2620 (as in countryName) and using the mask might result in no valid
2624 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2625 SSL_get_peer_finished to allow applications to obtain the latest
2626 Finished messages sent to the peer or expected from the peer,
2627 respectively. (SSL_get_peer_finished is usually the Finished message
2628 actually received from the peer, otherwise the protocol will be aborted.)
2630 As the Finished message are message digests of the complete handshake
2631 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2632 be used for external authentication procedures when the authentication
2633 provided by SSL/TLS is not desired or is not enough.
2636 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2637 the host supports BWX extension and if Compaq C is present on the
2638 $PATH. Just exploiting of the BWX extension results in 20-30%
2639 performance kick for some algorithms, e.g. DES and RC4 to mention
2640 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2644 *) Add support for MS "fast SGC". This is arguably a violation of the
2645 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2646 weak crypto and after checking the certificate is SGC a second one
2647 with strong crypto. MS SGC stops the first handshake after receiving
2648 the server certificate message and sends a second client hello. Since
2649 a server will typically do all the time consuming operations before
2650 expecting any further messages from the client (server key exchange
2651 is the most expensive) there is little difference between the two.
2653 To get OpenSSL to support MS SGC we have to permit a second client
2654 hello message after we have sent server done. In addition we have to
2655 reset the MAC if we do get this second client hello.
2658 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2659 if a DER encoded private key is RSA or DSA traditional format. Changed
2660 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2661 format DER encoded private key. Newer code should use PKCS#8 format which
2662 has the key type encoded in the ASN1 structure. Added DER private key
2663 support to pkcs8 application.
2666 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2667 ciphersuites has been selected (as required by the SSL 3/TLS 1
2668 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2669 is set, we interpret this as a request to violate the specification
2670 (the worst that can happen is a handshake failure, and 'correct'
2671 behaviour would result in a handshake failure anyway).
2674 *) In SSL_CTX_add_session, take into account that there might be multiple
2675 SSL_SESSION structures with the same session ID (e.g. when two threads
2676 concurrently obtain them from an external cache).
2677 The internal cache can handle only one SSL_SESSION with a given ID,
2678 so if there's a conflict, we now throw out the old one to achieve
2682 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2683 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2684 some routines that use cipher OIDs: some ciphers do not have OIDs
2685 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2689 *) Simplify the trust setting structure and code. Now we just have
2690 two sequences of OIDs for trusted and rejected settings. These will
2691 typically have values the same as the extended key usage extension
2692 and any application specific purposes.
2694 The trust checking code now has a default behaviour: it will just
2695 check for an object with the same NID as the passed id. Functions can
2696 be provided to override either the default behaviour or the behaviour
2697 for a given id. SSL client, server and email already have functions
2698 in place for compatibility: they check the NID and also return "trusted"
2699 if the certificate is self signed.
2702 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2703 traditional format into an EVP_PKEY structure.
2706 *) Add a password callback function PEM_cb() which either prompts for
2707 a password if usr_data is NULL or otherwise assumes it is a null
2708 terminated password. Allow passwords to be passed on command line
2709 environment or config files in a few more utilities.
2712 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2713 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2714 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2715 Update documentation.
2718 *) Support for ASN1 "NULL" type. This could be handled before by using
2719 ASN1_TYPE but there wasn't any function that would try to read a NULL
2720 and produce an error if it couldn't. For compatibility we also have
2721 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2722 don't allocate anything because they don't need to.
2725 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2727 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2729 *) Rebuild of the memory allocation routines used by OpenSSL code and
2730 possibly others as well. The purpose is to make an interface that
2731 provide hooks so anyone can build a separate set of allocation and
2732 deallocation routines to be used by OpenSSL, for example memory
2733 pool implementations, or something else, which was previously hard
2734 since Malloc(), Realloc() and Free() were defined as macros having
2735 the values malloc, realloc and free, respectively (except for Win32
2736 compilations). The same is provided for memory debugging code.
2737 OpenSSL already comes with functionality to find memory leaks, but
2738 this gives people a chance to debug other memory problems.
2740 With these changes, a new set of functions and macros have appeared:
2742 CRYPTO_set_mem_debug_functions() [F]
2743 CRYPTO_get_mem_debug_functions() [F]
2744 CRYPTO_dbg_set_options() [F]
2745 CRYPTO_dbg_get_options() [F]
2746 CRYPTO_malloc_debug_init() [M]
2748 The memory debug functions are NULL by default, unless the library
2749 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2750 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2751 gives the standard debugging functions that come with OpenSSL) or
2752 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2753 provided by the library user) must be used. When the standard
2754 debugging functions are used, CRYPTO_dbg_set_options can be used to
2755 request additional information:
2756 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2757 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2759 Also, things like CRYPTO_set_mem_functions will always give the
2760 expected result (the new set of functions is used for allocation
2761 and deallocation) at all times, regardless of platform and compiler
2764 To finish it up, some functions that were never use in any other
2765 way than through macros have a new API and new semantic:
2768 CRYPTO_dbg_realloc()
2771 All macros of value have retained their old syntax.
2772 [Richard Levitte and Bodo Moeller]
2774 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2775 ordering of SMIMECapabilities wasn't in "strength order" and there
2776 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2780 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2781 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2782 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2784 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2785 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2786 functionality to handle multipart/signed properly) and a utility
2787 called 'smime' to call all this stuff. This is based on code I
2788 originally wrote for Celo who have kindly allowed it to be
2789 included in OpenSSL.
2792 *) Add variants des_set_key_checked and des_set_key_unchecked of
2793 des_set_key (aka des_key_sched). Global variable des_check_key
2794 decides which of these is called by des_set_key; this way
2795 des_check_key behaves as it always did, but applications and
2796 the library itself, which was buggy for des_check_key == 1,
2797 have a cleaner way to pick the version they need.
2800 *) New function PKCS12_newpass() which changes the password of a
2804 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2805 dynamic mix. In both cases the ids can be used as an index into the
2806 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2807 functions so they accept a list of the field values and the
2808 application doesn't need to directly manipulate the X509_TRUST
2812 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2816 *) Modify the way the V3 extension code looks up extensions. This now
2817 works in a similar way to the object code: we have some "standard"
2818 extensions in a static table which is searched with OBJ_bsearch()
2819 and the application can add dynamic ones if needed. The file
2820 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2821 updated whenever a new extension is added to the core code and kept
2822 in ext_nid order. There is a simple program 'tabtest.c' which checks
2823 this. New extensions are not added too often so this file can readily
2824 be maintained manually.
2826 There are two big advantages in doing things this way. The extensions
2827 can be looked up immediately and no longer need to be "added" using
2828 X509V3_add_standard_extensions(): this function now does nothing.
2829 [Side note: I get *lots* of email saying the extension code doesn't
2830 work because people forget to call this function]
2831 Also no dynamic allocation is done unless new extensions are added:
2832 so if we don't add custom extensions there is no need to call
2833 X509V3_EXT_cleanup().
2836 *) Modify enc utility's salting as follows: make salting the default. Add a
2837 magic header, so unsalted files fail gracefully instead of just decrypting
2838 to garbage. This is because not salting is a big security hole, so people
2839 should be discouraged from doing it.
2842 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2843 digest to be passed on the command line but it only used this
2844 parameter when signing a certificate. Modified so all relevant
2845 operations are affected by the digest parameter including the
2846 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2847 DSA key was used because it didn't fix the digest.
2850 *) Initial certificate chain verify code. Currently tests the untrusted
2851 certificates for consistency with the verify purpose (which is set
2852 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2854 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2855 this is because it will reject chains with invalid extensions whereas
2856 every previous version of OpenSSL and SSLeay made no checks at all.
2858 Trust code: checks the root CA for the relevant trust settings. Trust
2859 settings have an initial value consistent with the verify purpose: e.g.
2860 if the verify purpose is for SSL client use it expects the CA to be
2861 trusted for SSL client use. However the default value can be changed to
2862 permit custom trust settings: one example of this would be to only trust
2863 certificates from a specific "secure" set of CAs.
2865 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2866 which should be used for version portability: especially since the
2867 verify structure is likely to change more often now.
2869 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2870 to set them. If not set then assume SSL clients will verify SSL servers
2873 Two new options to the verify program: -untrusted allows a set of
2874 untrusted certificates to be passed in and -purpose which sets the
2875 intended purpose of the certificate. If a purpose is set then the
2876 new chain verify code is used to check extension consistency.
2879 *) Support for the authority information access extension.
2882 *) Modify RSA and DSA PEM read routines to transparently handle
2883 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2884 public keys in a format compatible with certificate
2885 SubjectPublicKeyInfo structures. Unfortunately there were already
2886 functions called *_PublicKey_* which used various odd formats so
2887 these are retained for compatibility: however the DSA variants were
2888 never in a public release so they have been deleted. Changed dsa/rsa
2889 utilities to handle the new format: note no releases ever handled public
2890 keys so we should be OK.
2892 The primary motivation for this change is to avoid the same fiasco
2893 that dogs private keys: there are several incompatible private key
2894 formats some of which are standard and some OpenSSL specific and
2895 require various evil hacks to allow partial transparent handling and
2896 even then it doesn't work with DER formats. Given the option anything
2897 other than PKCS#8 should be dumped: but the other formats have to
2898 stay in the name of compatibility.
2900 With public keys and the benefit of hindsight one standard format
2901 is used which works with EVP_PKEY, RSA or DSA structures: though
2902 it clearly returns an error if you try to read the wrong kind of key.
2904 Added a -pubkey option to the 'x509' utility to output the public key.
2905 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2906 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2907 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2908 that do the same as the EVP_PKEY_assign_*() except they up the
2909 reference count of the added key (they don't "swallow" the
2913 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2914 CRLs would fail if the file contained no certificates or no CRLs:
2915 added a new function to read in both types and return the number
2916 read: this means that if none are read it will be an error. The
2917 DER versions of the certificate and CRL reader would always fail
2918 because it isn't possible to mix certificates and CRLs in DER format
2919 without choking one or the other routine. Changed this to just read
2920 a certificate: this is the best we can do. Also modified the code
2921 in apps/verify.c to take notice of return codes: it was previously
2922 attempting to read in certificates from NULL pointers and ignoring
2923 any errors: this is one reason why the cert and CRL reader seemed
2924 to work. It doesn't check return codes from the default certificate
2925 routines: these may well fail if the certificates aren't installed.
2928 *) Code to support otherName option in GeneralName.
2931 *) First update to verify code. Change the verify utility
2932 so it warns if it is passed a self signed certificate:
2933 for consistency with the normal behaviour. X509_verify
2934 has been modified to it will now verify a self signed
2935 certificate if *exactly* the same certificate appears
2936 in the store: it was previously impossible to trust a
2937 single self signed certificate. This means that:
2938 openssl verify ss.pem
2939 now gives a warning about a self signed certificate but
2940 openssl verify -CAfile ss.pem ss.pem
2944 *) For servers, store verify_result in SSL_SESSION data structure
2945 (and add it to external session representation).
2946 This is needed when client certificate verifications fails,
2947 but an application-provided verification callback (set by
2948 SSL_CTX_set_cert_verify_callback) allows accepting the session
2949 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2950 but returns 1): When the session is reused, we have to set
2951 ssl->verify_result to the appropriate error code to avoid
2953 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2955 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2956 case in PKCS7_dataInit() where the signed PKCS7 structure
2957 didn't contain any existing data because it was being created.
2958 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2960 *) Add a salt to the key derivation routines in enc.c. This
2961 forms the first 8 bytes of the encrypted file. Also add a
2962 -S option to allow a salt to be input on the command line.
2965 *) New function X509_cmp(). Oddly enough there wasn't a function
2966 to compare two certificates. We do this by working out the SHA1
2967 hash and comparing that. X509_cmp() will be needed by the trust
2971 *) SSL_get1_session() is like SSL_get_session(), but increments
2972 the reference count in the SSL_SESSION returned.
2973 [Geoff Thorpe <geoff@eu.c2.net>]
2975 *) Fix for 'req': it was adding a null to request attributes.
2976 Also change the X509_LOOKUP and X509_INFO code to handle
2977 certificate auxiliary information.
2980 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2984 *) Add the possibility to add extra information to the memory leak
2985 detecting output, to form tracebacks, showing from where each
2986 allocation was originated: CRYPTO_push_info("constant string") adds
2987 the string plus current file name and line number to a per-thread
2988 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2989 is like calling CYRPTO_pop_info() until the stack is empty.
2990 Also updated memory leak detection code to be multi-thread-safe.
2993 *) Add options -text and -noout to pkcs7 utility and delete the
2994 encryption options which never did anything. Update docs.
2997 *) Add options to some of the utilities to allow the pass phrase
2998 to be included on either the command line (not recommended on
2999 OSes like Unix) or read from the environment. Update the
3000 manpages and fix a few bugs.
3003 *) Add a few manpages for some of the openssl commands.
3006 *) Fix the -revoke option in ca. It was freeing up memory twice,
3007 leaking and not finding already revoked certificates.
3010 *) Extensive changes to support certificate auxiliary information.
3011 This involves the use of X509_CERT_AUX structure and X509_AUX
3012 functions. An X509_AUX function such as PEM_read_X509_AUX()
3013 can still read in a certificate file in the usual way but it
3014 will also read in any additional "auxiliary information". By
3015 doing things this way a fair degree of compatibility can be
3016 retained: existing certificates can have this information added
3017 using the new 'x509' options.
3019 Current auxiliary information includes an "alias" and some trust
3020 settings. The trust settings will ultimately be used in enhanced
3021 certificate chain verification routines: currently a certificate
3022 can only be trusted if it is self signed and then it is trusted
3026 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3027 The problem was that one of the replacement routines had not been working
3028 since SSLeay releases. For now the offending routine has been replaced
3029 with non-optimised assembler. Even so, this now gives around 95%
3030 performance improvement for 1024 bit RSA signs.
3033 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3034 handling. Most clients have the effective key size in bits equal to
3035 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3036 A few however don't do this and instead use the size of the decrypted key
3037 to determine the RC2 key length and the AlgorithmIdentifier to determine
3038 the effective key length. In this case the effective key length can still
3039 be 40 bits but the key length can be 168 bits for example. This is fixed
3040 by manually forcing an RC2 key into the EVP_PKEY structure because the
3041 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3042 the key length and effective key length are equal.
3045 *) Add a bunch of functions that should simplify the creation of
3046 X509_NAME structures. Now you should be able to do:
3047 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3048 and have it automatically work out the correct field type and fill in
3049 the structures. The more adventurous can try:
3050 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3051 and it will (hopefully) work out the correct multibyte encoding.
3054 *) Change the 'req' utility to use the new field handling and multibyte
3055 copy routines. Before the DN field creation was handled in an ad hoc
3056 way in req, ca, and x509 which was rather broken and didn't support
3057 BMPStrings or UTF8Strings. Since some software doesn't implement
3058 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3059 using the dirstring_type option. See the new comment in the default
3060 openssl.cnf for more info.
3063 *) Make crypto/rand/md_rand.c more robust:
3064 - Assure unique random numbers after fork().
3065 - Make sure that concurrent threads access the global counter and
3066 md serializably so that we never lose entropy in them
3067 or use exactly the same state in multiple threads.
3068 Access to the large state is not always serializable because
3069 the additional locking could be a performance killer, and
3070 md should be large enough anyway.
3073 *) New file apps/app_rand.c with commonly needed functionality
3074 for handling the random seed file.
3076 Use the random seed file in some applications that previously did not:
3078 dsaparam -genkey (which also ignored its '-rand' option),
3081 x509 (when signing).
3082 Except on systems with /dev/urandom, it is crucial to have a random
3083 seed file at least for key creation, DSA signing, and for DH exchanges;
3084 for RSA signatures we could do without one.
3086 gendh and gendsa (unlike genrsa) used to read only the first byte
3087 of each file listed in the '-rand' option. The function as previously
3088 found in genrsa is now in app_rand.c and is used by all programs
3089 that support '-rand'.
3092 *) In RAND_write_file, use mode 0600 for creating files;
3093 don't just chmod when it may be too late.
3096 *) Report an error from X509_STORE_load_locations
3097 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3100 *) New function ASN1_mbstring_copy() this copies a string in either
3101 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3102 into an ASN1_STRING type. A mask of permissible types is passed
3103 and it chooses the "minimal" type to use or an error if not type
3107 *) Add function equivalents to the various macros in asn1.h. The old
3108 macros are retained with an M_ prefix. Code inside the library can
3109 use the M_ macros. External code (including the openssl utility)
3110 should *NOT* in order to be "shared library friendly".
3113 *) Add various functions that can check a certificate's extensions
3114 to see if it usable for various purposes such as SSL client,
3115 server or S/MIME and CAs of these types. This is currently
3116 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3117 verification. Also added a -purpose flag to x509 utility to
3118 print out all the purposes.
3121 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3125 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3126 for, obtain and decode and extension and obtain its critical flag.
3127 This allows all the necessary extension code to be handled in a
3128 single function call.
3131 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3132 platforms. See crypto/rc4/rc4_enc.c for further details.
3135 *) New -noout option to asn1parse. This causes no output to be produced
3136 its main use is when combined with -strparse and -out to extract data
3137 from a file (which may not be in ASN.1 format).
3140 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3141 when producing the local key id.
3142 [Richard Levitte <levitte@stacken.kth.se>]
3144 *) New option -dhparam in s_server. This allows a DH parameter file to be
3145 stated explicitly. If it is not stated then it tries the first server
3146 certificate file. The previous behaviour hard coded the filename
3150 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3151 a public key to be input or output. For example:
3152 openssl rsa -in key.pem -pubout -out pubkey.pem
3153 Also added necessary DSA public key functions to handle this.
3156 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3157 in the message. This was handled by allowing
3158 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3159 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3161 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3162 to the end of the strings whereas this didn't. This would cause problems
3163 if strings read with d2i_ASN1_bytes() were later modified.
3164 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3166 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3167 data and it contains EOF it will end up returning an error. This is
3168 caused by input 46 bytes long. The cause is due to the way base64
3169 BIOs find the start of base64 encoded data. They do this by trying a
3170 trial decode on each line until they find one that works. When they
3171 do a flag is set and it starts again knowing it can pass all the
3172 data directly through the decoder. Unfortunately it doesn't reset
3173 the context it uses. This means that if EOF is reached an attempt
3174 is made to pass two EOFs through the context and this causes the
3175 resulting error. This can also cause other problems as well. As is
3176 usual with these problems it takes *ages* to find and the fix is
3177 trivial: move one line.
3178 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3180 *) Ugly workaround to get s_client and s_server working under Windows. The
3181 old code wouldn't work because it needed to select() on sockets and the
3182 tty (for keypresses and to see if data could be written). Win32 only
3183 supports select() on sockets so we select() with a 1s timeout on the
3184 sockets and then see if any characters are waiting to be read, if none
3185 are present then we retry, we also assume we can always write data to
3186 the tty. This isn't nice because the code then blocks until we've
3187 received a complete line of data and it is effectively polling the
3188 keyboard at 1s intervals: however it's quite a bit better than not
3189 working at all :-) A dedicated Windows application might handle this
3190 with an event loop for example.
3193 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3194 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3195 will be called when RSA_sign() and RSA_verify() are used. This is useful
3196 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3197 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3198 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3199 This necessitated the support of an extra signature type NID_md5_sha1
3200 for SSL signatures and modifications to the SSL library to use it instead
3201 of calling RSA_public_decrypt() and RSA_private_encrypt().
3204 *) Add new -verify -CAfile and -CApath options to the crl program, these
3205 will lookup a CRL issuers certificate and verify the signature in a
3206 similar way to the verify program. Tidy up the crl program so it
3207 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3208 less strict. It will now permit CRL extensions even if it is not
3209 a V2 CRL: this will allow it to tolerate some broken CRLs.
3212 *) Initialize all non-automatic variables each time one of the openssl
3213 sub-programs is started (this is necessary as they may be started
3214 multiple times from the "OpenSSL>" prompt).
3215 [Lennart Bang, Bodo Moeller]
3217 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3218 removing all other RSA functionality (this is what NO_RSA does). This
3219 is so (for example) those in the US can disable those operations covered
3220 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3224 *) Non-copying interface to BIO pairs.
3225 (still largely untested)
3228 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3229 ASCII string. This was handled independently in various places before.
3232 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3233 UTF8 strings a character at a time.
3236 *) Use client_version from client hello to select the protocol
3237 (s23_srvr.c) and for RSA client key exchange verification
3238 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3241 *) Add various utility functions to handle SPKACs, these were previously
3242 handled by poking round in the structure internals. Added new function
3243 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3244 print, verify and generate SPKACs. Based on an original idea from
3245 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3248 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3251 *) Allow the config file extension section to be overwritten on the
3252 command line. Based on an original idea from Massimiliano Pala
3253 <madwolf@comune.modena.it>. The new option is called -extensions
3254 and can be applied to ca, req and x509. Also -reqexts to override
3255 the request extensions in req and -crlexts to override the crl extensions
3259 *) Add new feature to the SPKAC handling in ca. Now you can include
3260 the same field multiple times by preceding it by "XXXX." for example:
3263 this is the same syntax as used in the req config file.
3266 *) Allow certificate extensions to be added to certificate requests. These
3267 are specified in a 'req_extensions' option of the req section of the
3268 config file. They can be printed out with the -text option to req but
3269 are otherwise ignored at present.
3272 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3273 data read consists of only the final block it would not decrypted because
3274 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3275 A misplaced 'break' also meant the decrypted final block might not be
3276 copied until the next read.
3279 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3280 a few extra parameters to the DH structure: these will be useful if
3281 for example we want the value of 'q' or implement X9.42 DH.
3284 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3285 provides hooks that allow the default DSA functions or functions on a
3286 "per key" basis to be replaced. This allows hardware acceleration and
3287 hardware key storage to be handled without major modification to the
3288 library. Also added low level modexp hooks and CRYPTO_EX structure and
3289 associated functions.
3292 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3293 as "read only": it can't be written to and the buffer it points to will
3294 not be freed. Reading from a read only BIO is much more efficient than
3295 a normal memory BIO. This was added because there are several times when
3296 an area of memory needs to be read from a BIO. The previous method was
3297 to create a memory BIO and write the data to it, this results in two
3298 copies of the data and an O(n^2) reading algorithm. There is a new
3299 function BIO_new_mem_buf() which creates a read only memory BIO from
3300 an area of memory. Also modified the PKCS#7 routines to use read only
3304 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3305 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3306 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3307 but a retry condition occured while trying to read the rest.
3310 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3311 NID_pkcs7_encrypted by default: this was wrong since this should almost
3312 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3313 the encrypted data type: this is a more sensible place to put it and it
3314 allows the PKCS#12 code to be tidied up that duplicated this
3318 *) Changed obj_dat.pl script so it takes its input and output files on
3319 the command line. This should avoid shell escape redirection problems
3323 *) Initial support for certificate extension requests, these are included
3324 in things like Xenroll certificate requests. Included functions to allow
3325 extensions to be obtained and added.
3328 *) -crlf option to s_client and s_server for sending newlines as
3329 CRLF (as required by many protocols).
3332 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3334 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3335 [Ralf S. Engelschall]
3337 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3338 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3340 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3344 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3345 DH parameters/keys (q is lost during that conversion, but the resulting
3346 DH parameters contain its length).
3348 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3349 much faster than DH_generate_parameters (which creates parameters
3350 where p = 2*q + 1), and also the smaller q makes DH computations
3351 much more efficient (160-bit exponentiation instead of 1024-bit
3352 exponentiation); so this provides a convenient way to support DHE
3353 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3354 utter importance to use
3355 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3357 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3358 when such DH parameters are used, because otherwise small subgroup
3359 attacks may become possible!
3362 *) Avoid memory leak in i2d_DHparams.
3365 *) Allow the -k option to be used more than once in the enc program:
3366 this allows the same encrypted message to be read by multiple recipients.
3369 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3370 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3371 it will always use the numerical form of the OID, even if it has a short
3375 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3376 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3377 otherwise bn_mod_exp was called. In the case of hardware keys for example
3378 no private key components need be present and it might store extra data
3379 in the RSA structure, which cannot be accessed from bn_mod_exp.
3380 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3381 private key operations.
3384 *) Added support for SPARC Linux.
3387 *) pem_password_cb function type incompatibly changed from
3388 typedef int pem_password_cb(char *buf, int size, int rwflag);
3390 ....(char *buf, int size, int rwflag, void *userdata);
3391 so that applications can pass data to their callbacks:
3392 The PEM[_ASN1]_{read,write}... functions and macros now take an
3393 additional void * argument, which is just handed through whenever
3394 the password callback is called.
3395 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3397 New function SSL_CTX_set_default_passwd_cb_userdata.
3399 Compatibility note: As many C implementations push function arguments
3400 onto the stack in reverse order, the new library version is likely to
3401 interoperate with programs that have been compiled with the old
3402 pem_password_cb definition (PEM_whatever takes some data that
3403 happens to be on the stack as its last argument, and the callback
3404 just ignores this garbage); but there is no guarantee whatsoever that
3407 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3408 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3409 problems not only on Windows, but also on some Unix platforms.
3410 To avoid problematic command lines, these definitions are now in an
3411 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3412 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3415 *) MIPS III/IV assembler module is reimplemented.
3418 *) More DES library cleanups: remove references to srand/rand and
3419 delete an unused file.
3422 *) Add support for the the free Netwide assembler (NASM) under Win32,
3423 since not many people have MASM (ml) and it can be hard to obtain.
3424 This is currently experimental but it seems to work OK and pass all
3425 the tests. Check out INSTALL.W32 for info.
3428 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3429 without temporary keys kept an extra copy of the server key,
3430 and connections with temporary keys did not free everything in case
3434 *) New function RSA_check_key and new openssl rsa option -check
3435 for verifying the consistency of RSA keys.
3436 [Ulf Moeller, Bodo Moeller]
3438 *) Various changes to make Win32 compile work:
3439 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3440 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3441 comparison" warnings.
3442 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3445 *) Add a debugging option to PKCS#5 v2 key generation function: when
3446 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3447 derived keys are printed to stderr.
3450 *) Copy the flags in ASN1_STRING_dup().
3451 [Roman E. Pavlov <pre@mo.msk.ru>]
3453 *) The x509 application mishandled signing requests containing DSA
3454 keys when the signing key was also DSA and the parameters didn't match.
3456 It was supposed to omit the parameters when they matched the signing key:
3457 the verifying software was then supposed to automatically use the CA's
3458 parameters if they were absent from the end user certificate.
3460 Omitting parameters is no longer recommended. The test was also
3461 the wrong way round! This was probably due to unusual behaviour in
3462 EVP_cmp_parameters() which returns 1 if the parameters match.
3463 This meant that parameters were omitted when they *didn't* match and
3464 the certificate was useless. Certificates signed with 'ca' didn't have
3466 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3468 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3469 The interface is as follows:
3470 Applications can use
3471 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
3472 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
3473 "off" is now the default.
3474 The library internally uses
3475 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
3476 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
3477 to disable memory-checking temporarily.
3479 Some inconsistent states that previously were possible (and were
3480 even the default) are now avoided.
3482 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
3483 with each memory chunk allocated; this is occasionally more helpful
3484 than just having a counter.
3486 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
3488 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
3492 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
3493 which largely parallels "options", but is for changing API behaviour,
3494 whereas "options" are about protocol behaviour.
3495 Initial "mode" flags are:
3497 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
3498 a single record has been written.
3499 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
3500 retries use the same buffer location.
3501 (But all of the contents must be
3505 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
3508 *) Fix problems with no-hmac etc.
3509 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3511 *) New functions RSA_get_default_method(), RSA_set_method() and
3512 RSA_get_method(). These allows replacement of RSA_METHODs without having
3513 to mess around with the internals of an RSA structure.
3516 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3517 Also really enable memory leak checks in openssl.c and in some
3519 [Chad C. Mulligan, Bodo Moeller]
3521 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3522 up the length of negative integers. This has now been simplified to just
3523 store the length when it is first determined and use it later, rather
3524 than trying to keep track of where data is copied and updating it to
3526 [Steve Henson, reported by Brien Wheeler
3527 <bwheeler@authentica-security.com>]
3529 *) Add a new function PKCS7_signatureVerify. This allows the verification
3530 of a PKCS#7 signature but with the signing certificate passed to the
3531 function itself. This contrasts with PKCS7_dataVerify which assumes the
3532 certificate is present in the PKCS#7 structure. This isn't always the
3533 case: certificates can be omitted from a PKCS#7 structure and be
3534 distributed by "out of band" means (such as a certificate database).
3537 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3538 function prototypes in pem.h, also change util/mkdef.pl to add the
3539 necessary function names.
3542 *) mk1mf.pl (used by Windows builds) did not properly read the
3543 options set by Configure in the top level Makefile, and Configure
3544 was not even able to write more than one option correctly.
3545 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3548 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3549 file to be loaded from a BIO or FILE pointer. The BIO version will
3550 for example allow memory BIOs to contain config info.
3553 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3554 Whoever hopes to achieve shared-library compatibility across versions
3555 must use this, not the compile-time macro.
3556 (Exercise 0.9.4: Which is the minimum library version required by
3558 Note: All this applies only to multi-threaded programs, others don't
3562 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3563 through a BIO pair triggered the default case, i.e.
3564 SSLerr(...,SSL_R_UNKNOWN_STATE).
3567 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3568 can use the SSL library even if none of the specific BIOs is
3572 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3573 for the encoded length.
3574 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3576 *) Add initial documentation of the X509V3 functions.
3579 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3580 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3581 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3582 secure PKCS#8 private key format with a high iteration count.
3585 *) Fix determination of Perl interpreter: A perl or perl5
3586 _directory_ in $PATH was also accepted as the interpreter.
3587 [Ralf S. Engelschall]
3589 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3590 wrong with it but it was very old and did things like calling
3591 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3595 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3596 to use the new extension code.
3599 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3600 with macros. This should make it easier to change their form, add extra
3601 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3605 *) Add to configuration table a new entry that can specify an alternative
3606 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3607 according to Mark Crispin <MRC@Panda.COM>.
3611 *) DES CBC did not update the IV. Weird.
3614 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3615 Changing the behaviour of the former might break existing programs --
3616 where IV updating is needed, des_ncbc_encrypt can be used.
3619 *) When bntest is run from "make test" it drives bc to check its
3620 calculations, as well as internally checking them. If an internal check
3621 fails, it needs to cause bc to give a non-zero result or make test carries
3622 on without noticing the failure. Fixed.
3625 *) DES library cleanups.
3628 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3629 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3630 ciphers. NOTE: although the key derivation function has been verified
3631 against some published test vectors it has not been extensively tested
3632 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3636 *) Instead of "mkdir -p", which is not fully portable, use new
3637 Perl script "util/mkdir-p.pl".
3640 *) Rewrite the way password based encryption (PBE) is handled. It used to
3641 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3642 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3643 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3644 the 'parameter' field of the AlgorithmIdentifier is passed to the
3645 underlying key generation function so it must do its own ASN1 parsing.
3646 This has also changed the EVP_PBE_CipherInit() function which now has a
3647 'parameter' argument instead of literal salt and iteration count values
3648 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3651 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3652 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3653 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3654 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3655 value was just used as a "magic string" and not used directly its
3656 value doesn't matter.
3659 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3663 *) "linux-sparc64" configuration (ultrapenguin).
3664 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3665 "linux-sparc" configuration.
3666 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3668 *) config now generates no-xxx options for missing ciphers.
3671 *) Support the EBCDIC character set (work in progress).
3672 File ebcdic.c not yet included because it has a different license.
3673 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3675 *) Support BS2000/OSD-POSIX.
3676 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3678 *) Make callbacks for key generation use void * instead of char *.
3681 *) Make S/MIME samples compile (not yet tested).
3684 *) Additional typesafe stacks.
3687 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3691 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3693 *) New configuration variant "sco5-gcc".
3695 *) Updated some demos.
3696 [Sean O Riordain, Wade Scholine]
3698 *) Add missing BIO_free at exit of pkcs12 application.
3701 *) Fix memory leak in conf.c.
3704 *) Updates for Win32 to assembler version of MD5.
3707 *) Set #! path to perl in apps/der_chop to where we found it
3708 instead of using a fixed path.
3711 *) SHA library changes for irix64-mips4-cc.
3714 *) Improvements for VMS support.
3718 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3720 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3721 This also avoids the problems with SC4.2 and unpatched SC5.
3722 [Andy Polyakov <appro@fy.chalmers.se>]
3724 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3725 These are required because of the typesafe stack would otherwise break
3726 existing code. If old code used a structure member which used to be STACK
3727 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3728 sk_num or sk_value it would produce an error because the num, data members
3729 are not present in STACK_OF. Now it just produces a warning. sk_set
3730 replaces the old method of assigning a value to sk_value
3731 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3732 that does this will no longer work (and should use sk_set instead) but
3733 this could be regarded as a "questionable" behaviour anyway.
3736 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3737 correctly handle encrypted S/MIME data.
3740 *) Change type of various DES function arguments from des_cblock
3741 (which means, in function argument declarations, pointer to char)
3742 to des_cblock * (meaning pointer to array with 8 char elements),
3743 which allows the compiler to do more typechecking; it was like
3744 that back in SSLeay, but with lots of ugly casts.
3746 Introduce new type const_des_cblock.
3749 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3750 problems: find RecipientInfo structure that matches recipient certificate
3751 and initialise the ASN1 structures properly based on passed cipher.
3754 *) Belatedly make the BN tests actually check the results.
3757 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3758 to and from BNs: it was completely broken. New compilation option
3759 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3760 key elements as negative integers.
3763 *) Reorganize and speed up MD5.
3764 [Andy Polyakov <appro@fy.chalmers.se>]
3767 [Richard Levitte <richard@levitte.org>]
3769 *) New option -out to asn1parse to allow the parsed structure to be
3770 output to a file. This is most useful when combined with the -strparse
3771 option to examine the output of things like OCTET STRINGS.
3774 *) Make SSL library a little more fool-proof by not requiring any longer
3775 that SSL_set_{accept,connect}_state be called before
3776 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3777 in many applications because usually everything *appeared* to work as
3778 intended anyway -- now it really works as intended).
3781 *) Move openssl.cnf out of lib/.
3784 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3785 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3786 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3787 [Ralf S. Engelschall]
3789 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3790 handle PKCS#7 enveloped data properly.
3791 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3793 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3794 copying pointers. The cert_st handling is changed by this in
3795 various ways (and thus what used to be known as ctx->default_cert
3796 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3797 any longer when s->cert does not give us what we need).
3798 ssl_cert_instantiate becomes obsolete by this change.
3799 As soon as we've got the new code right (possibly it already is?),
3800 we have solved a couple of bugs of the earlier code where s->cert
3801 was used as if it could not have been shared with other SSL structures.
3803 Note that using the SSL API in certain dirty ways now will result
3804 in different behaviour than observed with earlier library versions:
3805 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3806 does not influence s as it used to.
3808 In order to clean up things more thoroughly, inside SSL_SESSION
3809 we don't use CERT any longer, but a new structure SESS_CERT
3810 that holds per-session data (if available); currently, this is
3811 the peer's certificate chain and, for clients, the server's certificate
3812 and temporary key. CERT holds only those values that can have
3813 meaningful defaults in an SSL_CTX.
3816 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3817 from the internal representation. Various PKCS#7 fixes: remove some
3818 evil casts and set the enc_dig_alg field properly based on the signing
3822 *) Allow PKCS#12 password to be set from the command line or the
3823 environment. Let 'ca' get its config file name from the environment
3824 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3828 *) Allow certificate policies extension to use an IA5STRING for the
3829 organization field. This is contrary to the PKIX definition but
3830 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3834 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3835 without disallowing inline assembler and the like for non-pedantic builds.
3838 *) Support Borland C++ builder.
3839 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3844 *) SHA-1 cleanups and performance enhancements.
3845 [Andy Polyakov <appro@fy.chalmers.se>]
3847 *) Sparc v8plus assembler for the bignum library.
3848 [Andy Polyakov <appro@fy.chalmers.se>]
3850 *) Accept any -xxx and +xxx compiler options in Configure.
3853 *) Update HPUX configuration.
3856 *) Add missing sk_<type>_unshift() function to safestack.h
3857 [Ralf S. Engelschall]
3859 *) New function SSL_CTX_use_certificate_chain_file that sets the
3860 "extra_cert"s in addition to the certificate. (This makes sense
3861 only for "PEM" format files, as chains as a whole are not
3865 *) Support verify_depth from the SSL API.
3866 x509_vfy.c had what can be considered an off-by-one-error:
3867 Its depth (which was not part of the external interface)
3868 was actually counting the number of certificates in a chain;
3869 now it really counts the depth.
3872 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3873 instead of X509err, which often resulted in confusing error
3874 messages since the error codes are not globally unique
3875 (e.g. an alleged error in ssl3_accept when a certificate
3876 didn't match the private key).
3878 *) New function SSL_CTX_set_session_id_context that allows to set a default
3879 value (so that you don't need SSL_set_session_id_context for each
3880 connection using the SSL_CTX).
3883 *) OAEP decoding bug fix.
3886 *) Support INSTALL_PREFIX for package builders, as proposed by
3890 *) New Configure options "threads" and "no-threads". For systems
3891 where the proper compiler options are known (currently Solaris
3892 and Linux), "threads" is the default.
3895 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3898 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3899 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3900 such as /usr/local/bin.
3903 *) "make linux-shared" to build shared libraries.
3904 [Niels Poppe <niels@netbox.org>]
3906 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3909 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3910 extension adding in x509 utility.
3913 *) Remove NOPROTO sections and error code comments.
3916 *) Partial rewrite of the DEF file generator to now parse the ANSI
3920 *) New Configure options --prefix=DIR and --openssldir=DIR.
3923 *) Complete rewrite of the error code script(s). It is all now handled
3924 by one script at the top level which handles error code gathering,
3925 header rewriting and C source file generation. It should be much better
3926 than the old method: it now uses a modified version of Ulf's parser to
3927 read the ANSI prototypes in all header files (thus the old K&R definitions
3928 aren't needed for error creation any more) and do a better job of
3929 translating function codes into names. The old 'ASN1 error code imbedded
3930 in a comment' is no longer necessary and it doesn't use .err files which
3931 have now been deleted. Also the error code call doesn't have to appear all
3932 on one line (which resulted in some large lines...).
3935 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3938 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3939 0 (which usually indicates a closed connection), but continue reading.
3942 *) Fix some race conditions.
3945 *) Add support for CRL distribution points extension. Add Certificate
3946 Policies and CRL distribution points documentation.
3949 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3952 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3953 8 of keying material. Merlin has also confirmed interop with this fix
3954 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3955 [Merlin Hughes <merlin@baltimore.ie>]
3957 *) Fix lots of warnings.
3958 [Richard Levitte <levitte@stacken.kth.se>]
3960 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3961 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3962 [Richard Levitte <levitte@stacken.kth.se>]
3964 *) Fix problems with sizeof(long) == 8.
3965 [Andy Polyakov <appro@fy.chalmers.se>]
3967 *) Change functions to ANSI C.
3970 *) Fix typos in error codes.
3971 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3973 *) Remove defunct assembler files from Configure.
3976 *) SPARC v8 assembler BIGNUM implementation.
3977 [Andy Polyakov <appro@fy.chalmers.se>]
3979 *) Support for Certificate Policies extension: both print and set.
3980 Various additions to support the r2i method this uses.
3983 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3984 return a const string when you are expecting an allocated buffer.
3987 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3988 types DirectoryString and DisplayText.
3991 *) Add code to allow r2i extensions to access the configuration database,
3992 add an LHASH database driver and add several ctx helper functions.
3995 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3996 fail when they extended the size of a BIGNUM.
3999 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4000 support typesafe stack.
4003 *) Fix typo in SSL_[gs]et_options().
4004 [Nils Frostberg <nils@medcom.se>]
4006 *) Delete various functions and files that belonged to the (now obsolete)
4007 old X509V3 handling code.
4010 *) New Configure option "rsaref".
4013 *) Don't auto-generate pem.h.
4016 *) Introduce type-safe ASN.1 SETs.
4019 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4020 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4022 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4023 that links with OpenSSL (well at least cause lots of warnings), but fear
4024 not: the conversion is trivial, and it eliminates loads of evil casts. A
4025 few STACKed things have been converted already. Feel free to convert more.
4026 In the fullness of time, I'll do away with the STACK type altogether.
4029 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4030 specified in <certfile> by updating the entry in the index.txt file.
4031 This way one no longer has to edit the index.txt file manually for
4032 revoking a certificate. The -revoke option does the gory details now.
4033 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4035 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4036 `-text' option at all and this way the `-noout -text' combination was
4037 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4038 [Ralf S. Engelschall]
4040 *) Make sure a corresponding plain text error message exists for the
4041 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4042 verify callback function determined that a certificate was revoked.
4043 [Ralf S. Engelschall]
4045 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4046 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4047 all available cipers including rc5, which was forgotten until now.
4048 In order to let the testing shell script know which algorithms
4049 are available, a new (up to now undocumented) command
4050 "openssl list-cipher-commands" is used.
4053 *) Bugfix: s_client occasionally would sleep in select() when
4054 it should have checked SSL_pending() first.
4057 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4058 the raw DSA values prior to ASN.1 encoding.
4061 *) Tweaks to Configure
4062 [Niels Poppe <niels@netbox.org>]
4064 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4068 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4071 *) New config option to avoid instructions that are illegal on the 80386.
4072 The default code is faster, but requires at least a 486.
4075 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4076 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4077 same as SSL2_VERSION anyway.
4080 *) New "-showcerts" option for s_client.
4083 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4084 application. Various cleanups and fixes.
4087 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4088 modify error routines to work internally. Add error codes and PBE init
4089 to library startup routines.
4092 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4093 packing functions to asn1 and evp. Changed function names and error
4094 codes along the way.
4097 *) PKCS12 integration: and so it begins... First of several patches to
4098 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4099 objects to objects.h
4102 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4103 and display support for Thawte strong extranet extension.
4106 *) Add LinuxPPC support.
4107 [Jeff Dubrule <igor@pobox.org>]
4109 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4110 bn_div_words in alpha.s.
4111 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4113 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4114 OAEP isn't supported when OpenSSL is built with RSAref.
4115 [Ulf Moeller <ulf@fitug.de>]
4117 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4118 so they no longer are missing under -DNOPROTO.
4119 [Soren S. Jorvang <soren@t.dk>]
4122 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4124 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4125 doesn't work when the session is reused. Coming soon!
4128 *) Fix a security hole, that allows sessions to be reused in the wrong
4129 context thus bypassing client cert protection! All software that uses
4130 client certs and session caches in multiple contexts NEEDS PATCHING to
4131 allow session reuse! A fuller solution is in the works.
4132 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4134 *) Some more source tree cleanups (removed obsolete files
4135 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4136 permission on "config" script to be executable) and a fix for the INSTALL
4138 [Ulf Moeller <ulf@fitug.de>]
4140 *) Remove some legacy and erroneous uses of malloc, free instead of
4142 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4144 *) Make rsa_oaep_test return non-zero on error.
4145 [Ulf Moeller <ulf@fitug.de>]
4147 *) Add support for native Solaris shared libraries. Configure
4148 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4149 if someone would make that last step automatic.
4150 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4152 *) ctx_size was not built with the right compiler during "make links". Fixed.
4155 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4156 except NULL ciphers". This means the default cipher list will no longer
4157 enable NULL ciphers. They need to be specifically enabled e.g. with
4158 the string "DEFAULT:eNULL".
4161 *) Fix to RSA private encryption routines: if p < q then it would
4162 occasionally produce an invalid result. This will only happen with
4163 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4166 *) Be less restrictive and allow also `perl util/perlpath.pl
4167 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4168 because this way one can also use an interpreter named `perl5' (which is
4169 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4170 installed as `perl').
4171 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4173 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4174 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4176 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4177 advapi32.lib to Win32 build and change the pem test comparision
4178 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4179 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4180 and crypto/des/ede_cbcm_enc.c.
4183 *) DES quad checksum was broken on big-endian architectures. Fixed.
4186 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4187 Win32 test batch file so it (might) work again. The Win32 test batch file
4188 is horrible: I feel ill....
4191 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4192 in e_os.h. Audit of header files to check ANSI and non ANSI
4193 sections: 10 functions were absent from non ANSI section and not exported
4194 from Windows DLLs. Fixed up libeay.num for new functions.
4197 *) Make `openssl version' output lines consistent.
4198 [Ralf S. Engelschall]
4200 *) Fix Win32 symbol export lists for BIO functions: Added
4201 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4202 to ms/libeay{16,32}.def.
4203 [Ralf S. Engelschall]
4205 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4206 fine under Unix and passes some trivial tests I've now added. But the
4207 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4208 added to make sure no one expects that this stuff really works in the
4209 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4210 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4212 [Ralf S. Engelschall]
4214 *) Fix the generation of two part addresses in perl.
4215 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4217 *) Add config entry for Linux on MIPS.
4218 [John Tobey <jtobey@channel1.com>]
4220 *) Make links whenever Configure is run, unless we are on Windoze.
4223 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4224 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
projects / openssl.git / blob