5 Changes between 0.9.6 and 0.9.6a [xx XXX 2000]
7 *) rand_win.c fix for Borland C.
10 *) BN_rshift bugfix for n == 0.
13 *) Store verify_result within SSL_SESSION also for client side to
14 avoid potential security hole. (Re-used sessions on the client side
15 always resulted in verify_result==X509_V_OK, not using the original
16 result of the server certificate verification.)
19 *) Disable ssl2_peek and ssl3_peek (i.e., both implementations
20 of SSL_peek) because they both are completely broken.
21 They will be fixed RSN by adding an additional 'peek' parameter
22 to the internal read functions.
25 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
26 The previous value, 12, was not always sufficient for BN_mod_exp().
29 *) Fix typo in get_cert_by_subject() in by_dir.c
30 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
32 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
34 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
35 and not in SSL_clear because the latter is also used by the
36 accept/connect functions; previously, the settings made by
37 SSL_set_read_ahead would be lost during the handshake.
38 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
40 *) Correct util/mkdef.pl to be selective about disabled algorithms.
41 Previously, it would create entries for disableed algorithms no
45 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
47 *) In ssl23_get_client_hello, generate an error message when faced
48 with an initial SSL 3.0/TLS record that is too small to contain the
49 first two bytes of the ClientHello message, i.e. client_version.
50 (Note that this is a pathologic case that probably has never happened
51 in real life.) The previous approach was to use the version number
52 from the record header as a substitute; but our protocol choice
53 should not depend on that one because it is not authenticated
54 by the Finished messages.
57 *) More robust randomness gathering functions for Windows.
58 [Jeffrey Altman <jaltman@columbia.edu>]
60 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
61 not set then we don't setup the error code for issuer check errors
62 to avoid possibly overwriting other errors which the callback does
63 handle. If an application does set the flag then we assume it knows
64 what it is doing and can handle the new informational codes
68 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
69 a general "ANY" type, as such it should be able to decode anything
70 including tagged types. However it didn't check the class so it would
71 wrongly interpret tagged types in the same way as their universal
72 counterpart and unknown types were just rejected. Changed so that the
73 tagged and unknown types are handled in the same way as a SEQUENCE:
74 that is the encoding is stored intact. There is also a new type
75 "V_ASN1_OTHER" which is used when the class is not universal, in this
76 case we have no idea what the actual type is so we just lump them all
80 *) On VMS, stdout may very well lead to a file that is written to
81 in a record-oriented fashion. That means that every write() will
82 write a separate record, which will be read separately by the
83 programs trying to read from it. This can be very confusing.
85 The solution is to put a BIO filter in the way that will buffer
86 text until a linefeed is reached, and then write everything a
87 line at a time, so every record written will be an actual line,
88 not chunks of lines and not (usually doesn't happen, but I've
89 seen it once) several lines in one record. BIO_f_linebuffer() is
92 Currently, it's a VMS-only method, because that's where it has
93 been tested well enough.
96 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
97 it can return incorrect results.
98 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
99 but it was in 0.9.6-beta[12].)
102 *) Disable the check for content being present when verifying detached
103 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
104 include zero length content when signing messages.
107 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
108 BIO_ctrl (for BIO pairs).
111 *) Add DSO method for VMS.
114 *) Bug fix: Montgomery multiplication could produce results with the
118 *) Add RPM specification openssl.spec and modify it to build three
119 packages. The default package contains applications, application
120 documentation and run-time libraries. The devel package contains
121 include files, static libraries and function documentation. The
122 doc package contains the contents of the doc directory. The original
123 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
126 *) Add a large number of documentation files for many SSL routines.
127 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
129 *) Add a configuration entry for Sony News 4.
130 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
132 *) Don't set the two most significant bits to one when generating a
133 random number < q in the DSA library.
136 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
137 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
138 the underlying transport is blocking) if a handshake took place.
139 (The default behaviour is needed by applications such as s_client
140 and s_server that use select() to determine when to use SSL_read;
141 but for applications that know in advance when to expect data, it
142 just makes things more complicated.)
145 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
149 *) Add a few more EBCDIC conditionals that make `req' and `x509'
150 work better on such systems.
151 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
153 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
154 Update PKCS12_parse() so it copies the friendlyName and the
155 keyid to the certificates aux info.
158 *) Fix bug in PKCS7_verify() which caused an infinite loop
159 if there was more than one signature.
160 [Sven Uszpelkat <su@celocom.de>]
162 *) Major change in util/mkdef.pl to include extra information
163 about each symbol, as well as presentig variables as well
164 as functions. This change means that there's n more need
165 to rebuild the .num files when some algorithms are excluded.
168 *) Allow the verify time to be set by an application,
169 rather than always using the current time.
172 *) Phase 2 verify code reorganisation. The certificate
173 verify code now looks up an issuer certificate by a
174 number of criteria: subject name, authority key id
175 and key usage. It also verifies self signed certificates
176 by the same criteria. The main comparison function is
177 X509_check_issued() which performs these checks.
179 Lot of changes were necessary in order to support this
180 without completely rewriting the lookup code.
182 Authority and subject key identifier are now cached.
184 The LHASH 'certs' is X509_STORE has now been replaced
185 by a STACK_OF(X509_OBJECT). This is mainly because an
186 LHASH can't store or retrieve multiple objects with
189 As a result various functions (which were all internal
190 use only) have changed to handle the new X509_STORE
191 structure. This will break anything that messed round
192 with X509_STORE internally.
194 The functions X509_STORE_add_cert() now checks for an
195 exact match, rather than just subject name.
197 The X509_STORE API doesn't directly support the retrieval
198 of multiple certificates matching a given criteria, however
199 this can be worked round by performing a lookup first
200 (which will fill the cache with candidate certificates)
201 and then examining the cache for matches. This is probably
202 the best we can do without throwing out X509_LOOKUP
203 entirely (maybe later...).
205 The X509_VERIFY_CTX structure has been enhanced considerably.
207 All certificate lookup operations now go via a get_issuer()
208 callback. Although this currently uses an X509_STORE it
209 can be replaced by custom lookups. This is a simple way
210 to bypass the X509_STORE hackery necessary to make this
211 work and makes it possible to use more efficient techniques
212 in future. A very simple version which uses a simple
213 STACK for its trusted certificate store is also provided
214 using X509_STORE_CTX_trusted_stack().
216 The verify_cb() and verify() callbacks now have equivalents
217 in the X509_STORE_CTX structure.
219 X509_STORE_CTX also has a 'flags' field which can be used
220 to customise the verify behaviour.
223 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
224 excludes S/MIME capabilities.
227 *) When a certificate request is read in keep a copy of the
228 original encoding of the signed data and use it when outputing
229 again. Signatures then use the original encoding rather than
230 a decoded, encoded version which may cause problems if the
231 request is improperly encoded.
234 *) For consistency with other BIO_puts implementations, call
235 buffer_write(b, ...) directly in buffer_puts instead of calling
238 In BIO_puts, increment b->num_write as in BIO_write.
239 [Peter.Sylvester@EdelWeb.fr]
241 *) Fix BN_mul_word for the case where the word is 0. (We have to use
242 BN_zero, we may not return a BIGNUM with an array consisting of
246 *) Avoid calling abort() from within the library when problems are
247 detected, except if preprocessor symbols have been defined
248 (such as REF_CHECK, BN_DEBUG etc.).
251 *) New openssl application 'rsautl'. This utility can be
252 used for low level RSA operations. DER public key
253 BIO/fp routines also added.
256 *) New Configure entry and patches for compiling on QNX 4.
257 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
259 *) A demo state-machine implementation was sponsored by
260 Nuron (http://www.nuron.com/) and is now available in
264 *) New options added to the 'dgst' utility for signature
265 generation and verification.
268 *) Unrecognized PKCS#7 content types are now handled via a
269 catch all ASN1_TYPE structure. This allows unsupported
270 types to be stored as a "blob" and an application can
271 encode and decode it manually.
274 *) Fix various signed/unsigned issues to make a_strex.c
276 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
278 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
279 length if passed a buffer. ASN1_INTEGER_to_BN failed
280 if passed a NULL BN and its argument was negative.
281 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
283 *) Modification to PKCS#7 encoding routines to output definite
284 length encoding. Since currently the whole structures are in
285 memory there's not real point in using indefinite length
286 constructed encoding. However if OpenSSL is compiled with
287 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
290 *) Added BIO_vprintf() and BIO_vsnprintf().
293 *) Added more prefixes to parse for in the the strings written
294 through a logging bio, to cover all the levels that are available
295 through syslog. The prefixes are now:
297 PANIC, EMERG, EMR => LOG_EMERG
298 ALERT, ALR => LOG_ALERT
299 CRIT, CRI => LOG_CRIT
300 ERROR, ERR => LOG_ERR
301 WARNING, WARN, WAR => LOG_WARNING
302 NOTICE, NOTE, NOT => LOG_NOTICE
303 INFO, INF => LOG_INFO
304 DEBUG, DBG => LOG_DEBUG
306 and as before, if none of those prefixes are present at the
307 beginning of the string, LOG_ERR is chosen.
309 On Win32, the LOG_* levels are mapped according to this:
311 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
312 LOG_WARNING => EVENTLOG_WARNING_TYPE
313 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
317 *) Made it possible to reconfigure with just the configuration
318 argument "reconf" or "reconfigure". The command line arguments
319 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
320 and are retrieved from there when reconfiguring.
324 [Assar Westerlund <assar@sics.se>, Richard Levitte]
326 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
329 *) The obj_dat.pl script was messing up the sorting of object
330 names. The reason was that it compared the quoted version
331 of strings as a result "OCSP" > "OCSP Signing" because
332 " > SPACE. Changed script to store unquoted versions of
333 names and add quotes on output. It was also omitting some
334 names from the lookup table if they were given a default
335 value (that is if SN is missing it is given the same
336 value as LN and vice versa), these are now added on the
337 grounds that if an object has a name we should be able to
338 look it up. Finally added warning output when duplicate
339 short or long names are found.
342 *) Changes needed for Tandem NSK.
343 [Scott Uroff <scott@xypro.com>]
345 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
346 RSA_padding_check_SSLv23(), special padding was never detected
347 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
348 version rollback attacks was not effective.
350 In s23_clnt.c, don't use special rollback-attack detection padding
351 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
352 client; similarly, in s23_srvr.c, don't do the rollback check if
353 SSL 2.0 is the only protocol enabled in the server.
356 *) Make it possible to get hexdumps of unprintable data with 'openssl
357 asn1parse'. By implication, the functions ASN1_parse_dump() and
358 BIO_dump_indent() are added.
361 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
362 these print out strings and name structures based on various
363 flags including RFC2253 support and proper handling of
364 multibyte characters. Added options to the 'x509' utility
365 to allow the various flags to be set.
368 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
369 Also change the functions X509_cmp_current_time() and
370 X509_gmtime_adj() work with an ASN1_TIME structure,
371 this will enable certificates using GeneralizedTime in validity
375 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
376 negative public key encodings) on by default,
377 NO_NEG_PUBKEY_BUG can be set to disable it.
380 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
381 content octets. An i2c_ASN1_OBJECT is unnecessary because
382 the encoding can be trivially obtained from the structure.
385 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
386 not read locks (CRYPTO_r_[un]lock).
389 *) A first attempt at creating official support for shared
390 libraries through configuration. I've kept it so the
391 default is static libraries only, and the OpenSSL programs
392 are always statically linked for now, but there are
393 preparations for dynamic linking in place.
394 This has been tested on Linux and True64.
397 *) Randomness polling function for Win9x, as described in:
398 Peter Gutmann, Software Generation of Practically Strong
402 *) Fix so PRNG is seeded in req if using an already existing
406 *) New options to smime application. -inform and -outform
407 allow alternative formats for the S/MIME message including
408 PEM and DER. The -content option allows the content to be
409 specified separately. This should allow things like Netscape
410 form signing output easier to verify.
413 *) Fix the ASN1 encoding of tags using the 'long form'.
416 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
417 STRING types. These convert content octets to and from the
418 underlying type. The actual tag and length octets are
419 already assumed to have been read in and checked. These
420 are needed because all other string types have virtually
421 identical handling apart from the tag. By having versions
422 of the ASN1 functions that just operate on content octets
423 IMPLICIT tagging can be handled properly. It also allows
424 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
425 and ASN1_INTEGER are identical apart from the tag.
428 *) Change the handling of OID objects as follows:
430 - New object identifiers are inserted in objects.txt, following
431 the syntax given in objects.README.
432 - objects.pl is used to process obj_mac.num and create a new
434 - obj_dat.pl is used to create a new obj_dat.h, using the data in
437 This is currently kind of a hack, and the perl code in objects.pl
438 isn't very elegant, but it works as I intended. The simplest way
439 to check that it worked correctly is to look in obj_dat.h and
440 check the array nid_objs and make sure the objects haven't moved
441 around (this is important!). Additions are OK, as well as
442 consistent name changes.
445 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
448 *) Addition of the command line parameter '-rand file' to 'openssl req'.
449 The given file adds to whatever has already been seeded into the
450 random pool through the RANDFILE configuration file option or
451 environment variable, or the default random state file.
454 *) mkstack.pl now sorts each macro group into lexical order.
455 Previously the output order depended on the order the files
456 appeared in the directory, resulting in needless rewriting
460 *) Patches to make OpenSSL compile under Win32 again. Mostly
461 work arounds for the VC++ problem that it treats func() as
462 func(void). Also stripped out the parts of mkdef.pl that
463 added extra typesafe functions: these no longer exist.
466 *) Reorganisation of the stack code. The macros are now all
467 collected in safestack.h . Each macro is defined in terms of
468 a "stack macro" of the form SKM_<name>(type, a, b). The
469 DEBUG_SAFESTACK is now handled in terms of function casts,
470 this has the advantage of retaining type safety without the
471 use of additional functions. If DEBUG_SAFESTACK is not defined
472 then the non typesafe macros are used instead. Also modified the
473 mkstack.pl script to handle the new form. Needs testing to see
474 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
475 the default if no major problems. Similar behaviour for ASN1_SET_OF
479 *) When some versions of IIS use the 'NET' form of private key the
480 key derivation algorithm is different. Normally MD5(password) is
481 used as a 128 bit RC4 key. In the modified case
482 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
483 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
484 as the old Netscape_RSA functions except they have an additional
485 'sgckey' parameter which uses the modified algorithm. Also added
486 an -sgckey command line option to the rsa utility. Thanks to
487 Adrian Peck <bertie@ncipher.com> for posting details of the modified
488 algorithm to openssl-dev.
491 *) The evp_local.h macros were using 'c.##kname' which resulted in
492 invalid expansion on some systems (SCO 5.0.5 for example).
493 Corrected to 'c.kname'.
494 [Phillip Porch <root@theporch.com>]
496 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
497 a STACK of email addresses from a certificate or request, these look
498 in the subject name and the subject alternative name extensions and
499 omit any duplicate addresses.
502 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
503 This makes DSA verification about 2 % faster.
506 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
507 (meaning that now 2^5 values will be precomputed, which is only 4 KB
508 plus overhead for 1024 bit moduli).
509 This makes exponentiations about 0.5 % faster for 1024 bit
510 exponents (as measured by "openssl speed rsa2048").
513 *) Rename memory handling macros to avoid conflicts with other
515 Malloc => OPENSSL_malloc
516 Malloc_locked => OPENSSL_malloc_locked
517 Realloc => OPENSSL_realloc
521 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
522 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
526 [John Jarvie <jjarvie@newsguy.com>]
528 *) The type-safe stack code has been rejigged. It is now only compiled
529 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
530 by default all type-specific stack functions are "#define"d back to
531 standard stack functions. This results in more streamlined output
532 but retains the type-safety checking possibilities of the original
536 *) The STACK code has been cleaned up, and certain type declarations
537 that didn't make a lot of sense have been brought in line. This has
538 also involved a cleanup of sorts in safestack.h to more correctly
539 map type-safe stack functions onto their plain stack counterparts.
540 This work has also resulted in a variety of "const"ifications of
541 lots of the code, especially "_cmp" operations which should normally
542 be prototyped with "const" parameters anyway.
545 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
546 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
547 (The PRNG state consists of two parts, the large pool 'state' and 'md',
548 where all of 'md' is used each time the PRNG is used, but 'state'
549 is used only indexed by a cyclic counter. As entropy may not be
550 well distributed from the beginning, 'md' is important as a
551 chaining variable. However, the output function chains only half
552 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
553 all of 'md', and seeding with STATE_SIZE dummy bytes will result
554 in all of 'state' being rewritten, with the new values depending
555 on virtually all of 'md'. This overcomes the 80 bit limitation.)
558 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
559 the handshake is continued after ssl_verify_cert_chain();
560 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
561 can lead to 'unexplainable' connection aborts later.
562 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
564 *) Major EVP API cipher revision.
565 Add hooks for extra EVP features. This allows various cipher
566 parameters to be set in the EVP interface. Support added for variable
567 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
568 setting of RC2 and RC5 parameters.
570 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
573 Remove lots of duplicated code from the EVP library. For example *every*
574 cipher init() function handles the 'iv' in the same way according to the
575 cipher mode. They also all do nothing if the 'key' parameter is NULL and
576 for CFB and OFB modes they zero ctx->num.
578 New functionality allows removal of S/MIME code RC2 hack.
580 Most of the routines have the same form and so can be declared in terms
583 By shifting this to the top level EVP_CipherInit() it can be removed from
584 all individual ciphers. If the cipher wants to handle IVs or keys
585 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
588 Change lots of functions like EVP_EncryptUpdate() to now return a
589 value: although software versions of the algorithms cannot fail
590 any installed hardware versions can.
593 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
594 this option is set, tolerate broken clients that send the negotiated
595 protocol version number instead of the requested protocol version
599 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
600 i.e. non-zero for export ciphersuites, zero otherwise.
601 Previous versions had this flag inverted, inconsistent with
602 rsa_tmp_cb (..._TMP_RSA_CB).
603 [Bodo Moeller; problem reported by Amit Chopra]
605 *) Add missing DSA library text string. Work around for some IIS
606 key files with invalid SEQUENCE encoding.
609 *) Add a document (doc/standards.txt) that list all kinds of standards
610 and so on that are implemented in OpenSSL.
613 *) Enhance c_rehash script. Old version would mishandle certificates
614 with the same subject name hash and wouldn't handle CRLs at all.
615 Added -fingerprint option to crl utility, to support new c_rehash
619 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
622 *) Fix for SSL server purpose checking. Server checking was
623 rejecting certificates which had extended key usage present
624 but no ssl client purpose.
625 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
627 *) Make PKCS#12 code work with no password. The PKCS#12 spec
628 is a little unclear about how a blank password is handled.
629 Since the password in encoded as a BMPString with terminating
630 double NULL a zero length password would end up as just the
631 double NULL. However no password at all is different and is
632 handled differently in the PKCS#12 key generation code. NS
633 treats a blank password as zero length. MSIE treats it as no
634 password on export: but it will try both on import. We now do
635 the same: PKCS12_parse() tries zero length and no password if
636 the password is set to "" or NULL (NULL is now a valid password:
637 it wasn't before) as does the pkcs12 application.
640 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
641 perror when PEM_read_bio_X509_REQ fails, the error message must
642 be obtained from the error queue.
645 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
646 it in ERR_remove_state if appropriate, and change ERR_get_state
647 accordingly to avoid race conditions (this is necessary because
648 thread_hash is no longer constant once set).
651 *) Bugfix for linux-elf makefile.one.
654 *) RSA_get_default_method() will now cause a default
655 RSA_METHOD to be chosen if one doesn't exist already.
656 Previously this was only set during a call to RSA_new()
657 or RSA_new_method(NULL) meaning it was possible for
658 RSA_get_default_method() to return NULL.
661 *) Added native name translation to the existing DSO code
662 that will convert (if the flag to do so is set) filenames
663 that are sufficiently small and have no path information
664 into a canonical native form. Eg. "blah" converted to
665 "libblah.so" or "blah.dll" etc.
668 *) New function ERR_error_string_n(e, buf, len) which is like
669 ERR_error_string(e, buf), but writes at most 'len' bytes
670 including the 0 terminator. For ERR_error_string_n, 'buf'
672 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
674 *) CONF library reworked to become more general. A new CONF
675 configuration file reader "class" is implemented as well as a
676 new functions (NCONF_*, for "New CONF") to handle it. The now
677 old CONF_* functions are still there, but are reimplemented to
678 work in terms of the new functions. Also, a set of functions
679 to handle the internal storage of the configuration data is
680 provided to make it easier to write new configuration file
681 reader "classes" (I can definitely see something reading a
682 configuration file in XML format, for example), called _CONF_*,
683 or "the configuration storage API"...
685 The new configuration file reading functions are:
687 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
688 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
690 NCONF_default, NCONF_WIN32
692 NCONF_dump_fp, NCONF_dump_bio
694 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
695 NCONF_new creates a new CONF object. This works in the same way
696 as other interfaces in OpenSSL, like the BIO interface.
697 NCONF_dump_* dump the internal storage of the configuration file,
698 which is useful for debugging. All other functions take the same
699 arguments as the old CONF_* functions wth the exception of the
700 first that must be a `CONF *' instead of a `LHASH *'.
702 To make it easer to use the new classes with the old CONF_* functions,
703 the function CONF_set_default_method is provided.
706 *) Add '-tls1' option to 'openssl ciphers', which was already
707 mentioned in the documentation but had not been implemented.
708 (This option is not yet really useful because even the additional
709 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
712 *) Initial DSO code added into libcrypto for letting OpenSSL (and
713 OpenSSL-based applications) load shared libraries and bind to
714 them in a portable way.
715 [Geoff Thorpe, with contributions from Richard Levitte]
717 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
719 *) Make sure _lrotl and _lrotr are only used with MSVC.
721 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
722 (the default implementation of RAND_status).
724 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
725 to '-clrext' (= clear extensions), as intended and documented.
726 [Bodo Moeller; inconsistency pointed out by Michael Attili
729 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
730 was larger than the MD block size.
731 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
733 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
734 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
735 using the passed key: if the passed key was a private key the result
736 of X509_print(), for example, would be to print out all the private key
740 *) des_quad_cksum() byte order bug fix.
741 [Ulf Möller, using the problem description in krb4-0.9.7, where
742 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
744 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
746 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
748 *) For easily testing in shell scripts whether some command
749 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
750 returns with exit code 0 iff no command of the given name is available.
751 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
752 the output goes to stdout and nothing is printed to stderr.
753 Additional arguments are always ignored.
755 Since for each cipher there is a command of the same name,
756 the 'no-cipher' compilation switches can be tested this way.
758 ('openssl no-XXX' is not able to detect pseudo-commands such
759 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
762 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
765 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
766 is set; it will be thrown away anyway because each handshake creates
768 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
769 to parameters -- in previous versions (since OpenSSL 0.9.3) the
770 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
771 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
774 *) New s_client option -ign_eof: EOF at stdin is ignored, and
775 'Q' and 'R' lose their special meanings (quit/renegotiate).
776 This is part of what -quiet does; unlike -quiet, -ign_eof
777 does not suppress any output.
780 *) Add compatibility options to the purpose and trust code. The
781 purpose X509_PURPOSE_ANY is "any purpose" which automatically
782 accepts a certificate or CA, this was the previous behaviour,
783 with all the associated security issues.
785 X509_TRUST_COMPAT is the old trust behaviour: only and
786 automatically trust self signed roots in certificate store. A
787 new trust setting X509_TRUST_DEFAULT is used to specify that
788 a purpose has no associated trust setting and it should instead
789 use the value in the default purpose.
792 *) Fix the PKCS#8 DSA private key code so it decodes keys again
793 and fix a memory leak.
796 *) In util/mkerr.pl (which implements 'make errors'), preserve
797 reason strings from the previous version of the .c file, as
798 the default to have only downcase letters (and digits) in
799 automatically generated reasons codes is not always appropriate.
802 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
803 using strerror. Previously, ERR_reason_error_string() returned
804 library names as reason strings for SYSerr; but SYSerr is a special
805 case where small numbers are errno values, not library numbers.
808 *) Add '-dsaparam' option to 'openssl dhparam' application. This
809 converts DSA parameters into DH parameters. (When creating parameters,
810 DSA_generate_parameters is used.)
813 *) Include 'length' (recommended exponent length) in C code generated
814 by 'openssl dhparam -C'.
817 *) The second argument to set_label in perlasm was already being used
818 so couldn't be used as a "file scope" flag. Moved to third argument
822 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
823 instead of RAND_bytes for encryption IVs and salts.
826 *) Include RAND_status() into RAND_METHOD instead of implementing
827 it only for md_rand.c Otherwise replacing the PRNG by calling
828 RAND_set_rand_method would be impossible.
831 *) Don't let DSA_generate_key() enter an infinite loop if the random
832 number generation fails.
835 *) New 'rand' application for creating pseudo-random output.
838 *) Added configuration support for Linux/IA64
839 [Rolf Haberrecker <rolf@suse.de>]
841 *) Assembler module support for Mingw32.
844 *) Shared library support for HPUX (in shlib/).
845 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
847 *) Shared library support for Solaris gcc.
848 [Lutz Behnke <behnke@trustcenter.de>]
850 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
852 *) PKCS7_encrypt() was adding text MIME headers twice because they
853 were added manually and by SMIME_crlf_copy().
856 *) In bntest.c don't call BN_rand with zero bits argument.
857 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
859 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
860 case was implemented. This caused BN_div_recp() to fail occasionally.
863 *) Add an optional second argument to the set_label() in the perl
864 assembly language builder. If this argument exists and is set
865 to 1 it signals that the assembler should use a symbol whose
866 scope is the entire file, not just the current function. This
867 is needed with MASM which uses the format label:: for this scope.
868 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
870 *) Change the ASN1 types so they are typedefs by default. Before
871 almost all types were #define'd to ASN1_STRING which was causing
872 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
876 *) Change names of new functions to the new get1/get0 naming
877 convention: After 'get1', the caller owns a reference count
878 and has to call ..._free; 'get0' returns a pointer to some
879 data structure without incrementing reference counters.
880 (Some of the existing 'get' functions increment a reference
881 counter, some don't.)
882 Similarly, 'set1' and 'add1' functions increase reference
883 counters or duplicate objects.
886 *) Allow for the possibility of temp RSA key generation failure:
887 the code used to assume it always worked and crashed on failure.
890 *) Fix potential buffer overrun problem in BIO_printf().
891 [Ulf Möller, using public domain code by Patrick Powell; problem
892 pointed out by David Sacerdote <das33@cornell.edu>]
894 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
895 RAND_egd() and RAND_status(). In the command line application,
896 the EGD socket can be specified like a seed file using RANDFILE
900 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
901 Some CAs (e.g. Verisign) distribute certificates in this form.
904 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
905 list to exclude them. This means that no special compilation option
906 is needed to use anonymous DH: it just needs to be included in the
910 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
911 EVP_MD_type. The old functionality is available in a new macro called
912 EVP_MD_md(). Change code that uses it and update docs.
915 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
916 where the 'void *' argument is replaced by a function pointer argument.
917 Previously 'void *' was abused to point to functions, which works on
918 many platforms, but is not correct. As these functions are usually
919 called by macros defined in OpenSSL header files, most source code
920 should work without changes.
923 *) <openssl/opensslconf.h> (which is created by Configure) now contains
924 sections with information on -D... compiler switches used for
925 compiling the library so that applications can see them. To enable
926 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
927 must be defined. E.g.,
928 #define OPENSSL_ALGORITHM_DEFINES
929 #include <openssl/opensslconf.h>
930 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
931 [Richard Levitte, Ulf and Bodo Möller]
933 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
937 *) Change the 'other' type in certificate aux info to a STACK_OF
938 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
939 the required ASN1 format: arbitrary types determined by an OID.
942 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
943 argument to 'req'. This is not because the function is newer or
944 better than others it just uses the work 'NEW' in the certificate
945 request header lines. Some software needs this.
948 *) Reorganise password command line arguments: now passwords can be
949 obtained from various sources. Delete the PEM_cb function and make
950 it the default behaviour: i.e. if the callback is NULL and the
951 usrdata argument is not NULL interpret it as a null terminated pass
952 phrase. If usrdata and the callback are NULL then the pass phrase
953 is prompted for as usual.
956 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
957 the support is automatically enabled. The resulting binaries will
958 autodetect the card and use it if present.
959 [Ben Laurie and Compaq Inc.]
961 *) Work around for Netscape hang bug. This sends certificate request
962 and server done in one record. Since this is perfectly legal in the
963 SSL/TLS protocol it isn't a "bug" option and is on by default. See
964 the bugs/SSLv3 entry for more info.
967 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
970 *) Add -rand argument to smime and pkcs12 applications and read/write
974 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
977 *) Add command line password options to the remaining applications.
980 *) Bug fix for BN_div_recp() for numerators with an even number of
984 *) More tests in bntest.c, and changed test_bn output.
987 *) ./config recognizes MacOS X now.
990 *) Bug fix for BN_div() when the first words of num and divsor are
991 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
994 *) Add support for various broken PKCS#8 formats, and command line
995 options to produce them.
998 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
999 get temporary BIGNUMs from a BN_CTX.
1002 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1006 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1007 include a #define from the old name to the new. The original intent
1008 was that statically linked binaries could for example just call
1009 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1010 link with digests. This never worked becayse SSLeay_add_all_digests()
1011 and SSLeay_add_all_ciphers() were in the same source file so calling
1012 one would link with the other. They are now in separate source files.
1015 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1018 *) Use a less unusual form of the Miller-Rabin primality test (it used
1019 a binary algorithm for exponentiation integrated into the Miller-Rabin
1020 loop, our standard modexp algorithms are faster).
1023 *) Support for the EBCDIC character set completed.
1024 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1026 *) Source code cleanups: use const where appropriate, eliminate casts,
1027 use void * instead of char * in lhash.
1030 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1031 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1032 this the server could overwrite ephemeral keys that the client
1036 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1037 using 50 iterations of the Rabin-Miller test.
1039 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1040 iterations of the Rabin-Miller test as required by the appendix
1041 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1042 As BN_is_prime_fasttest includes trial division, DSA parameter
1043 generation becomes much faster.
1045 This implies a change for the callback functions in DSA_is_prime
1046 and DSA_generate_parameters: The callback function is called once
1047 for each positive witness in the Rabin-Miller test, not just
1048 occasionally in the inner loop; and the parameters to the
1049 callback function now provide an iteration count for the outer
1050 loop rather than for the current invocation of the inner loop.
1051 DSA_generate_parameters additionally can call the callback
1052 function with an 'iteration count' of -1, meaning that a
1053 candidate has passed the trial division test (when q is generated
1054 from an application-provided seed, trial division is skipped).
1057 *) New function BN_is_prime_fasttest that optionally does trial
1058 division before starting the Rabin-Miller test and has
1059 an additional BN_CTX * argument (whereas BN_is_prime always
1060 has to allocate at least one BN_CTX).
1061 'callback(1, -1, cb_arg)' is called when a number has passed the
1062 trial division stage.
1065 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1069 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1072 *) New function BN_pseudo_rand().
1075 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1076 bignum version of BN_from_montgomery() with the working code from
1077 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1081 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1082 made it impossible to use the same SSL_SESSION data structure in
1083 SSL2 clients in multiple threads.
1086 *) The return value of RAND_load_file() no longer counts bytes obtained
1087 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1088 to seed the PRNG (previously an explicit byte count was required).
1089 [Ulf Möller, Bodo Möller]
1091 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1092 used (char *) instead of (void *) and had casts all over the place.
1095 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1098 *) Retain source code compatibility for BN_prime_checks macro:
1099 BN_is_prime(..., BN_prime_checks, ...) now uses
1100 BN_prime_checks_for_size to determine the appropriate number of
1101 Rabin-Miller iterations.
1104 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1105 DH_CHECK_P_NOT_SAFE_PRIME.
1106 (Check if this is true? OpenPGP calls them "strong".)
1109 *) Merge the functionality of "dh" and "gendh" programs into a new program
1110 "dhparam". The old programs are retained for now but will handle DH keys
1111 (instead of parameters) in future.
1114 *) Make the ciphers, s_server and s_client programs check the return values
1115 when a new cipher list is set.
1118 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1119 ciphers. Before when the 56bit ciphers were enabled the sorting was
1122 The syntax for the cipher sorting has been extended to support sorting by
1123 cipher-strength (using the strength_bits hard coded in the tables).
1124 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1126 Fix a bug in the cipher-command parser: when supplying a cipher command
1127 string with an "undefined" symbol (neither command nor alphanumeric
1128 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1129 an error is flagged.
1131 Due to the strength-sorting extension, the code of the
1132 ssl_create_cipher_list() function was completely rearranged. I hope that
1133 the readability was also increased :-)
1134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1136 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1137 for the first serial number and places 2 in the serial number file. This
1138 avoids problems when the root CA is created with serial number zero and
1139 the first user certificate has the same issuer name and serial number
1143 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1144 the new code. Add documentation for this stuff.
1147 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1148 X509_*() to X509at_*() on the grounds that they don't handle X509
1149 structures and behave in an analagous way to the X509v3 functions:
1150 they shouldn't be called directly but wrapper functions should be used
1153 So we also now have some wrapper functions that call the X509at functions
1154 when passed certificate requests. (TO DO: similar things can be done with
1155 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1156 things. Some of these need some d2i or i2d and print functionality
1157 because they handle more complex structures.)
1160 *) Add missing #ifndefs that caused missing symbols when building libssl
1161 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1162 NO_RSA in ssl/s2*.c.
1163 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1165 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1166 has a return value which indicates the quality of the random data
1167 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1168 error queue. New function RAND_pseudo_bytes() generates output that is
1169 guaranteed to be unique but not unpredictable. RAND_add is like
1170 RAND_seed, but takes an extra argument for an entropy estimate
1171 (RAND_seed always assumes full entropy).
1174 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1175 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1176 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1177 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1178 false-positive rate of at most 2^-80 for random input.
1181 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1184 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1185 in the 0.9.5 release), this returns the chain
1186 from an X509_CTX structure with a dup of the stack and all
1187 the X509 reference counts upped: so the stack will exist
1188 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1191 Also make SSL_SESSION_print() print out the verify return
1195 *) Add manpage for the pkcs12 command. Also change the default
1196 behaviour so MAC iteration counts are used unless the new
1197 -nomaciter option is used. This improves file security and
1198 only older versions of MSIE (4.0 for example) need it.
1201 *) Honor the no-xxx Configure options when creating .DEF files.
1204 *) Add PKCS#10 attributes to field table: challengePassword,
1205 unstructuredName and unstructuredAddress. These are taken from
1206 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1207 international characters are used.
1209 More changes to X509_ATTRIBUTE code: allow the setting of types
1210 based on strings. Remove the 'loc' parameter when adding
1211 attributes because these will be a SET OF encoding which is sorted
1215 *) Initial changes to the 'req' utility to allow request generation
1216 automation. This will allow an application to just generate a template
1217 file containing all the field values and have req construct the
1220 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1221 used all over the place including certificate requests and PKCS#7
1222 structures. They are currently handled manually where necessary with
1223 some primitive wrappers for PKCS#7. The new functions behave in a
1224 manner analogous to the X509 extension functions: they allow
1225 attributes to be looked up by NID and added.
1227 Later something similar to the X509V3 code would be desirable to
1228 automatically handle the encoding, decoding and printing of the
1229 more complex types. The string types like challengePassword can
1230 be handled by the string table functions.
1232 Also modified the multi byte string table handling. Now there is
1233 a 'global mask' which masks out certain types. The table itself
1234 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1235 is useful when for example there is only one permissible type
1236 (as in countryName) and using the mask might result in no valid
1240 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1241 SSL_get_peer_finished to allow applications to obtain the latest
1242 Finished messages sent to the peer or expected from the peer,
1243 respectively. (SSL_get_peer_finished is usually the Finished message
1244 actually received from the peer, otherwise the protocol will be aborted.)
1246 As the Finished message are message digests of the complete handshake
1247 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1248 be used for external authentication procedures when the authentication
1249 provided by SSL/TLS is not desired or is not enough.
1252 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1253 the host supports BWX extension and if Compaq C is present on the
1254 $PATH. Just exploiting of the BWX extension results in 20-30%
1255 performance kick for some algorithms, e.g. DES and RC4 to mention
1256 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1260 *) Add support for MS "fast SGC". This is arguably a violation of the
1261 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1262 weak crypto and after checking the certificate is SGC a second one
1263 with strong crypto. MS SGC stops the first handshake after receiving
1264 the server certificate message and sends a second client hello. Since
1265 a server will typically do all the time consuming operations before
1266 expecting any further messages from the client (server key exchange
1267 is the most expensive) there is little difference between the two.
1269 To get OpenSSL to support MS SGC we have to permit a second client
1270 hello message after we have sent server done. In addition we have to
1271 reset the MAC if we do get this second client hello.
1274 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1275 if a DER encoded private key is RSA or DSA traditional format. Changed
1276 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1277 format DER encoded private key. Newer code should use PKCS#8 format which
1278 has the key type encoded in the ASN1 structure. Added DER private key
1279 support to pkcs8 application.
1282 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1283 ciphersuites has been selected (as required by the SSL 3/TLS 1
1284 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1285 is set, we interpret this as a request to violate the specification
1286 (the worst that can happen is a handshake failure, and 'correct'
1287 behaviour would result in a handshake failure anyway).
1290 *) In SSL_CTX_add_session, take into account that there might be multiple
1291 SSL_SESSION structures with the same session ID (e.g. when two threads
1292 concurrently obtain them from an external cache).
1293 The internal cache can handle only one SSL_SESSION with a given ID,
1294 so if there's a conflict, we now throw out the old one to achieve
1298 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1299 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1300 some routines that use cipher OIDs: some ciphers do not have OIDs
1301 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1305 *) Simplify the trust setting structure and code. Now we just have
1306 two sequences of OIDs for trusted and rejected settings. These will
1307 typically have values the same as the extended key usage extension
1308 and any application specific purposes.
1310 The trust checking code now has a default behaviour: it will just
1311 check for an object with the same NID as the passed id. Functions can
1312 be provided to override either the default behaviour or the behaviour
1313 for a given id. SSL client, server and email already have functions
1314 in place for compatibility: they check the NID and also return "trusted"
1315 if the certificate is self signed.
1318 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1319 traditional format into an EVP_PKEY structure.
1322 *) Add a password callback function PEM_cb() which either prompts for
1323 a password if usr_data is NULL or otherwise assumes it is a null
1324 terminated password. Allow passwords to be passed on command line
1325 environment or config files in a few more utilities.
1328 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1329 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1330 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1331 Update documentation.
1334 *) Support for ASN1 "NULL" type. This could be handled before by using
1335 ASN1_TYPE but there wasn't any function that would try to read a NULL
1336 and produce an error if it couldn't. For compatibility we also have
1337 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1338 don't allocate anything because they don't need to.
1341 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1343 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1345 *) Rebuild of the memory allocation routines used by OpenSSL code and
1346 possibly others as well. The purpose is to make an interface that
1347 provide hooks so anyone can build a separate set of allocation and
1348 deallocation routines to be used by OpenSSL, for example memory
1349 pool implementations, or something else, which was previously hard
1350 since Malloc(), Realloc() and Free() were defined as macros having
1351 the values malloc, realloc and free, respectively (except for Win32
1352 compilations). The same is provided for memory debugging code.
1353 OpenSSL already comes with functionality to find memory leaks, but
1354 this gives people a chance to debug other memory problems.
1356 With these changes, a new set of functions and macros have appeared:
1358 CRYPTO_set_mem_debug_functions() [F]
1359 CRYPTO_get_mem_debug_functions() [F]
1360 CRYPTO_dbg_set_options() [F]
1361 CRYPTO_dbg_get_options() [F]
1362 CRYPTO_malloc_debug_init() [M]
1364 The memory debug functions are NULL by default, unless the library
1365 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1366 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1367 gives the standard debugging functions that come with OpenSSL) or
1368 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1369 provided by the library user) must be used. When the standard
1370 debugging functions are used, CRYPTO_dbg_set_options can be used to
1371 request additional information:
1372 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1373 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1375 Also, things like CRYPTO_set_mem_functions will always give the
1376 expected result (the new set of functions is used for allocation
1377 and deallocation) at all times, regardless of platform and compiler
1380 To finish it up, some functions that were never use in any other
1381 way than through macros have a new API and new semantic:
1384 CRYPTO_dbg_realloc()
1387 All macros of value have retained their old syntax.
1388 [Richard Levitte and Bodo Moeller]
1390 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1391 ordering of SMIMECapabilities wasn't in "strength order" and there
1392 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1396 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1397 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1398 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1400 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1401 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1402 functionality to handle multipart/signed properly) and a utility
1403 called 'smime' to call all this stuff. This is based on code I
1404 originally wrote for Celo who have kindly allowed it to be
1405 included in OpenSSL.
1408 *) Add variants des_set_key_checked and des_set_key_unchecked of
1409 des_set_key (aka des_key_sched). Global variable des_check_key
1410 decides which of these is called by des_set_key; this way
1411 des_check_key behaves as it always did, but applications and
1412 the library itself, which was buggy for des_check_key == 1,
1413 have a cleaner way to pick the version they need.
1416 *) New function PKCS12_newpass() which changes the password of a
1420 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1421 dynamic mix. In both cases the ids can be used as an index into the
1422 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1423 functions so they accept a list of the field values and the
1424 application doesn't need to directly manipulate the X509_TRUST
1428 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1432 *) Modify the way the V3 extension code looks up extensions. This now
1433 works in a similar way to the object code: we have some "standard"
1434 extensions in a static table which is searched with OBJ_bsearch()
1435 and the application can add dynamic ones if needed. The file
1436 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1437 updated whenever a new extension is added to the core code and kept
1438 in ext_nid order. There is a simple program 'tabtest.c' which checks
1439 this. New extensions are not added too often so this file can readily
1440 be maintained manually.
1442 There are two big advantages in doing things this way. The extensions
1443 can be looked up immediately and no longer need to be "added" using
1444 X509V3_add_standard_extensions(): this function now does nothing.
1445 [Side note: I get *lots* of email saying the extension code doesn't
1446 work because people forget to call this function]
1447 Also no dynamic allocation is done unless new extensions are added:
1448 so if we don't add custom extensions there is no need to call
1449 X509V3_EXT_cleanup().
1452 *) Modify enc utility's salting as follows: make salting the default. Add a
1453 magic header, so unsalted files fail gracefully instead of just decrypting
1454 to garbage. This is because not salting is a big security hole, so people
1455 should be discouraged from doing it.
1458 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1459 digest to be passed on the command line but it only used this
1460 parameter when signing a certificate. Modified so all relevant
1461 operations are affected by the digest parameter including the
1462 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1463 DSA key was used because it didn't fix the digest.
1466 *) Initial certificate chain verify code. Currently tests the untrusted
1467 certificates for consistency with the verify purpose (which is set
1468 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1470 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1471 this is because it will reject chains with invalid extensions whereas
1472 every previous version of OpenSSL and SSLeay made no checks at all.
1474 Trust code: checks the root CA for the relevant trust settings. Trust
1475 settings have an initial value consistent with the verify purpose: e.g.
1476 if the verify purpose is for SSL client use it expects the CA to be
1477 trusted for SSL client use. However the default value can be changed to
1478 permit custom trust settings: one example of this would be to only trust
1479 certificates from a specific "secure" set of CAs.
1481 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1482 which should be used for version portability: especially since the
1483 verify structure is likely to change more often now.
1485 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1486 to set them. If not set then assume SSL clients will verify SSL servers
1489 Two new options to the verify program: -untrusted allows a set of
1490 untrusted certificates to be passed in and -purpose which sets the
1491 intended purpose of the certificate. If a purpose is set then the
1492 new chain verify code is used to check extension consistency.
1495 *) Support for the authority information access extension.
1498 *) Modify RSA and DSA PEM read routines to transparently handle
1499 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
1500 public keys in a format compatible with certificate
1501 SubjectPublicKeyInfo structures. Unfortunately there were already
1502 functions called *_PublicKey_* which used various odd formats so
1503 these are retained for compatibility: however the DSA variants were
1504 never in a public release so they have been deleted. Changed dsa/rsa
1505 utilities to handle the new format: note no releases ever handled public
1506 keys so we should be OK.
1508 The primary motivation for this change is to avoid the same fiasco
1509 that dogs private keys: there are several incompatible private key
1510 formats some of which are standard and some OpenSSL specific and
1511 require various evil hacks to allow partial transparent handling and
1512 even then it doesn't work with DER formats. Given the option anything
1513 other than PKCS#8 should be dumped: but the other formats have to
1514 stay in the name of compatibility.
1516 With public keys and the benefit of hindsight one standard format
1517 is used which works with EVP_PKEY, RSA or DSA structures: though
1518 it clearly returns an error if you try to read the wrong kind of key.
1520 Added a -pubkey option to the 'x509' utility to output the public key.
1521 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
1522 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
1523 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
1524 that do the same as the EVP_PKEY_assign_*() except they up the
1525 reference count of the added key (they don't "swallow" the
1529 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
1530 CRLs would fail if the file contained no certificates or no CRLs:
1531 added a new function to read in both types and return the number
1532 read: this means that if none are read it will be an error. The
1533 DER versions of the certificate and CRL reader would always fail
1534 because it isn't possible to mix certificates and CRLs in DER format
1535 without choking one or the other routine. Changed this to just read
1536 a certificate: this is the best we can do. Also modified the code
1537 in apps/verify.c to take notice of return codes: it was previously
1538 attempting to read in certificates from NULL pointers and ignoring
1539 any errors: this is one reason why the cert and CRL reader seemed
1540 to work. It doesn't check return codes from the default certificate
1541 routines: these may well fail if the certificates aren't installed.
1544 *) Code to support otherName option in GeneralName.
1547 *) First update to verify code. Change the verify utility
1548 so it warns if it is passed a self signed certificate:
1549 for consistency with the normal behaviour. X509_verify
1550 has been modified to it will now verify a self signed
1551 certificate if *exactly* the same certificate appears
1552 in the store: it was previously impossible to trust a
1553 single self signed certificate. This means that:
1554 openssl verify ss.pem
1555 now gives a warning about a self signed certificate but
1556 openssl verify -CAfile ss.pem ss.pem
1560 *) For servers, store verify_result in SSL_SESSION data structure
1561 (and add it to external session representation).
1562 This is needed when client certificate verifications fails,
1563 but an application-provided verification callback (set by
1564 SSL_CTX_set_cert_verify_callback) allows accepting the session
1565 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
1566 but returns 1): When the session is reused, we have to set
1567 ssl->verify_result to the appropriate error code to avoid
1569 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
1571 *) Fix a bug in the new PKCS#7 code: it didn't consider the
1572 case in PKCS7_dataInit() where the signed PKCS7 structure
1573 didn't contain any existing data because it was being created.
1574 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
1576 *) Add a salt to the key derivation routines in enc.c. This
1577 forms the first 8 bytes of the encrypted file. Also add a
1578 -S option to allow a salt to be input on the command line.
1581 *) New function X509_cmp(). Oddly enough there wasn't a function
1582 to compare two certificates. We do this by working out the SHA1
1583 hash and comparing that. X509_cmp() will be needed by the trust
1587 *) SSL_get1_session() is like SSL_get_session(), but increments
1588 the reference count in the SSL_SESSION returned.
1589 [Geoff Thorpe <geoff@eu.c2.net>]
1591 *) Fix for 'req': it was adding a null to request attributes.
1592 Also change the X509_LOOKUP and X509_INFO code to handle
1593 certificate auxiliary information.
1596 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
1600 *) Add the possibility to add extra information to the memory leak
1601 detecting output, to form tracebacks, showing from where each
1602 allocation was originated: CRYPTO_push_info("constant string") adds
1603 the string plus current file name and line number to a per-thread
1604 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
1605 is like calling CYRPTO_pop_info() until the stack is empty.
1606 Also updated memory leak detection code to be multi-thread-safe.
1609 *) Add options -text and -noout to pkcs7 utility and delete the
1610 encryption options which never did anything. Update docs.
1613 *) Add options to some of the utilities to allow the pass phrase
1614 to be included on either the command line (not recommended on
1615 OSes like Unix) or read from the environment. Update the
1616 manpages and fix a few bugs.
1619 *) Add a few manpages for some of the openssl commands.
1622 *) Fix the -revoke option in ca. It was freeing up memory twice,
1623 leaking and not finding already revoked certificates.
1626 *) Extensive changes to support certificate auxiliary information.
1627 This involves the use of X509_CERT_AUX structure and X509_AUX
1628 functions. An X509_AUX function such as PEM_read_X509_AUX()
1629 can still read in a certificate file in the usual way but it
1630 will also read in any additional "auxiliary information". By
1631 doing things this way a fair degree of compatibility can be
1632 retained: existing certificates can have this information added
1633 using the new 'x509' options.
1635 Current auxiliary information includes an "alias" and some trust
1636 settings. The trust settings will ultimately be used in enhanced
1637 certificate chain verification routines: currently a certificate
1638 can only be trusted if it is self signed and then it is trusted
1642 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
1643 The problem was that one of the replacement routines had not been working
1644 since SSLeay releases. For now the offending routine has been replaced
1645 with non-optimised assembler. Even so, this now gives around 95%
1646 performance improvement for 1024 bit RSA signs.
1649 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
1650 handling. Most clients have the effective key size in bits equal to
1651 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
1652 A few however don't do this and instead use the size of the decrypted key
1653 to determine the RC2 key length and the AlgorithmIdentifier to determine
1654 the effective key length. In this case the effective key length can still
1655 be 40 bits but the key length can be 168 bits for example. This is fixed
1656 by manually forcing an RC2 key into the EVP_PKEY structure because the
1657 EVP code can't currently handle unusual RC2 key sizes: it always assumes
1658 the key length and effective key length are equal.
1661 *) Add a bunch of functions that should simplify the creation of
1662 X509_NAME structures. Now you should be able to do:
1663 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
1664 and have it automatically work out the correct field type and fill in
1665 the structures. The more adventurous can try:
1666 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
1667 and it will (hopefully) work out the correct multibyte encoding.
1670 *) Change the 'req' utility to use the new field handling and multibyte
1671 copy routines. Before the DN field creation was handled in an ad hoc
1672 way in req, ca, and x509 which was rather broken and didn't support
1673 BMPStrings or UTF8Strings. Since some software doesn't implement
1674 BMPStrings or UTF8Strings yet, they can be enabled using the config file
1675 using the dirstring_type option. See the new comment in the default
1676 openssl.cnf for more info.
1679 *) Make crypto/rand/md_rand.c more robust:
1680 - Assure unique random numbers after fork().
1681 - Make sure that concurrent threads access the global counter and
1682 md serializably so that we never lose entropy in them
1683 or use exactly the same state in multiple threads.
1684 Access to the large state is not always serializable because
1685 the additional locking could be a performance killer, and
1686 md should be large enough anyway.
1689 *) New file apps/app_rand.c with commonly needed functionality
1690 for handling the random seed file.
1692 Use the random seed file in some applications that previously did not:
1694 dsaparam -genkey (which also ignored its '-rand' option),
1697 x509 (when signing).
1698 Except on systems with /dev/urandom, it is crucial to have a random
1699 seed file at least for key creation, DSA signing, and for DH exchanges;
1700 for RSA signatures we could do without one.
1702 gendh and gendsa (unlike genrsa) used to read only the first byte
1703 of each file listed in the '-rand' option. The function as previously
1704 found in genrsa is now in app_rand.c and is used by all programs
1705 that support '-rand'.
1708 *) In RAND_write_file, use mode 0600 for creating files;
1709 don't just chmod when it may be too late.
1712 *) Report an error from X509_STORE_load_locations
1713 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
1716 *) New function ASN1_mbstring_copy() this copies a string in either
1717 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
1718 into an ASN1_STRING type. A mask of permissible types is passed
1719 and it chooses the "minimal" type to use or an error if not type
1723 *) Add function equivalents to the various macros in asn1.h. The old
1724 macros are retained with an M_ prefix. Code inside the library can
1725 use the M_ macros. External code (including the openssl utility)
1726 should *NOT* in order to be "shared library friendly".
1729 *) Add various functions that can check a certificate's extensions
1730 to see if it usable for various purposes such as SSL client,
1731 server or S/MIME and CAs of these types. This is currently
1732 VERY EXPERIMENTAL but will ultimately be used for certificate chain
1733 verification. Also added a -purpose flag to x509 utility to
1734 print out all the purposes.
1737 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
1741 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
1742 for, obtain and decode and extension and obtain its critical flag.
1743 This allows all the necessary extension code to be handled in a
1744 single function call.
1747 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
1748 platforms. See crypto/rc4/rc4_enc.c for further details.
1751 *) New -noout option to asn1parse. This causes no output to be produced
1752 its main use is when combined with -strparse and -out to extract data
1753 from a file (which may not be in ASN.1 format).
1756 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
1757 when producing the local key id.
1758 [Richard Levitte <levitte@stacken.kth.se>]
1760 *) New option -dhparam in s_server. This allows a DH parameter file to be
1761 stated explicitly. If it is not stated then it tries the first server
1762 certificate file. The previous behaviour hard coded the filename
1766 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
1767 a public key to be input or output. For example:
1768 openssl rsa -in key.pem -pubout -out pubkey.pem
1769 Also added necessary DSA public key functions to handle this.
1772 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
1773 in the message. This was handled by allowing
1774 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
1775 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
1777 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
1778 to the end of the strings whereas this didn't. This would cause problems
1779 if strings read with d2i_ASN1_bytes() were later modified.
1780 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
1782 *) Fix for base64 decode bug. When a base64 bio reads only one line of
1783 data and it contains EOF it will end up returning an error. This is
1784 caused by input 46 bytes long. The cause is due to the way base64
1785 BIOs find the start of base64 encoded data. They do this by trying a
1786 trial decode on each line until they find one that works. When they
1787 do a flag is set and it starts again knowing it can pass all the
1788 data directly through the decoder. Unfortunately it doesn't reset
1789 the context it uses. This means that if EOF is reached an attempt
1790 is made to pass two EOFs through the context and this causes the
1791 resulting error. This can also cause other problems as well. As is
1792 usual with these problems it takes *ages* to find and the fix is
1793 trivial: move one line.
1794 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
1796 *) Ugly workaround to get s_client and s_server working under Windows. The
1797 old code wouldn't work because it needed to select() on sockets and the
1798 tty (for keypresses and to see if data could be written). Win32 only
1799 supports select() on sockets so we select() with a 1s timeout on the
1800 sockets and then see if any characters are waiting to be read, if none
1801 are present then we retry, we also assume we can always write data to
1802 the tty. This isn't nice because the code then blocks until we've
1803 received a complete line of data and it is effectively polling the
1804 keyboard at 1s intervals: however it's quite a bit better than not
1805 working at all :-) A dedicated Windows application might handle this
1806 with an event loop for example.
1809 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
1810 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
1811 will be called when RSA_sign() and RSA_verify() are used. This is useful
1812 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
1813 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
1814 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
1815 This necessitated the support of an extra signature type NID_md5_sha1
1816 for SSL signatures and modifications to the SSL library to use it instead
1817 of calling RSA_public_decrypt() and RSA_private_encrypt().
1820 *) Add new -verify -CAfile and -CApath options to the crl program, these
1821 will lookup a CRL issuers certificate and verify the signature in a
1822 similar way to the verify program. Tidy up the crl program so it
1823 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
1824 less strict. It will now permit CRL extensions even if it is not
1825 a V2 CRL: this will allow it to tolerate some broken CRLs.
1828 *) Initialize all non-automatic variables each time one of the openssl
1829 sub-programs is started (this is necessary as they may be started
1830 multiple times from the "OpenSSL>" prompt).
1831 [Lennart Bang, Bodo Moeller]
1833 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
1834 removing all other RSA functionality (this is what NO_RSA does). This
1835 is so (for example) those in the US can disable those operations covered
1836 by the RSA patent while allowing storage and parsing of RSA keys and RSA
1840 *) Non-copying interface to BIO pairs.
1841 (still largely untested)
1844 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
1845 ASCII string. This was handled independently in various places before.
1848 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
1849 UTF8 strings a character at a time.
1852 *) Use client_version from client hello to select the protocol
1853 (s23_srvr.c) and for RSA client key exchange verification
1854 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
1857 *) Add various utility functions to handle SPKACs, these were previously
1858 handled by poking round in the structure internals. Added new function
1859 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
1860 print, verify and generate SPKACs. Based on an original idea from
1861 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
1864 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
1867 *) Allow the config file extension section to be overwritten on the
1868 command line. Based on an original idea from Massimiliano Pala
1869 <madwolf@comune.modena.it>. The new option is called -extensions
1870 and can be applied to ca, req and x509. Also -reqexts to override
1871 the request extensions in req and -crlexts to override the crl extensions
1875 *) Add new feature to the SPKAC handling in ca. Now you can include
1876 the same field multiple times by preceding it by "XXXX." for example:
1879 this is the same syntax as used in the req config file.
1882 *) Allow certificate extensions to be added to certificate requests. These
1883 are specified in a 'req_extensions' option of the req section of the
1884 config file. They can be printed out with the -text option to req but
1885 are otherwise ignored at present.
1888 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
1889 data read consists of only the final block it would not decrypted because
1890 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
1891 A misplaced 'break' also meant the decrypted final block might not be
1892 copied until the next read.
1895 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
1896 a few extra parameters to the DH structure: these will be useful if
1897 for example we want the value of 'q' or implement X9.42 DH.
1900 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
1901 provides hooks that allow the default DSA functions or functions on a
1902 "per key" basis to be replaced. This allows hardware acceleration and
1903 hardware key storage to be handled without major modification to the
1904 library. Also added low level modexp hooks and CRYPTO_EX structure and
1905 associated functions.
1908 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
1909 as "read only": it can't be written to and the buffer it points to will
1910 not be freed. Reading from a read only BIO is much more efficient than
1911 a normal memory BIO. This was added because there are several times when
1912 an area of memory needs to be read from a BIO. The previous method was
1913 to create a memory BIO and write the data to it, this results in two
1914 copies of the data and an O(n^2) reading algorithm. There is a new
1915 function BIO_new_mem_buf() which creates a read only memory BIO from
1916 an area of memory. Also modified the PKCS#7 routines to use read only
1920 *) Bugfix: ssl23_get_client_hello did not work properly when called in
1921 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
1922 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
1923 but a retry condition occured while trying to read the rest.
1926 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
1927 NID_pkcs7_encrypted by default: this was wrong since this should almost
1928 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
1929 the encrypted data type: this is a more sensible place to put it and it
1930 allows the PKCS#12 code to be tidied up that duplicated this
1934 *) Changed obj_dat.pl script so it takes its input and output files on
1935 the command line. This should avoid shell escape redirection problems
1939 *) Initial support for certificate extension requests, these are included
1940 in things like Xenroll certificate requests. Included functions to allow
1941 extensions to be obtained and added.
1944 *) -crlf option to s_client and s_server for sending newlines as
1945 CRLF (as required by many protocols).
1948 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
1950 *) Install libRSAglue.a when OpenSSL is built with RSAref.
1951 [Ralf S. Engelschall]
1953 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
1954 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
1956 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
1960 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
1961 DH parameters/keys (q is lost during that conversion, but the resulting
1962 DH parameters contain its length).
1964 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
1965 much faster than DH_generate_parameters (which creates parameters
1966 where p = 2*q + 1), and also the smaller q makes DH computations
1967 much more efficient (160-bit exponentiation instead of 1024-bit
1968 exponentiation); so this provides a convenient way to support DHE
1969 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
1970 utter importance to use
1971 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
1973 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
1974 when such DH parameters are used, because otherwise small subgroup
1975 attacks may become possible!
1978 *) Avoid memory leak in i2d_DHparams.
1981 *) Allow the -k option to be used more than once in the enc program:
1982 this allows the same encrypted message to be read by multiple recipients.
1985 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
1986 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
1987 it will always use the numerical form of the OID, even if it has a short
1991 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
1992 method only got called if p,q,dmp1,dmq1,iqmp components were present,
1993 otherwise bn_mod_exp was called. In the case of hardware keys for example
1994 no private key components need be present and it might store extra data
1995 in the RSA structure, which cannot be accessed from bn_mod_exp.
1996 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
1997 private key operations.
2000 *) Added support for SPARC Linux.
2003 *) pem_password_cb function type incompatibly changed from
2004 typedef int pem_password_cb(char *buf, int size, int rwflag);
2006 ....(char *buf, int size, int rwflag, void *userdata);
2007 so that applications can pass data to their callbacks:
2008 The PEM[_ASN1]_{read,write}... functions and macros now take an
2009 additional void * argument, which is just handed through whenever
2010 the password callback is called.
2011 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2013 New function SSL_CTX_set_default_passwd_cb_userdata.
2015 Compatibility note: As many C implementations push function arguments
2016 onto the stack in reverse order, the new library version is likely to
2017 interoperate with programs that have been compiled with the old
2018 pem_password_cb definition (PEM_whatever takes some data that
2019 happens to be on the stack as its last argument, and the callback
2020 just ignores this garbage); but there is no guarantee whatsoever that
2023 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2024 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2025 problems not only on Windows, but also on some Unix platforms.
2026 To avoid problematic command lines, these definitions are now in an
2027 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2028 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2031 *) MIPS III/IV assembler module is reimplemented.
2034 *) More DES library cleanups: remove references to srand/rand and
2035 delete an unused file.
2038 *) Add support for the the free Netwide assembler (NASM) under Win32,
2039 since not many people have MASM (ml) and it can be hard to obtain.
2040 This is currently experimental but it seems to work OK and pass all
2041 the tests. Check out INSTALL.W32 for info.
2044 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2045 without temporary keys kept an extra copy of the server key,
2046 and connections with temporary keys did not free everything in case
2050 *) New function RSA_check_key and new openssl rsa option -check
2051 for verifying the consistency of RSA keys.
2052 [Ulf Moeller, Bodo Moeller]
2054 *) Various changes to make Win32 compile work:
2055 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2056 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2057 comparison" warnings.
2058 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2061 *) Add a debugging option to PKCS#5 v2 key generation function: when
2062 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2063 derived keys are printed to stderr.
2066 *) Copy the flags in ASN1_STRING_dup().
2067 [Roman E. Pavlov <pre@mo.msk.ru>]
2069 *) The x509 application mishandled signing requests containing DSA
2070 keys when the signing key was also DSA and the parameters didn't match.
2072 It was supposed to omit the parameters when they matched the signing key:
2073 the verifying software was then supposed to automatically use the CA's
2074 parameters if they were absent from the end user certificate.
2076 Omitting parameters is no longer recommended. The test was also
2077 the wrong way round! This was probably due to unusual behaviour in
2078 EVP_cmp_parameters() which returns 1 if the parameters match.
2079 This meant that parameters were omitted when they *didn't* match and
2080 the certificate was useless. Certificates signed with 'ca' didn't have
2082 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2084 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2085 The interface is as follows:
2086 Applications can use
2087 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2088 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2089 "off" is now the default.
2090 The library internally uses
2091 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2092 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2093 to disable memory-checking temporarily.
2095 Some inconsistent states that previously were possible (and were
2096 even the default) are now avoided.
2098 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2099 with each memory chunk allocated; this is occasionally more helpful
2100 than just having a counter.
2102 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2104 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2108 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2109 which largely parallels "options", but is for changing API behaviour,
2110 whereas "options" are about protocol behaviour.
2111 Initial "mode" flags are:
2113 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2114 a single record has been written.
2115 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2116 retries use the same buffer location.
2117 (But all of the contents must be
2121 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2124 *) Fix problems with no-hmac etc.
2125 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2127 *) New functions RSA_get_default_method(), RSA_set_method() and
2128 RSA_get_method(). These allows replacement of RSA_METHODs without having
2129 to mess around with the internals of an RSA structure.
2132 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2133 Also really enable memory leak checks in openssl.c and in some
2135 [Chad C. Mulligan, Bodo Moeller]
2137 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2138 up the length of negative integers. This has now been simplified to just
2139 store the length when it is first determined and use it later, rather
2140 than trying to keep track of where data is copied and updating it to
2142 [Steve Henson, reported by Brien Wheeler
2143 <bwheeler@authentica-security.com>]
2145 *) Add a new function PKCS7_signatureVerify. This allows the verification
2146 of a PKCS#7 signature but with the signing certificate passed to the
2147 function itself. This contrasts with PKCS7_dataVerify which assumes the
2148 certificate is present in the PKCS#7 structure. This isn't always the
2149 case: certificates can be omitted from a PKCS#7 structure and be
2150 distributed by "out of band" means (such as a certificate database).
2153 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2154 function prototypes in pem.h, also change util/mkdef.pl to add the
2155 necessary function names.
2158 *) mk1mf.pl (used by Windows builds) did not properly read the
2159 options set by Configure in the top level Makefile, and Configure
2160 was not even able to write more than one option correctly.
2161 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2164 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2165 file to be loaded from a BIO or FILE pointer. The BIO version will
2166 for example allow memory BIOs to contain config info.
2169 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2170 Whoever hopes to achieve shared-library compatibility across versions
2171 must use this, not the compile-time macro.
2172 (Exercise 0.9.4: Which is the minimum library version required by
2174 Note: All this applies only to multi-threaded programs, others don't
2178 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2179 through a BIO pair triggered the default case, i.e.
2180 SSLerr(...,SSL_R_UNKNOWN_STATE).
2183 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2184 can use the SSL library even if none of the specific BIOs is
2188 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2189 for the encoded length.
2190 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2192 *) Add initial documentation of the X509V3 functions.
2195 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2196 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2197 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2198 secure PKCS#8 private key format with a high iteration count.
2201 *) Fix determination of Perl interpreter: A perl or perl5
2202 _directory_ in $PATH was also accepted as the interpreter.
2203 [Ralf S. Engelschall]
2205 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2206 wrong with it but it was very old and did things like calling
2207 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2211 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2212 to use the new extension code.
2215 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2216 with macros. This should make it easier to change their form, add extra
2217 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2221 *) Add to configuration table a new entry that can specify an alternative
2222 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2223 according to Mark Crispin <MRC@Panda.COM>.
2227 *) DES CBC did not update the IV. Weird.
2230 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2231 Changing the behaviour of the former might break existing programs --
2232 where IV updating is needed, des_ncbc_encrypt can be used.
2235 *) When bntest is run from "make test" it drives bc to check its
2236 calculations, as well as internally checking them. If an internal check
2237 fails, it needs to cause bc to give a non-zero result or make test carries
2238 on without noticing the failure. Fixed.
2241 *) DES library cleanups.
2244 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2245 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2246 ciphers. NOTE: although the key derivation function has been verified
2247 against some published test vectors it has not been extensively tested
2248 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2252 *) Instead of "mkdir -p", which is not fully portable, use new
2253 Perl script "util/mkdir-p.pl".
2256 *) Rewrite the way password based encryption (PBE) is handled. It used to
2257 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2258 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2259 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2260 the 'parameter' field of the AlgorithmIdentifier is passed to the
2261 underlying key generation function so it must do its own ASN1 parsing.
2262 This has also changed the EVP_PBE_CipherInit() function which now has a
2263 'parameter' argument instead of literal salt and iteration count values
2264 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2267 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2268 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2269 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2270 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2271 value was just used as a "magic string" and not used directly its
2272 value doesn't matter.
2275 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2279 *) "linux-sparc64" configuration (ultrapenguin).
2280 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2281 "linux-sparc" configuration.
2282 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2284 *) config now generates no-xxx options for missing ciphers.
2287 *) Support the EBCDIC character set (work in progress).
2288 File ebcdic.c not yet included because it has a different license.
2289 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2291 *) Support BS2000/OSD-POSIX.
2292 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2294 *) Make callbacks for key generation use void * instead of char *.
2297 *) Make S/MIME samples compile (not yet tested).
2300 *) Additional typesafe stacks.
2303 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2307 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2309 *) New configuration variant "sco5-gcc".
2311 *) Updated some demos.
2312 [Sean O Riordain, Wade Scholine]
2314 *) Add missing BIO_free at exit of pkcs12 application.
2317 *) Fix memory leak in conf.c.
2320 *) Updates for Win32 to assembler version of MD5.
2323 *) Set #! path to perl in apps/der_chop to where we found it
2324 instead of using a fixed path.
2327 *) SHA library changes for irix64-mips4-cc.
2330 *) Improvements for VMS support.
2334 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2336 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2337 This also avoids the problems with SC4.2 and unpatched SC5.
2338 [Andy Polyakov <appro@fy.chalmers.se>]
2340 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2341 These are required because of the typesafe stack would otherwise break
2342 existing code. If old code used a structure member which used to be STACK
2343 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2344 sk_num or sk_value it would produce an error because the num, data members
2345 are not present in STACK_OF. Now it just produces a warning. sk_set
2346 replaces the old method of assigning a value to sk_value
2347 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2348 that does this will no longer work (and should use sk_set instead) but
2349 this could be regarded as a "questionable" behaviour anyway.
2352 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2353 correctly handle encrypted S/MIME data.
2356 *) Change type of various DES function arguments from des_cblock
2357 (which means, in function argument declarations, pointer to char)
2358 to des_cblock * (meaning pointer to array with 8 char elements),
2359 which allows the compiler to do more typechecking; it was like
2360 that back in SSLeay, but with lots of ugly casts.
2362 Introduce new type const_des_cblock.
2365 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2366 problems: find RecipientInfo structure that matches recipient certificate
2367 and initialise the ASN1 structures properly based on passed cipher.
2370 *) Belatedly make the BN tests actually check the results.
2373 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2374 to and from BNs: it was completely broken. New compilation option
2375 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2376 key elements as negative integers.
2379 *) Reorganize and speed up MD5.
2380 [Andy Polyakov <appro@fy.chalmers.se>]
2383 [Richard Levitte <richard@levitte.org>]
2385 *) New option -out to asn1parse to allow the parsed structure to be
2386 output to a file. This is most useful when combined with the -strparse
2387 option to examine the output of things like OCTET STRINGS.
2390 *) Make SSL library a little more fool-proof by not requiring any longer
2391 that SSL_set_{accept,connect}_state be called before
2392 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2393 in many applications because usually everything *appeared* to work as
2394 intended anyway -- now it really works as intended).
2397 *) Move openssl.cnf out of lib/.
2400 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2401 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2402 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2403 [Ralf S. Engelschall]
2405 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2406 handle PKCS#7 enveloped data properly.
2407 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2409 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2410 copying pointers. The cert_st handling is changed by this in
2411 various ways (and thus what used to be known as ctx->default_cert
2412 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2413 any longer when s->cert does not give us what we need).
2414 ssl_cert_instantiate becomes obsolete by this change.
2415 As soon as we've got the new code right (possibly it already is?),
2416 we have solved a couple of bugs of the earlier code where s->cert
2417 was used as if it could not have been shared with other SSL structures.
2419 Note that using the SSL API in certain dirty ways now will result
2420 in different behaviour than observed with earlier library versions:
2421 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2422 does not influence s as it used to.
2424 In order to clean up things more thoroughly, inside SSL_SESSION
2425 we don't use CERT any longer, but a new structure SESS_CERT
2426 that holds per-session data (if available); currently, this is
2427 the peer's certificate chain and, for clients, the server's certificate
2428 and temporary key. CERT holds only those values that can have
2429 meaningful defaults in an SSL_CTX.
2432 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2433 from the internal representation. Various PKCS#7 fixes: remove some
2434 evil casts and set the enc_dig_alg field properly based on the signing
2438 *) Allow PKCS#12 password to be set from the command line or the
2439 environment. Let 'ca' get its config file name from the environment
2440 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2444 *) Allow certificate policies extension to use an IA5STRING for the
2445 organization field. This is contrary to the PKIX definition but
2446 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2450 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2451 without disallowing inline assembler and the like for non-pedantic builds.
2454 *) Support Borland C++ builder.
2455 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2460 *) SHA-1 cleanups and performance enhancements.
2461 [Andy Polyakov <appro@fy.chalmers.se>]
2463 *) Sparc v8plus assembler for the bignum library.
2464 [Andy Polyakov <appro@fy.chalmers.se>]
2466 *) Accept any -xxx and +xxx compiler options in Configure.
2469 *) Update HPUX configuration.
2472 *) Add missing sk_<type>_unshift() function to safestack.h
2473 [Ralf S. Engelschall]
2475 *) New function SSL_CTX_use_certificate_chain_file that sets the
2476 "extra_cert"s in addition to the certificate. (This makes sense
2477 only for "PEM" format files, as chains as a whole are not
2481 *) Support verify_depth from the SSL API.
2482 x509_vfy.c had what can be considered an off-by-one-error:
2483 Its depth (which was not part of the external interface)
2484 was actually counting the number of certificates in a chain;
2485 now it really counts the depth.
2488 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
2489 instead of X509err, which often resulted in confusing error
2490 messages since the error codes are not globally unique
2491 (e.g. an alleged error in ssl3_accept when a certificate
2492 didn't match the private key).
2494 *) New function SSL_CTX_set_session_id_context that allows to set a default
2495 value (so that you don't need SSL_set_session_id_context for each
2496 connection using the SSL_CTX).
2499 *) OAEP decoding bug fix.
2502 *) Support INSTALL_PREFIX for package builders, as proposed by
2506 *) New Configure options "threads" and "no-threads". For systems
2507 where the proper compiler options are known (currently Solaris
2508 and Linux), "threads" is the default.
2511 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
2514 *) Install various scripts to $(OPENSSLDIR)/misc, not to
2515 $(INSTALLTOP)/bin -- they shouldn't clutter directories
2516 such as /usr/local/bin.
2519 *) "make linux-shared" to build shared libraries.
2520 [Niels Poppe <niels@netbox.org>]
2522 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
2525 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
2526 extension adding in x509 utility.
2529 *) Remove NOPROTO sections and error code comments.
2532 *) Partial rewrite of the DEF file generator to now parse the ANSI
2536 *) New Configure options --prefix=DIR and --openssldir=DIR.
2539 *) Complete rewrite of the error code script(s). It is all now handled
2540 by one script at the top level which handles error code gathering,
2541 header rewriting and C source file generation. It should be much better
2542 than the old method: it now uses a modified version of Ulf's parser to
2543 read the ANSI prototypes in all header files (thus the old K&R definitions
2544 aren't needed for error creation any more) and do a better job of
2545 translating function codes into names. The old 'ASN1 error code imbedded
2546 in a comment' is no longer necessary and it doesn't use .err files which
2547 have now been deleted. Also the error code call doesn't have to appear all
2548 on one line (which resulted in some large lines...).
2551 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
2554 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
2555 0 (which usually indicates a closed connection), but continue reading.
2558 *) Fix some race conditions.
2561 *) Add support for CRL distribution points extension. Add Certificate
2562 Policies and CRL distribution points documentation.
2565 *) Move the autogenerated header file parts to crypto/opensslconf.h.
2568 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
2569 8 of keying material. Merlin has also confirmed interop with this fix
2570 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
2571 [Merlin Hughes <merlin@baltimore.ie>]
2573 *) Fix lots of warnings.
2574 [Richard Levitte <levitte@stacken.kth.se>]
2576 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
2577 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
2578 [Richard Levitte <levitte@stacken.kth.se>]
2580 *) Fix problems with sizeof(long) == 8.
2581 [Andy Polyakov <appro@fy.chalmers.se>]
2583 *) Change functions to ANSI C.
2586 *) Fix typos in error codes.
2587 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
2589 *) Remove defunct assembler files from Configure.
2592 *) SPARC v8 assembler BIGNUM implementation.
2593 [Andy Polyakov <appro@fy.chalmers.se>]
2595 *) Support for Certificate Policies extension: both print and set.
2596 Various additions to support the r2i method this uses.
2599 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
2600 return a const string when you are expecting an allocated buffer.
2603 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
2604 types DirectoryString and DisplayText.
2607 *) Add code to allow r2i extensions to access the configuration database,
2608 add an LHASH database driver and add several ctx helper functions.
2611 *) Fix an evil bug in bn_expand2() which caused various BN functions to
2612 fail when they extended the size of a BIGNUM.
2615 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
2616 support typesafe stack.
2619 *) Fix typo in SSL_[gs]et_options().
2620 [Nils Frostberg <nils@medcom.se>]
2622 *) Delete various functions and files that belonged to the (now obsolete)
2623 old X509V3 handling code.
2626 *) New Configure option "rsaref".
2629 *) Don't auto-generate pem.h.
2632 *) Introduce type-safe ASN.1 SETs.
2635 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
2636 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
2638 *) Introduce type-safe STACKs. This will almost certainly break lots of code
2639 that links with OpenSSL (well at least cause lots of warnings), but fear
2640 not: the conversion is trivial, and it eliminates loads of evil casts. A
2641 few STACKed things have been converted already. Feel free to convert more.
2642 In the fullness of time, I'll do away with the STACK type altogether.
2645 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
2646 specified in <certfile> by updating the entry in the index.txt file.
2647 This way one no longer has to edit the index.txt file manually for
2648 revoking a certificate. The -revoke option does the gory details now.
2649 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
2651 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
2652 `-text' option at all and this way the `-noout -text' combination was
2653 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
2654 [Ralf S. Engelschall]
2656 *) Make sure a corresponding plain text error message exists for the
2657 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
2658 verify callback function determined that a certificate was revoked.
2659 [Ralf S. Engelschall]
2661 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
2662 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
2663 all available cipers including rc5, which was forgotten until now.
2664 In order to let the testing shell script know which algorithms
2665 are available, a new (up to now undocumented) command
2666 "openssl list-cipher-commands" is used.
2669 *) Bugfix: s_client occasionally would sleep in select() when
2670 it should have checked SSL_pending() first.
2673 *) New functions DSA_do_sign and DSA_do_verify to provide access to
2674 the raw DSA values prior to ASN.1 encoding.
2677 *) Tweaks to Configure
2678 [Niels Poppe <niels@netbox.org>]
2680 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
2684 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
2687 *) New config option to avoid instructions that are illegal on the 80386.
2688 The default code is faster, but requires at least a 486.
2691 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
2692 SSL2_SERVER_VERSION (not used at all) macros, which are now the
2693 same as SSL2_VERSION anyway.
2696 *) New "-showcerts" option for s_client.
2699 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
2700 application. Various cleanups and fixes.
2703 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
2704 modify error routines to work internally. Add error codes and PBE init
2705 to library startup routines.
2708 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
2709 packing functions to asn1 and evp. Changed function names and error
2710 codes along the way.
2713 *) PKCS12 integration: and so it begins... First of several patches to
2714 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
2715 objects to objects.h
2718 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
2719 and display support for Thawte strong extranet extension.
2722 *) Add LinuxPPC support.
2723 [Jeff Dubrule <igor@pobox.org>]
2725 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
2726 bn_div_words in alpha.s.
2727 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
2729 *) Make sure the RSA OAEP test is skipped under -DRSAref because
2730 OAEP isn't supported when OpenSSL is built with RSAref.
2731 [Ulf Moeller <ulf@fitug.de>]
2733 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
2734 so they no longer are missing under -DNOPROTO.
2735 [Soren S. Jorvang <soren@t.dk>]
2738 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
2740 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
2741 doesn't work when the session is reused. Coming soon!
2744 *) Fix a security hole, that allows sessions to be reused in the wrong
2745 context thus bypassing client cert protection! All software that uses
2746 client certs and session caches in multiple contexts NEEDS PATCHING to
2747 allow session reuse! A fuller solution is in the works.
2748 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
2750 *) Some more source tree cleanups (removed obsolete files
2751 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
2752 permission on "config" script to be executable) and a fix for the INSTALL
2754 [Ulf Moeller <ulf@fitug.de>]
2756 *) Remove some legacy and erroneous uses of malloc, free instead of
2758 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
2760 *) Make rsa_oaep_test return non-zero on error.
2761 [Ulf Moeller <ulf@fitug.de>]
2763 *) Add support for native Solaris shared libraries. Configure
2764 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
2765 if someone would make that last step automatic.
2766 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
2768 *) ctx_size was not built with the right compiler during "make links". Fixed.
2771 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
2772 except NULL ciphers". This means the default cipher list will no longer
2773 enable NULL ciphers. They need to be specifically enabled e.g. with
2774 the string "DEFAULT:eNULL".
2777 *) Fix to RSA private encryption routines: if p < q then it would
2778 occasionally produce an invalid result. This will only happen with
2779 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
2782 *) Be less restrictive and allow also `perl util/perlpath.pl
2783 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
2784 because this way one can also use an interpreter named `perl5' (which is
2785 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
2786 installed as `perl').
2787 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2789 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
2790 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2792 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
2793 advapi32.lib to Win32 build and change the pem test comparision
2794 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
2795 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
2796 and crypto/des/ede_cbcm_enc.c.
2799 *) DES quad checksum was broken on big-endian architectures. Fixed.
2802 *) Comment out two functions in bio.h that aren't implemented. Fix up the
2803 Win32 test batch file so it (might) work again. The Win32 test batch file
2804 is horrible: I feel ill....
2807 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
2808 in e_os.h. Audit of header files to check ANSI and non ANSI
2809 sections: 10 functions were absent from non ANSI section and not exported
2810 from Windows DLLs. Fixed up libeay.num for new functions.
2813 *) Make `openssl version' output lines consistent.
2814 [Ralf S. Engelschall]
2816 *) Fix Win32 symbol export lists for BIO functions: Added
2817 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
2818 to ms/libeay{16,32}.def.
2819 [Ralf S. Engelschall]
2821 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
2822 fine under Unix and passes some trivial tests I've now added. But the
2823 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
2824 added to make sure no one expects that this stuff really works in the
2825 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
2826 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
2828 [Ralf S. Engelschall]
2830 *) Fix the generation of two part addresses in perl.
2831 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
2833 *) Add config entry for Linux on MIPS.
2834 [John Tobey <jtobey@channel1.com>]
2836 *) Make links whenever Configure is run, unless we are on Windoze.
2839 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
2840 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
2844 *) Add a useful kludge to allow package maintainers to specify compiler and
2845 other platforms details on the command line without having to patch the
2846 Configure script everytime: One now can use ``perl Configure
2847 <id>:<details>'', i.e. platform ids are allowed to have details appended
2848 to them (seperated by colons). This is treated as there would be a static
2849 pre-configured entry in Configure's %table under key <id> with value
2850 <details> and ``perl Configure <id>'' is called. So, when you want to
2851 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
2852 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
2853 now, which overrides the FreeBSD-elf entry on-the-fly.
2854 [Ralf S. Engelschall]
2856 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
2859 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
2860 on the `perl Configure ...' command line. This way one can compile
2861 OpenSSL libraries with Position Independent Code (PIC) which is needed
2862 for linking it into DSOs.
2863 [Ralf S. Engelschall]
2865 *) Remarkably, export ciphers were totally broken and no-one had noticed!
2869 *) Cleaned up the LICENSE document: The official contact for any license
2870 questions now is the OpenSSL core team under openssl-core@openssl.org.
2871 And add a paragraph about the dual-license situation to make sure people
2872 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
2873 to the OpenSSL toolkit.
2874 [Ralf S. Engelschall]
2876 *) General source tree makefile cleanups: Made `making xxx in yyy...'
2877 display consistent in the source tree and replaced `/bin/rm' by `rm'.
2878 Additonally cleaned up the `make links' target: Remove unnecessary
2879 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
2880 to speed processing and no longer clutter the display with confusing
2881 stuff. Instead only the actually done links are displayed.
2882 [Ralf S. Engelschall]
2884 *) Permit null encryption ciphersuites, used for authentication only. It used
2885 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
2886 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
2890 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
2891 signed attributes when verifying signatures (this would break them),
2892 the detached data encoding was wrong and public keys obtained using
2893 X509_get_pubkey() weren't freed.
2896 *) Add text documentation for the BUFFER functions. Also added a work around
2897 to a Win95 console bug. This was triggered by the password read stuff: the
2898 last character typed gets carried over to the next fread(). If you were
2899 generating a new cert request using 'req' for example then the last
2900 character of the passphrase would be CR which would then enter the first
2904 *) Added the new `Includes OpenSSL Cryptography Software' button as
2905 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
2906 button and can be used by applications based on OpenSSL to show the
2907 relationship to the OpenSSL project.
2908 [Ralf S. Engelschall]
2910 *) Remove confusing variables in function signatures in files
2911 ssl/ssl_lib.c and ssl/ssl.h.
2912 [Lennart Bong <lob@kulthea.stacken.kth.se>]
2914 *) Don't install bss_file.c under PREFIX/include/
2915 [Lennart Bong <lob@kulthea.stacken.kth.se>]
2917 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
2918 functions that return function pointers and has support for NT specific
2919 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
2920 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
2921 unsigned to signed types: this was killing the Win32 compile.
2924 *) Add new certificate file to stack functions,
2925 SSL_add_dir_cert_subjects_to_stack() and
2926 SSL_add_file_cert_subjects_to_stack(). These largely supplant
2927 SSL_load_client_CA_file(), and can be used to add multiple certs easily
2928 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
2929 This means that Apache-SSL and similar packages don't have to mess around
2930 to add as many CAs as they want to the preferred list.
2933 *) Experiment with doxygen documentation. Currently only partially applied to
2935 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
2936 openssl.doxy as the configuration file.
2939 *) Get rid of remaining C++-style comments which strict C compilers hate.
2940 [Ralf S. Engelschall, pointed out by Carlos Amengual]
2942 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
2943 compiled in by default: it has problems with large keys.
2946 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
2947 DH private keys and/or callback functions which directly correspond to
2948 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
2949 is needed for applications which have to configure certificates on a
2950 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
2952 For the RSA certificate situation is makes no difference, but
2953 for the DSA certificate situation this fixes the "no shared cipher"
2954 problem where the OpenSSL cipher selection procedure failed because the
2955 temporary keys were not overtaken from the context and the API provided
2956 no way to reconfigure them.
2957 The new functions now let applications reconfigure the stuff and they
2958 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
2959 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
2960 non-public-API function ssl_cert_instantiate() is used as a helper
2961 function and also to reduce code redundancy inside ssl_rsa.c.
2962 [Ralf S. Engelschall]
2964 *) Move s_server -dcert and -dkey options out of the undocumented feature
2965 area because they are useful for the DSA situation and should be
2966 recognized by the users.
2967 [Ralf S. Engelschall]
2969 *) Fix the cipher decision scheme for export ciphers: the export bits are
2970 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
2971 SSL_EXP_MASK. So, the original variable has to be used instead of the
2972 already masked variable.
2973 [Richard Levitte <levitte@stacken.kth.se>]
2975 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
2976 [Richard Levitte <levitte@stacken.kth.se>]
2978 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
2979 from `int' to `unsigned int' because it's a length and initialized by
2980 EVP_DigestFinal() which expects an `unsigned int *'.
2981 [Richard Levitte <levitte@stacken.kth.se>]
2983 *) Don't hard-code path to Perl interpreter on shebang line of Configure
2984 script. Instead use the usual Shell->Perl transition trick.
2985 [Ralf S. Engelschall]
2987 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
2988 (in addition to RSA certificates) to match the behaviour of `openssl dsa
2989 -noout -modulus' as it's already the case for `openssl rsa -noout
2990 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
2991 currently the public key is printed (a decision which was already done by
2992 `openssl dsa -modulus' in the past) which serves a similar purpose.
2993 Additionally the NO_RSA no longer completely removes the whole -modulus
2994 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
2996 [Ralf S. Engelschall]
2998 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
2999 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3000 [Arne Ansper <arne@ats.cyber.ee>]
3002 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3003 to be added. Now both 'req' and 'ca' can use new objects defined in the
3007 *) Add cool BIO that does syslog (or event log on NT).
3008 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3010 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3011 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3012 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3013 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3016 *) Add preliminary config info for new extension code.
3019 *) Make RSA_NO_PADDING really use no padding.
3020 [Ulf Moeller <ulf@fitug.de>]
3022 *) Generate errors when private/public key check is done.
3025 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3026 for some CRL extensions and new objects added.
3029 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3030 key usage extension and fuller support for authority key id.
3033 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3034 padding method for RSA, which is recommended for new applications in PKCS
3035 #1 v2.0 (RFC 2437, October 1998).
3036 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3037 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3038 against Bleichbacher's attack on RSA.
3039 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3042 *) Updates to the new SSL compression code
3043 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3045 *) Fix so that the version number in the master secret, when passed
3046 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3047 (because the server will not accept higher), that the version number
3048 is 0x03,0x01, not 0x03,0x00
3049 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3051 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3052 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3053 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3056 *) Support for RAW extensions where an arbitrary extension can be
3057 created by including its DER encoding. See apps/openssl.cnf for
3061 *) Make sure latest Perl versions don't interpret some generated C array
3062 code as Perl array code in the crypto/err/err_genc.pl script.
3063 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3065 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3066 not many people have the assembler. Various Win32 compilation fixes and
3067 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3071 *) Modify configure script 'Configure' to automatically create crypto/date.h
3072 file under Win32 and also build pem.h from pem.org. New script
3073 util/mkfiles.pl to create the MINFO file on environments that can't do a
3074 'make files': perl util/mkfiles.pl >MINFO should work.
3077 *) Major rework of DES function declarations, in the pursuit of correctness
3078 and purity. As a result, many evil casts evaporated, and some weirdness,
3079 too. You may find this causes warnings in your code. Zapping your evil
3080 casts will probably fix them. Mostly.
3083 *) Fix for a typo in asn1.h. Bug fix to object creation script
3084 obj_dat.pl. It considered a zero in an object definition to mean
3085 "end of object": none of the objects in objects.h have any zeros
3086 so it wasn't spotted.
3087 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3089 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3090 Masking (CBCM). In the absence of test vectors, the best I have been able
3091 to do is check that the decrypt undoes the encrypt, so far. Send me test
3092 vectors if you have them.
3095 *) Correct calculation of key length for export ciphers (too much space was
3096 allocated for null ciphers). This has not been tested!
3099 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3100 message is now correct (it understands "crypto" and "ssl" on its
3101 command line). There is also now an "update" option. This will update
3102 the util/ssleay.num and util/libeay.num files with any new functions.
3104 perl util/mkdef.pl crypto ssl update
3105 it will update them.
3108 *) Overhauled the Perl interface (perl/*):
3109 - ported BN stuff to OpenSSL's different BN library
3110 - made the perl/ source tree CVS-aware
3111 - renamed the package from SSLeay to OpenSSL (the files still contain
3112 their history because I've copied them in the repository)
3113 - removed obsolete files (the test scripts will be replaced
3114 by better Test::Harness variants in the future)
3115 [Ralf S. Engelschall]
3117 *) First cut for a very conservative source tree cleanup:
3118 1. merge various obsolete readme texts into doc/ssleay.txt
3119 where we collect the old documents and readme texts.
3120 2. remove the first part of files where I'm already sure that we no
3121 longer need them because of three reasons: either they are just temporary
3122 files which were left by Eric or they are preserved original files where
3123 I've verified that the diff is also available in the CVS via "cvs diff
3124 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3125 the crypto/md/ stuff).
3126 [Ralf S. Engelschall]
3128 *) More extension code. Incomplete support for subject and issuer alt
3129 name, issuer and authority key id. Change the i2v function parameters
3130 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3131 what that's for :-) Fix to ASN1 macro which messed up
3132 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3135 *) Preliminary support for ENUMERATED type. This is largely copied from the
3139 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3140 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3142 *) Make sure `make rehash' target really finds the `openssl' program.
3143 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3145 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3146 like to hear about it if this slows down other processors.
3149 *) Add CygWin32 platform information to Configure script.
3150 [Alan Batie <batie@aahz.jf.intel.com>]
3152 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3153 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3155 *) New program nseq to manipulate netscape certificate sequences
3158 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3162 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3163 but the BN code had some problems that would cause failures when
3164 doing certificate verification and some other functions.
3165 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3167 *) Add ASN1 and PEM code to support netscape certificate sequences.
3170 *) Add ASN1 and PEM code to support netscape certificate sequences.
3173 *) Add several PKIX and private extended key usage OIDs.
3176 *) Modify the 'ca' program to handle the new extension code. Modify
3177 openssl.cnf for new extension format, add comments.
3180 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3181 and add a sample to openssl.cnf so req -x509 now adds appropriate
3185 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3186 error code, add initial support to X509_print() and x509 application.
3189 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3190 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3191 stuff is currently isolated and isn't even compiled yet.
3194 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3195 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3196 Removed the versions check from X509 routines when loading extensions:
3197 this allows certain broken certificates that don't set the version
3198 properly to be processed.
3201 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3202 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3203 can still be regenerated with "make depend".
3206 *) Spelling mistake in C version of CAST-128.
3207 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3209 *) Changes to the error generation code. The perl script err-code.pl
3210 now reads in the old error codes and retains the old numbers, only
3211 adding new ones if necessary. It also only changes the .err files if new
3212 codes are added. The makefiles have been modified to only insert errors
3213 when needed (to avoid needlessly modifying header files). This is done
3214 by only inserting errors if the .err file is newer than the auto generated
3215 C file. To rebuild all the error codes from scratch (the old behaviour)
3216 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3217 or delete all the .err files.
3220 *) CAST-128 was incorrectly implemented for short keys. The C version has
3221 been fixed, but is untested. The assembler versions are also fixed, but
3222 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3223 to regenerate it if needed.
3224 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3225 Hagino <itojun@kame.net>]
3227 *) File was opened incorrectly in randfile.c.
3228 [Ulf Möller <ulf@fitug.de>]
3230 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3231 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3232 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3233 al: it's just almost always a UTCTime. Note this patch adds new error
3234 codes so do a "make errors" if there are problems.
3237 *) Correct Linux 1 recognition in config.
3238 [Ulf Möller <ulf@fitug.de>]
3240 *) Remove pointless MD5 hash when using DSA keys in ca.
3241 [Anonymous <nobody@replay.com>]
3243 *) Generate an error if given an empty string as a cert directory. Also
3244 generate an error if handed NULL (previously returned 0 to indicate an
3245 error, but didn't set one).
3246 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3248 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3251 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3252 parameters. This was causing a warning which killed off the Win32 compile.
3255 *) Remove C++ style comments from crypto/bn/bn_local.h.
3256 [Neil Costigan <neil.costigan@celocom.com>]
3258 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3259 based on a text string, looking up short and long names and finally
3260 "dot" format. The "dot" format stuff didn't work. Added new function
3261 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3262 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3263 OID is not part of the table.
3266 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3267 X509_LOOKUP_by_alias().
3270 *) Sort openssl functions by name.
3273 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3274 encryption from sample DSA keys (in case anyone is interested the password
3278 *) Make _all_ *_free functions accept a NULL pointer.
3279 [Frans Heymans <fheymans@isaserver.be>]
3281 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3283 [Anonymous <nobody@replay.com>]
3285 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3286 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3288 *) Don't blow it for numeric -newkey arguments to apps/req.
3289 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3291 *) Temp key "for export" tests were wrong in s3_srvr.c.
3292 [Anonymous <nobody@replay.com>]
3294 *) Add prototype for temp key callback functions
3295 SSL_CTX_set_tmp_{rsa,dh}_callback().
3298 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3299 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3302 *) X509_name_add_entry() freed the wrong thing after an error.
3303 [Arne Ansper <arne@ats.cyber.ee>]
3305 *) rsa_eay.c would attempt to free a NULL context.
3306 [Arne Ansper <arne@ats.cyber.ee>]
3308 *) BIO_s_socket() had a broken should_retry() on Windoze.
3309 [Arne Ansper <arne@ats.cyber.ee>]
3311 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3312 [Arne Ansper <arne@ats.cyber.ee>]
3314 *) Make sure the already existing X509_STORE->depth variable is initialized
3315 in X509_STORE_new(), but document the fact that this variable is still
3316 unused in the certificate verification process.
3317 [Ralf S. Engelschall]
3319 *) Fix the various library and apps files to free up pkeys obtained from
3320 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3323 *) Fix reference counting in X509_PUBKEY_get(). This makes
3324 demos/maurice/example2.c work, amongst others, probably.
3325 [Steve Henson and Ben Laurie]
3327 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3328 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3329 are no longer created. This way we have a single and consistent command
3330 line interface `openssl <command>', similar to `cvs <command>'.
3331 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3333 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3334 BIT STRING wrapper always have zero unused bits.
3337 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3340 *) Make the top-level INSTALL documentation easier to understand.
3343 *) Makefiles updated to exit if an error occurs in a sub-directory
3344 make (including if user presses ^C) [Paul Sutton]
3346 *) Make Montgomery context stuff explicit in RSA data structure.
3349 *) Fix build order of pem and err to allow for generated pem.h.
3352 *) Fix renumbering bug in X509_NAME_delete_entry().
3355 *) Enhanced the err-ins.pl script so it makes the error library number
3356 global and can add a library name. This is needed for external ASN1 and
3357 other error libraries.
3360 *) Fixed sk_insert which never worked properly.
3363 *) Fix ASN1 macros so they can handle indefinite length construted
3364 EXPLICIT tags. Some non standard certificates use these: they can now
3368 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3369 into a single doc/ssleay.txt bundle. This way the information is still
3370 preserved but no longer messes up this directory. Now it's new room for
3371 the new set of documenation files.
3372 [Ralf S. Engelschall]
3374 *) SETs were incorrectly DER encoded. This was a major pain, because they
3375 shared code with SEQUENCEs, which aren't coded the same. This means that
3376 almost everything to do with SETs or SEQUENCEs has either changed name or
3377 number of arguments.
3378 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
3380 *) Fix test data to work with the above.
3383 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
3384 was already fixed by Eric for 0.9.1 it seems.
3385 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
3387 *) Autodetect FreeBSD3.
3390 *) Fix various bugs in Configure. This affects the following platforms:
3394 unixware-2.0-pentium
3398 *) Eliminate generated files from CVS. Reorder tests to regenerate files
3399 before they are needed.
3402 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
3406 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
3408 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
3409 changed SSLeay to OpenSSL in version strings.
3410 [Ralf S. Engelschall]
3412 *) Some fixups to the top-level documents.
3415 *) Fixed the nasty bug where rsaref.h was not found under compile-time
3416 because the symlink to include/ was missing.
3417 [Ralf S. Engelschall]
3419 *) Incorporated the popular no-RSA/DSA-only patches
3420 which allow to compile a RSA-free SSLeay.
3421 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
3423 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
3424 when "ssleay" is still not found.
3425 [Ralf S. Engelschall]
3427 *) Added more platforms to Configure: Cray T3E, HPUX 11,
3428 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
3430 *) Updated the README file.
3431 [Ralf S. Engelschall]
3433 *) Added various .cvsignore files in the CVS repository subdirs
3434 to make a "cvs update" really silent.
3435 [Ralf S. Engelschall]
3437 *) Recompiled the error-definition header files and added
3438 missing symbols to the Win32 linker tables.
3439 [Ralf S. Engelschall]
3441 *) Cleaned up the top-level documents;
3442 o new files: CHANGES and LICENSE
3443 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
3444 o merged COPYRIGHT into LICENSE
3445 o removed obsolete TODO file
3446 o renamed MICROSOFT to INSTALL.W32
3447 [Ralf S. Engelschall]
3449 *) Removed dummy files from the 0.9.1b source tree:
3450 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
3451 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
3452 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
3453 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
3454 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
3455 [Ralf S. Engelschall]
3457 *) Added various platform portability fixes.
3460 *) The Genesis of the OpenSSL rpject:
3461 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
3462 Young and Tim J. Hudson created while they were working for C2Net until
3464 [The OpenSSL Project]
3467 Changes between 0.9.0b and 0.9.1b [not released]
3469 *) Updated a few CA certificates under certs/
3472 *) Changed some BIGNUM api stuff.
3475 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
3476 DGUX x86, Linux Alpha, etc.
3479 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
3480 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
3484 *) Add -strparse option to asn1pars program which parses nested
3486 [Dr Stephen Henson <shenson@bigfoot.com>]
3488 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
3491 *) DSA fix for "ca" program.
3494 *) Added "-genkey" option to "dsaparam" program.
3497 *) Added RIPE MD160 (rmd160) message digest.
3500 *) Added -a (all) option to "ssleay version" command.
3503 *) Added PLATFORM define which is the id given to Configure.
3506 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
3509 *) Extended the ASN.1 parser routines.
3512 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
3515 *) Added a BN_CTX to the BN library.
3518 *) Fixed the weak key values in DES library
3521 *) Changed API in EVP library for cipher aliases.
3524 *) Added support for RC2/64bit cipher.
3527 *) Converted the lhash library to the crypto/mem.c functions.
3530 *) Added more recognized ASN.1 object ids.
3533 *) Added more RSA padding checks for SSL/TLS.
3536 *) Added BIO proxy/filter functionality.
3539 *) Added extra_certs to SSL_CTX which can be used
3540 send extra CA certificates to the client in the CA cert chain sending
3541 process. It can be configured with SSL_CTX_add_extra_chain_cert().
3544 *) Now Fortezza is denied in the authentication phase because
3545 this is key exchange mechanism is not supported by SSLeay at all.
3548 *) Additional PKCS1 checks.
3551 *) Support the string "TLSv1" for all TLS v1 ciphers.
3554 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
3555 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
3558 *) Fixed a few memory leaks.
3561 *) Fixed various code and comment typos.
3564 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
3565 bytes sent in the client random.
3566 [Edward Bishop <ebishop@spyglass.com>]