4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7 call failed, free the DSA structure.
10 *) Add another call level for memory allocation routines, thereby
11 allowing memory allocation callbacks that can be given file
12 name and line number information. The basic functionality
13 remains, as well as the original possibility to just replace
14 malloc(), realloc() and free(). The new functions that can be
15 registered help users provide variants of malloc(), realloc()
16 and free() that take two extra arguments, a const char* and an
17 int. To register and find out the current settings for those
18 hooks, the following functions are provided:
20 CRYPTO_set_mem_ex_functions
21 CRYPTO_set_locked_mem_ex_functions
22 CRYPTO_get_mem_ex_functions
23 CRYPTO_get_locked_mem_ex_functions
25 They work the same way as the corresponding CRYPTO_set_mem_functions
26 and friends with one exception: giving NULL as arguments will restore
27 the internal hooks to internal routines and will still make the above
28 functions return 1 and not 0.
30 This functionality was created as a direct request to add the
31 possibility to interface with the Windows debugging routines
32 _malloc_dbg, _realloc_dbg and _free_dbg.
35 *) Fix to uni2asc() to cope with zero length Unicode strings.
36 These are present in some PKCS#12 files.
39 *) Finish off removing the remaining LHASH function pointer casts.
40 There should no longer be any prototype-casting required when using
41 the LHASH abstraction, and any casts that remain are "bugs". See
42 the callback types and macros at the head of lhash.h for details
43 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
46 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
47 If an EGD or PRNGD is running and enough entropy is returned, automatic
48 seeding like with /dev/[u]random will be performed.
49 Positions tried are: /etc/entropy, /var/run/egd-pool.
52 *) Change the Unix RAND_poll() variant to be able to poll several
53 random devices and only read data for a small fragment of time
54 to avoid hangs. Also separate out the Unix variant to it's own
55 file, rand_unix.c. For VMS, there's a currently-empty rand_vms.c.
58 *) Move OCSP client related routines to ocsp_cl.c. These
59 provide utility functions which an application needing
60 to issue a request to an OCSP responder and analyse the
61 response will typically need: as opposed to those which an
62 OCSP responder itself would need which will be added later.
64 OCSP_request_sign() signs an OCSP request with an API similar
65 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
66 response. OCSP_response_get1_basic() extracts basic response
67 from response. OCSP_resp_find_status(): finds and extracts status
68 information from an OCSP_CERTID structure (which will be created
69 when the request structure is built). These are built from lower
70 level functions which work on OCSP_SINGLERESP structures but
71 wont normally be used unless the application wishes to examine
72 extensions in the OCSP response for example.
74 Replace nonce routines with a pair of functions.
75 OCSP_request_add1_nonce() adds a nonce value and optionally
76 generates a random value. OCSP_check_nonce() checks the
77 validity of the nonce in an OCSP response.
80 *) Change function OCSP_request_add() to OCSP_request_add0_id().
81 This doesn't copy the supplied OCSP_CERTID and avoids the
82 need to free up the newly created id. Change return type
83 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
84 This can then be used to add extensions to the request.
85 Deleted OCSP_request_new(), since most of its functionality
86 is now in OCSP_REQUEST_new() (and the case insensitive name
87 clash) apart from the ability to set the request name which
88 will be added elsewhere.
91 *) Update OCSP API. Remove obsolete extensions argument from
92 various functions. Extensions are now handled using the new
93 OCSP extension code. New simple OCSP HTTP function which
94 can be used to send requests and parse the response.
97 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
98 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
99 uses the special reorder version of SET OF to sort the attributes
100 and reorder them to match the encoded order. This resolves a long
101 standing problem: a verify on a PKCS7 structure just after signing
102 it used to fail because the attribute order did not match the
103 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
104 it uses the received order. This is necessary to tolerate some broken
105 software that does not order SET OF. This is handled by encoding
106 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
107 to produce the required SET OF.
110 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
111 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
112 files to get correct declarations of the ASN.1 item variables.
115 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
116 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
117 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
118 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
119 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
120 ASN1_ITEM and no wrapper functions.
123 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
124 replace the old function pointer based I/O routines. Change most of
125 the *_d2i_bio() and *_d2i_fp() functions to use these.
128 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
129 lines, recognice more "algorithms" that can be deselected, and make
130 it complain about algorithm deselection that isn't recognised.
133 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
134 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
135 to use new functions. Add NO_ASN1_OLD which can be set to remove
136 some old style ASN1 functions: this can be used to determine if old
137 code will still work when these eventually go away.
140 *) New extension functions for OCSP structures, these follow the
141 same conventions as certificates and CRLs.
144 *) New function X509V3_add1_i2d(). This automatically encodes and
145 adds an extension. Its behaviour can be customised with various
146 flags to append, replace or delete. Various wrappers added for
147 certifcates and CRLs.
150 *) Fix to avoid calling the underlying ASN1 print routine when
151 an extension cannot be parsed. Correct a typo in the
152 OCSP_SERVICELOC extension. Tidy up print OCSP format.
155 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
156 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
157 when writing a 32767 byte record.
158 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
160 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
161 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
163 (RSA objects have a reference count access to which is protected
164 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
165 so they are meant to be shared between threads.)
166 [Bodo Moeller, Geoff Thorpe; original patch submitted by
167 "Reddie, Steven" <Steven.Reddie@ca.com>]
169 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
170 entries for variables.
173 *) Fix a deadlock in CRYPTO_mem_leaks().
176 *) Add functionality to apps/openssl.c for detecting locking
177 problems: As the program is single-threaded, all we have
178 to do is register a locking callback using an array for
179 storing which locks are currently held by the program.
182 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
183 SSL_get_ex_data_X509_STORE_idx(), which is used in
184 ssl_verify_cert_chain() and thus can be called at any time
185 during TLS/SSL handshakes so that thread-safety is essential.
186 Unfortunately, the ex_data design is not at all suited
187 for multi-threaded use, so it probably should be abolished.
190 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
191 [Broadcom, tweaked and integrated by Geoff Thorpe]
193 *) Move common extension printing code to new function
194 X509V3_print_extensions(). Reorganise OCSP print routines and
195 implement some needed OCSP ASN1 functions. Add OCSP extensions.
198 *) New function X509_signature_print() to remove duplication in some
202 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
203 set (this was treated exactly the same as SET OF previously). This
204 is used to reorder the STACK representing the structure to match the
205 encoding. This will be used to get round a problem where a PKCS7
206 structure which was signed could not be verified because the STACK
207 order did not reflect the encoded order.
210 *) Reimplement the OCSP ASN1 module using the new code.
213 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
214 for its ASN1 operations. The old style function pointers still exist
215 for now but they will eventually go away.
218 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
219 completely replaces the old ASN1 functionality with a table driven
220 encoder and decoder which interprets an ASN1_ITEM structure describing
221 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
222 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
223 has also been converted to the new form.
226 *) Change BN_mod_exp_recp so that negative moduli are tolerated
227 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
228 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
232 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
233 of not touching the result's sign bit.
236 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
240 *) Changed the LHASH code to use prototypes for callbacks, and created
241 macros to declare and implement thin (optionally static) functions
242 that provide type-safety and avoid function pointer casting for the
243 type-specific callbacks.
246 *) Use better test patterns in bntest.
249 *) Added Kerberos Cipher Suites to be used with TLS, as written in
251 [Veers Staats <staatsvr@asc.hpc.mil>,
252 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
254 *) rand_win.c fix for Borland C.
257 *) BN_rshift bugfix for n == 0.
260 *) Reformat the FAQ so the different questions and answers can be divided
261 in sections depending on the subject.
264 *) Have the zlib compression code load ZLIB.DLL dynamically under
268 *) New function BN_mod_sqrt for computing square roots modulo a prime
269 (using the probabilistic Tonelli-Shanks algorithm unless
270 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
271 be handled deterministically).
272 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
274 *) Store verify_result within SSL_SESSION also for client side to
275 avoid potential security hole. (Re-used sessions on the client side
276 always resulted in verify_result==X509_V_OK, not using the original
277 result of the server certificate verification.)
280 *) Make BN_mod_inverse faster by explicitly handling small quotients
281 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
282 512 bits], about 30% for larger ones [1024 or 2048 bits].)
285 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
286 SSL3_RT_APPLICATION_DATA, return 0.
287 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
291 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
292 releases, have been re-implemented by renaming the previous
293 implementations of ssl2_read and ssl3_read to ssl2_read_internal
294 and ssl3_read_internal, respectively, and adding 'peek' parameters
295 to them. The new ssl[23]_{read,peek} functions are calls to
296 ssl[23]_read_internal with the 'peek' flag set appropriately.
297 A 'peek' parameter has also been added to ssl3_read_bytes, which
298 does the actual work for ssl3_read_internal.
301 *) New function BN_kronecker.
304 *) Fix BN_gcd so that it works on negative inputs; the result is
305 positive unless both parameters are zero.
306 Previously something reasonably close to an infinite loop was
307 possible because numbers could be growing instead of shrinking
308 in the implementation of Euclid's algorithm.
311 *) Fix BN_is_word() and BN_is_one() macros to take into account the
312 sign of the number in question.
314 Fix BN_is_word(a,w) to work correctly for w == 0.
316 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
317 because its test if the absolute value of 'a' equals 'w'.
318 Note that BN_abs_is_word does *not* handle w == 0 reliably;
319 it exists mostly for use in the implementations of BN_is_zero(),
320 BN_is_one(), and BN_is_word().
323 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
324 the method-specific "init()" handler. Also clean up ex_data after
325 calling the method-specific "finish()" handler. Previously, this was
326 happening the other way round.
329 *) New function BN_swap.
332 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
333 the exponentiation functions are more likely to produce reasonable
334 results on negative inputs.
337 *) Change BN_mod_mul so that the result is always non-negative.
338 Previously, it could be negative if one of the factors was negative;
339 I don't think anyone really wanted that behaviour.
342 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
343 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
344 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
345 and add new functions:
358 These functions always generate non-negative results.
360 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
361 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
363 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
364 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
366 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
368 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
369 was actually never needed) and in BN_mul(). The removal in BN_mul()
370 required a small change in bn_mul_part_recursive() and the addition
371 of the functions bn_cmp_part_words(), bn_sub_part_words() and
372 bn_add_part_words(), which do the same thing as bn_cmp_words(),
373 bn_sub_words() and bn_add_words() except they take arrays with
377 *) In 'openssl passwd', verify passwords read from the terminal
378 unless the '-salt' option is used (which usually means that
379 verification would just waste user's time since the resulting
380 hash is going to be compared with some given password hash)
381 or the new '-noverify' option is used.
383 This is an incompatible change, but it does not affect
384 non-interactive use of 'openssl passwd' (passwords on the command
385 line, '-stdin' option, '-in ...' option) and thus should not
389 *) Remove all references to RSAref, since there's no more need for it.
392 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
393 The previous value, 12, was not always sufficient for BN_mod_exp().
396 *) Make DSO load along a path given through an environment variable
397 (SHLIB_PATH) with shl_load().
400 *) Constify the ENGINE code as a result of BIGNUM constification.
401 Also constify the RSA code and most things related to it. In a
402 few places, most notable in the depth of the ASN.1 code, ugly
403 casts back to non-const were required (to be solved at a later
407 *) Make it so the openssl application has all engines loaded by default.
410 *) Constify the BIGNUM routines a little more.
413 *) Make sure that shared libraries get the internal name engine with
414 the full version number and not just 0. This should mark the
415 shared libraries as not backward compatible. Of course, this should
416 be changed again when we can guarantee backward binary compatibility.
419 *) Add the following functions:
425 ENGINE_load_builtin_engines()
427 That way, an application can itself choose if external engines that
428 are built-in in OpenSSL shall ever be used or not. The benefit is
429 that applications won't have to be linked with libdl or other dso
430 libraries unless it's really needed.
432 Changed 'openssl engine' to load all engines on demand.
433 Changed the engine header files to avoid the duplication of some
434 declarations (they differed!).
437 *) 'openssl engine' can now list capabilities.
440 *) Better error reporting in 'openssl engine'.
443 *) Never call load_dh_param(NULL) in s_server.
446 *) Add engine application. It can currently list engines by name and
447 identity, and test if they are actually available.
450 *) Add support for shared libraries under Irix.
451 [Albert Chin-A-Young <china@thewrittenword.com>]
453 *) Improve RPM specification file by forcing symbolic linking and making
454 sure the installed documentation is also owned by root.root.
455 [Damien Miller <djm@mindrot.org>]
457 *) Add configuration option to build on Linux on both big-endian and
459 [Ralf Baechle <ralf@uni-koblenz.de>]
461 *) Give the OpenSSL applications more possibilities to make use of
462 keys (public as well as private) handled by engines.
465 *) Add OCSP code that comes from CertCo.
468 *) Add VMS support for the Rijndael code.
471 *) Added untested support for Nuron crypto accelerator.
474 *) Add support for external cryptographic devices. This code was
475 previously distributed separately as the "engine" branch.
476 [Geoff Thorpe, Richard Levitte]
478 *) Rework the filename-translation in the DSO code. It is now possible to
479 have far greater control over how a "name" is turned into a filename
480 depending on the operating environment and any oddities about the
481 different shared library filenames on each system.
484 *) Support threads on FreeBSD-elf in Configure.
487 *) Add the possibility to create shared libraries on HP-UX
490 *) Fix for SHA1 assembly problem with MASM: it produces
491 warnings about corrupt line number information when assembling
492 with debugging information. This is caused by the overlapping
494 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
497 NCONF_get_number() has no error checking at all. As a replacement,
498 NCONF_get_number_e() is defined (_e for "error checking") and is
499 promoted strongly. The old NCONF_get_number is kept around for
500 binary backward compatibility.
501 Make it possible for methods to load from something other than a BIO,
502 by providing a function pointer that is given a name instead of a BIO.
503 For example, this could be used to load configuration data from an
507 *) Fix typo in get_cert_by_subject() in by_dir.c
508 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
510 *) Rework the system to generate shared libraries:
512 - Make note of the expected extension for the shared libraries and
513 if there is a need for symbolic links from for example libcrypto.so.0
514 to libcrypto.so.0.9.7. There is extended info in Configure for
517 - Make as few rebuilds of the shared libraries as possible.
519 - Still avoid linking the OpenSSL programs with the shared libraries.
521 - When installing, install the shared libraries separately from the
525 *) Fix for non blocking accept BIOs. Added new I/O special reason
526 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
527 with non blocking I/O was not possible because no retry code was
528 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
532 *) Added the beginnings of Rijndael support.
535 *) Fix for bug in DirectoryString mask setting. Add support for
536 X509_NAME_print_ex() in 'req' and X509_print_ex() function
537 to allow certificate printing to more controllable, additional
538 'certopt' option to 'x509' to allow new printing options to be
542 *) Clean old EAY MD5 hack from e_os.h.
545 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
547 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
548 and not in SSL_clear because the latter is also used by the
549 accept/connect functions; previously, the settings made by
550 SSL_set_read_ahead would be lost during the handshake.
551 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
553 *) Correct util/mkdef.pl to be selective about disabled algorithms.
554 Previously, it would create entries for disableed algorithms no
558 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
560 *) In ssl23_get_client_hello, generate an error message when faced
561 with an initial SSL 3.0/TLS record that is too small to contain the
562 first two bytes of the ClientHello message, i.e. client_version.
563 (Note that this is a pathologic case that probably has never happened
564 in real life.) The previous approach was to use the version number
565 from the record header as a substitute; but our protocol choice
566 should not depend on that one because it is not authenticated
567 by the Finished messages.
570 *) More robust randomness gathering functions for Windows.
571 [Jeffrey Altman <jaltman@columbia.edu>]
573 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
574 not set then we don't setup the error code for issuer check errors
575 to avoid possibly overwriting other errors which the callback does
576 handle. If an application does set the flag then we assume it knows
577 what it is doing and can handle the new informational codes
581 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
582 a general "ANY" type, as such it should be able to decode anything
583 including tagged types. However it didn't check the class so it would
584 wrongly interpret tagged types in the same way as their universal
585 counterpart and unknown types were just rejected. Changed so that the
586 tagged and unknown types are handled in the same way as a SEQUENCE:
587 that is the encoding is stored intact. There is also a new type
588 "V_ASN1_OTHER" which is used when the class is not universal, in this
589 case we have no idea what the actual type is so we just lump them all
593 *) On VMS, stdout may very well lead to a file that is written to
594 in a record-oriented fashion. That means that every write() will
595 write a separate record, which will be read separately by the
596 programs trying to read from it. This can be very confusing.
598 The solution is to put a BIO filter in the way that will buffer
599 text until a linefeed is reached, and then write everything a
600 line at a time, so every record written will be an actual line,
601 not chunks of lines and not (usually doesn't happen, but I've
602 seen it once) several lines in one record. BIO_f_linebuffer() is
605 Currently, it's a VMS-only method, because that's where it has
606 been tested well enough.
609 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
610 it can return incorrect results.
611 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
612 but it was in 0.9.6-beta[12].)
615 *) Disable the check for content being present when verifying detached
616 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
617 include zero length content when signing messages.
620 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
621 BIO_ctrl (for BIO pairs).
624 *) Add DSO method for VMS.
627 *) Bug fix: Montgomery multiplication could produce results with the
631 *) Add RPM specification openssl.spec and modify it to build three
632 packages. The default package contains applications, application
633 documentation and run-time libraries. The devel package contains
634 include files, static libraries and function documentation. The
635 doc package contains the contents of the doc directory. The original
636 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
639 *) Add a large number of documentation files for many SSL routines.
640 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
642 *) Add a configuration entry for Sony News 4.
643 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
645 *) Don't set the two most significant bits to one when generating a
646 random number < q in the DSA library.
649 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
650 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
651 the underlying transport is blocking) if a handshake took place.
652 (The default behaviour is needed by applications such as s_client
653 and s_server that use select() to determine when to use SSL_read;
654 but for applications that know in advance when to expect data, it
655 just makes things more complicated.)
658 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
662 *) Add a few more EBCDIC conditionals that make `req' and `x509'
663 work better on such systems.
664 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
666 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
667 Update PKCS12_parse() so it copies the friendlyName and the
668 keyid to the certificates aux info.
671 *) Fix bug in PKCS7_verify() which caused an infinite loop
672 if there was more than one signature.
673 [Sven Uszpelkat <su@celocom.de>]
675 *) Major change in util/mkdef.pl to include extra information
676 about each symbol, as well as presentig variables as well
677 as functions. This change means that there's n more need
678 to rebuild the .num files when some algorithms are excluded.
681 *) Allow the verify time to be set by an application,
682 rather than always using the current time.
685 *) Phase 2 verify code reorganisation. The certificate
686 verify code now looks up an issuer certificate by a
687 number of criteria: subject name, authority key id
688 and key usage. It also verifies self signed certificates
689 by the same criteria. The main comparison function is
690 X509_check_issued() which performs these checks.
692 Lot of changes were necessary in order to support this
693 without completely rewriting the lookup code.
695 Authority and subject key identifier are now cached.
697 The LHASH 'certs' is X509_STORE has now been replaced
698 by a STACK_OF(X509_OBJECT). This is mainly because an
699 LHASH can't store or retrieve multiple objects with
702 As a result various functions (which were all internal
703 use only) have changed to handle the new X509_STORE
704 structure. This will break anything that messed round
705 with X509_STORE internally.
707 The functions X509_STORE_add_cert() now checks for an
708 exact match, rather than just subject name.
710 The X509_STORE API doesn't directly support the retrieval
711 of multiple certificates matching a given criteria, however
712 this can be worked round by performing a lookup first
713 (which will fill the cache with candidate certificates)
714 and then examining the cache for matches. This is probably
715 the best we can do without throwing out X509_LOOKUP
716 entirely (maybe later...).
718 The X509_VERIFY_CTX structure has been enhanced considerably.
720 All certificate lookup operations now go via a get_issuer()
721 callback. Although this currently uses an X509_STORE it
722 can be replaced by custom lookups. This is a simple way
723 to bypass the X509_STORE hackery necessary to make this
724 work and makes it possible to use more efficient techniques
725 in future. A very simple version which uses a simple
726 STACK for its trusted certificate store is also provided
727 using X509_STORE_CTX_trusted_stack().
729 The verify_cb() and verify() callbacks now have equivalents
730 in the X509_STORE_CTX structure.
732 X509_STORE_CTX also has a 'flags' field which can be used
733 to customise the verify behaviour.
736 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
737 excludes S/MIME capabilities.
740 *) When a certificate request is read in keep a copy of the
741 original encoding of the signed data and use it when outputing
742 again. Signatures then use the original encoding rather than
743 a decoded, encoded version which may cause problems if the
744 request is improperly encoded.
747 *) For consistency with other BIO_puts implementations, call
748 buffer_write(b, ...) directly in buffer_puts instead of calling
751 In BIO_puts, increment b->num_write as in BIO_write.
752 [Peter.Sylvester@EdelWeb.fr]
754 *) Fix BN_mul_word for the case where the word is 0. (We have to use
755 BN_zero, we may not return a BIGNUM with an array consisting of
759 *) Avoid calling abort() from within the library when problems are
760 detected, except if preprocessor symbols have been defined
761 (such as REF_CHECK, BN_DEBUG etc.).
764 *) New openssl application 'rsautl'. This utility can be
765 used for low level RSA operations. DER public key
766 BIO/fp routines also added.
769 *) New Configure entry and patches for compiling on QNX 4.
770 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
772 *) A demo state-machine implementation was sponsored by
773 Nuron (http://www.nuron.com/) and is now available in
777 *) New options added to the 'dgst' utility for signature
778 generation and verification.
781 *) Unrecognized PKCS#7 content types are now handled via a
782 catch all ASN1_TYPE structure. This allows unsupported
783 types to be stored as a "blob" and an application can
784 encode and decode it manually.
787 *) Fix various signed/unsigned issues to make a_strex.c
789 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
791 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
792 length if passed a buffer. ASN1_INTEGER_to_BN failed
793 if passed a NULL BN and its argument was negative.
794 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
796 *) Modification to PKCS#7 encoding routines to output definite
797 length encoding. Since currently the whole structures are in
798 memory there's not real point in using indefinite length
799 constructed encoding. However if OpenSSL is compiled with
800 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
803 *) Added BIO_vprintf() and BIO_vsnprintf().
806 *) Added more prefixes to parse for in the the strings written
807 through a logging bio, to cover all the levels that are available
808 through syslog. The prefixes are now:
810 PANIC, EMERG, EMR => LOG_EMERG
811 ALERT, ALR => LOG_ALERT
812 CRIT, CRI => LOG_CRIT
813 ERROR, ERR => LOG_ERR
814 WARNING, WARN, WAR => LOG_WARNING
815 NOTICE, NOTE, NOT => LOG_NOTICE
816 INFO, INF => LOG_INFO
817 DEBUG, DBG => LOG_DEBUG
819 and as before, if none of those prefixes are present at the
820 beginning of the string, LOG_ERR is chosen.
822 On Win32, the LOG_* levels are mapped according to this:
824 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
825 LOG_WARNING => EVENTLOG_WARNING_TYPE
826 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
830 *) Made it possible to reconfigure with just the configuration
831 argument "reconf" or "reconfigure". The command line arguments
832 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
833 and are retrieved from there when reconfiguring.
837 [Assar Westerlund <assar@sics.se>, Richard Levitte]
839 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
842 *) The obj_dat.pl script was messing up the sorting of object
843 names. The reason was that it compared the quoted version
844 of strings as a result "OCSP" > "OCSP Signing" because
845 " > SPACE. Changed script to store unquoted versions of
846 names and add quotes on output. It was also omitting some
847 names from the lookup table if they were given a default
848 value (that is if SN is missing it is given the same
849 value as LN and vice versa), these are now added on the
850 grounds that if an object has a name we should be able to
851 look it up. Finally added warning output when duplicate
852 short or long names are found.
855 *) Changes needed for Tandem NSK.
856 [Scott Uroff <scott@xypro.com>]
858 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
859 RSA_padding_check_SSLv23(), special padding was never detected
860 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
861 version rollback attacks was not effective.
863 In s23_clnt.c, don't use special rollback-attack detection padding
864 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
865 client; similarly, in s23_srvr.c, don't do the rollback check if
866 SSL 2.0 is the only protocol enabled in the server.
869 *) Make it possible to get hexdumps of unprintable data with 'openssl
870 asn1parse'. By implication, the functions ASN1_parse_dump() and
871 BIO_dump_indent() are added.
874 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
875 these print out strings and name structures based on various
876 flags including RFC2253 support and proper handling of
877 multibyte characters. Added options to the 'x509' utility
878 to allow the various flags to be set.
881 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
882 Also change the functions X509_cmp_current_time() and
883 X509_gmtime_adj() work with an ASN1_TIME structure,
884 this will enable certificates using GeneralizedTime in validity
888 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
889 negative public key encodings) on by default,
890 NO_NEG_PUBKEY_BUG can be set to disable it.
893 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
894 content octets. An i2c_ASN1_OBJECT is unnecessary because
895 the encoding can be trivially obtained from the structure.
898 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
899 not read locks (CRYPTO_r_[un]lock).
902 *) A first attempt at creating official support for shared
903 libraries through configuration. I've kept it so the
904 default is static libraries only, and the OpenSSL programs
905 are always statically linked for now, but there are
906 preparations for dynamic linking in place.
907 This has been tested on Linux and True64.
910 *) Randomness polling function for Win9x, as described in:
911 Peter Gutmann, Software Generation of Practically Strong
915 *) Fix so PRNG is seeded in req if using an already existing
919 *) New options to smime application. -inform and -outform
920 allow alternative formats for the S/MIME message including
921 PEM and DER. The -content option allows the content to be
922 specified separately. This should allow things like Netscape
923 form signing output easier to verify.
926 *) Fix the ASN1 encoding of tags using the 'long form'.
929 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
930 STRING types. These convert content octets to and from the
931 underlying type. The actual tag and length octets are
932 already assumed to have been read in and checked. These
933 are needed because all other string types have virtually
934 identical handling apart from the tag. By having versions
935 of the ASN1 functions that just operate on content octets
936 IMPLICIT tagging can be handled properly. It also allows
937 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
938 and ASN1_INTEGER are identical apart from the tag.
941 *) Change the handling of OID objects as follows:
943 - New object identifiers are inserted in objects.txt, following
944 the syntax given in objects.README.
945 - objects.pl is used to process obj_mac.num and create a new
947 - obj_dat.pl is used to create a new obj_dat.h, using the data in
950 This is currently kind of a hack, and the perl code in objects.pl
951 isn't very elegant, but it works as I intended. The simplest way
952 to check that it worked correctly is to look in obj_dat.h and
953 check the array nid_objs and make sure the objects haven't moved
954 around (this is important!). Additions are OK, as well as
955 consistent name changes.
958 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
961 *) Addition of the command line parameter '-rand file' to 'openssl req'.
962 The given file adds to whatever has already been seeded into the
963 random pool through the RANDFILE configuration file option or
964 environment variable, or the default random state file.
967 *) mkstack.pl now sorts each macro group into lexical order.
968 Previously the output order depended on the order the files
969 appeared in the directory, resulting in needless rewriting
973 *) Patches to make OpenSSL compile under Win32 again. Mostly
974 work arounds for the VC++ problem that it treats func() as
975 func(void). Also stripped out the parts of mkdef.pl that
976 added extra typesafe functions: these no longer exist.
979 *) Reorganisation of the stack code. The macros are now all
980 collected in safestack.h . Each macro is defined in terms of
981 a "stack macro" of the form SKM_<name>(type, a, b). The
982 DEBUG_SAFESTACK is now handled in terms of function casts,
983 this has the advantage of retaining type safety without the
984 use of additional functions. If DEBUG_SAFESTACK is not defined
985 then the non typesafe macros are used instead. Also modified the
986 mkstack.pl script to handle the new form. Needs testing to see
987 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
988 the default if no major problems. Similar behaviour for ASN1_SET_OF
992 *) When some versions of IIS use the 'NET' form of private key the
993 key derivation algorithm is different. Normally MD5(password) is
994 used as a 128 bit RC4 key. In the modified case
995 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
996 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
997 as the old Netscape_RSA functions except they have an additional
998 'sgckey' parameter which uses the modified algorithm. Also added
999 an -sgckey command line option to the rsa utility. Thanks to
1000 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1001 algorithm to openssl-dev.
1004 *) The evp_local.h macros were using 'c.##kname' which resulted in
1005 invalid expansion on some systems (SCO 5.0.5 for example).
1006 Corrected to 'c.kname'.
1007 [Phillip Porch <root@theporch.com>]
1009 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1010 a STACK of email addresses from a certificate or request, these look
1011 in the subject name and the subject alternative name extensions and
1012 omit any duplicate addresses.
1015 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1016 This makes DSA verification about 2 % faster.
1019 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1020 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1021 plus overhead for 1024 bit moduli).
1022 This makes exponentiations about 0.5 % faster for 1024 bit
1023 exponents (as measured by "openssl speed rsa2048").
1026 *) Rename memory handling macros to avoid conflicts with other
1028 Malloc => OPENSSL_malloc
1029 Malloc_locked => OPENSSL_malloc_locked
1030 Realloc => OPENSSL_realloc
1031 Free => OPENSSL_free
1034 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1035 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1038 *) CygWin32 support.
1039 [John Jarvie <jjarvie@newsguy.com>]
1041 *) The type-safe stack code has been rejigged. It is now only compiled
1042 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1043 by default all type-specific stack functions are "#define"d back to
1044 standard stack functions. This results in more streamlined output
1045 but retains the type-safety checking possibilities of the original
1049 *) The STACK code has been cleaned up, and certain type declarations
1050 that didn't make a lot of sense have been brought in line. This has
1051 also involved a cleanup of sorts in safestack.h to more correctly
1052 map type-safe stack functions onto their plain stack counterparts.
1053 This work has also resulted in a variety of "const"ifications of
1054 lots of the code, especially "_cmp" operations which should normally
1055 be prototyped with "const" parameters anyway.
1058 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1059 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1060 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1061 where all of 'md' is used each time the PRNG is used, but 'state'
1062 is used only indexed by a cyclic counter. As entropy may not be
1063 well distributed from the beginning, 'md' is important as a
1064 chaining variable. However, the output function chains only half
1065 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1066 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1067 in all of 'state' being rewritten, with the new values depending
1068 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1071 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1072 the handshake is continued after ssl_verify_cert_chain();
1073 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1074 can lead to 'unexplainable' connection aborts later.
1075 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1077 *) Major EVP API cipher revision.
1078 Add hooks for extra EVP features. This allows various cipher
1079 parameters to be set in the EVP interface. Support added for variable
1080 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1081 setting of RC2 and RC5 parameters.
1083 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1086 Remove lots of duplicated code from the EVP library. For example *every*
1087 cipher init() function handles the 'iv' in the same way according to the
1088 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1089 for CFB and OFB modes they zero ctx->num.
1091 New functionality allows removal of S/MIME code RC2 hack.
1093 Most of the routines have the same form and so can be declared in terms
1096 By shifting this to the top level EVP_CipherInit() it can be removed from
1097 all individual ciphers. If the cipher wants to handle IVs or keys
1098 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1101 Change lots of functions like EVP_EncryptUpdate() to now return a
1102 value: although software versions of the algorithms cannot fail
1103 any installed hardware versions can.
1106 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1107 this option is set, tolerate broken clients that send the negotiated
1108 protocol version number instead of the requested protocol version
1112 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1113 i.e. non-zero for export ciphersuites, zero otherwise.
1114 Previous versions had this flag inverted, inconsistent with
1115 rsa_tmp_cb (..._TMP_RSA_CB).
1116 [Bodo Moeller; problem reported by Amit Chopra]
1118 *) Add missing DSA library text string. Work around for some IIS
1119 key files with invalid SEQUENCE encoding.
1122 *) Add a document (doc/standards.txt) that list all kinds of standards
1123 and so on that are implemented in OpenSSL.
1126 *) Enhance c_rehash script. Old version would mishandle certificates
1127 with the same subject name hash and wouldn't handle CRLs at all.
1128 Added -fingerprint option to crl utility, to support new c_rehash
1132 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1135 *) Fix for SSL server purpose checking. Server checking was
1136 rejecting certificates which had extended key usage present
1137 but no ssl client purpose.
1138 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1140 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1141 is a little unclear about how a blank password is handled.
1142 Since the password in encoded as a BMPString with terminating
1143 double NULL a zero length password would end up as just the
1144 double NULL. However no password at all is different and is
1145 handled differently in the PKCS#12 key generation code. NS
1146 treats a blank password as zero length. MSIE treats it as no
1147 password on export: but it will try both on import. We now do
1148 the same: PKCS12_parse() tries zero length and no password if
1149 the password is set to "" or NULL (NULL is now a valid password:
1150 it wasn't before) as does the pkcs12 application.
1153 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1154 perror when PEM_read_bio_X509_REQ fails, the error message must
1155 be obtained from the error queue.
1158 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1159 it in ERR_remove_state if appropriate, and change ERR_get_state
1160 accordingly to avoid race conditions (this is necessary because
1161 thread_hash is no longer constant once set).
1164 *) Bugfix for linux-elf makefile.one.
1167 *) RSA_get_default_method() will now cause a default
1168 RSA_METHOD to be chosen if one doesn't exist already.
1169 Previously this was only set during a call to RSA_new()
1170 or RSA_new_method(NULL) meaning it was possible for
1171 RSA_get_default_method() to return NULL.
1174 *) Added native name translation to the existing DSO code
1175 that will convert (if the flag to do so is set) filenames
1176 that are sufficiently small and have no path information
1177 into a canonical native form. Eg. "blah" converted to
1178 "libblah.so" or "blah.dll" etc.
1181 *) New function ERR_error_string_n(e, buf, len) which is like
1182 ERR_error_string(e, buf), but writes at most 'len' bytes
1183 including the 0 terminator. For ERR_error_string_n, 'buf'
1185 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1187 *) CONF library reworked to become more general. A new CONF
1188 configuration file reader "class" is implemented as well as a
1189 new functions (NCONF_*, for "New CONF") to handle it. The now
1190 old CONF_* functions are still there, but are reimplemented to
1191 work in terms of the new functions. Also, a set of functions
1192 to handle the internal storage of the configuration data is
1193 provided to make it easier to write new configuration file
1194 reader "classes" (I can definitely see something reading a
1195 configuration file in XML format, for example), called _CONF_*,
1196 or "the configuration storage API"...
1198 The new configuration file reading functions are:
1200 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1201 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1203 NCONF_default, NCONF_WIN32
1205 NCONF_dump_fp, NCONF_dump_bio
1207 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1208 NCONF_new creates a new CONF object. This works in the same way
1209 as other interfaces in OpenSSL, like the BIO interface.
1210 NCONF_dump_* dump the internal storage of the configuration file,
1211 which is useful for debugging. All other functions take the same
1212 arguments as the old CONF_* functions wth the exception of the
1213 first that must be a `CONF *' instead of a `LHASH *'.
1215 To make it easer to use the new classes with the old CONF_* functions,
1216 the function CONF_set_default_method is provided.
1219 *) Add '-tls1' option to 'openssl ciphers', which was already
1220 mentioned in the documentation but had not been implemented.
1221 (This option is not yet really useful because even the additional
1222 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1225 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1226 OpenSSL-based applications) load shared libraries and bind to
1227 them in a portable way.
1228 [Geoff Thorpe, with contributions from Richard Levitte]
1230 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1232 *) Make sure _lrotl and _lrotr are only used with MSVC.
1234 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1235 (the default implementation of RAND_status).
1237 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1238 to '-clrext' (= clear extensions), as intended and documented.
1239 [Bodo Moeller; inconsistency pointed out by Michael Attili
1242 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1243 was larger than the MD block size.
1244 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1246 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1247 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1248 using the passed key: if the passed key was a private key the result
1249 of X509_print(), for example, would be to print out all the private key
1253 *) des_quad_cksum() byte order bug fix.
1254 [Ulf Möller, using the problem description in krb4-0.9.7, where
1255 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1257 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1259 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1261 *) For easily testing in shell scripts whether some command
1262 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1263 returns with exit code 0 iff no command of the given name is available.
1264 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1265 the output goes to stdout and nothing is printed to stderr.
1266 Additional arguments are always ignored.
1268 Since for each cipher there is a command of the same name,
1269 the 'no-cipher' compilation switches can be tested this way.
1271 ('openssl no-XXX' is not able to detect pseudo-commands such
1272 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1275 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1278 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1279 is set; it will be thrown away anyway because each handshake creates
1281 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1282 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1283 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1284 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1287 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1288 'Q' and 'R' lose their special meanings (quit/renegotiate).
1289 This is part of what -quiet does; unlike -quiet, -ign_eof
1290 does not suppress any output.
1293 *) Add compatibility options to the purpose and trust code. The
1294 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1295 accepts a certificate or CA, this was the previous behaviour,
1296 with all the associated security issues.
1298 X509_TRUST_COMPAT is the old trust behaviour: only and
1299 automatically trust self signed roots in certificate store. A
1300 new trust setting X509_TRUST_DEFAULT is used to specify that
1301 a purpose has no associated trust setting and it should instead
1302 use the value in the default purpose.
1305 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1306 and fix a memory leak.
1309 *) In util/mkerr.pl (which implements 'make errors'), preserve
1310 reason strings from the previous version of the .c file, as
1311 the default to have only downcase letters (and digits) in
1312 automatically generated reasons codes is not always appropriate.
1315 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1316 using strerror. Previously, ERR_reason_error_string() returned
1317 library names as reason strings for SYSerr; but SYSerr is a special
1318 case where small numbers are errno values, not library numbers.
1321 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1322 converts DSA parameters into DH parameters. (When creating parameters,
1323 DSA_generate_parameters is used.)
1326 *) Include 'length' (recommended exponent length) in C code generated
1327 by 'openssl dhparam -C'.
1330 *) The second argument to set_label in perlasm was already being used
1331 so couldn't be used as a "file scope" flag. Moved to third argument
1335 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1336 instead of RAND_bytes for encryption IVs and salts.
1339 *) Include RAND_status() into RAND_METHOD instead of implementing
1340 it only for md_rand.c Otherwise replacing the PRNG by calling
1341 RAND_set_rand_method would be impossible.
1344 *) Don't let DSA_generate_key() enter an infinite loop if the random
1345 number generation fails.
1348 *) New 'rand' application for creating pseudo-random output.
1351 *) Added configuration support for Linux/IA64
1352 [Rolf Haberrecker <rolf@suse.de>]
1354 *) Assembler module support for Mingw32.
1357 *) Shared library support for HPUX (in shlib/).
1358 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1360 *) Shared library support for Solaris gcc.
1361 [Lutz Behnke <behnke@trustcenter.de>]
1363 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1365 *) PKCS7_encrypt() was adding text MIME headers twice because they
1366 were added manually and by SMIME_crlf_copy().
1369 *) In bntest.c don't call BN_rand with zero bits argument.
1370 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1372 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1373 case was implemented. This caused BN_div_recp() to fail occasionally.
1376 *) Add an optional second argument to the set_label() in the perl
1377 assembly language builder. If this argument exists and is set
1378 to 1 it signals that the assembler should use a symbol whose
1379 scope is the entire file, not just the current function. This
1380 is needed with MASM which uses the format label:: for this scope.
1381 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1383 *) Change the ASN1 types so they are typedefs by default. Before
1384 almost all types were #define'd to ASN1_STRING which was causing
1385 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1389 *) Change names of new functions to the new get1/get0 naming
1390 convention: After 'get1', the caller owns a reference count
1391 and has to call ..._free; 'get0' returns a pointer to some
1392 data structure without incrementing reference counters.
1393 (Some of the existing 'get' functions increment a reference
1394 counter, some don't.)
1395 Similarly, 'set1' and 'add1' functions increase reference
1396 counters or duplicate objects.
1399 *) Allow for the possibility of temp RSA key generation failure:
1400 the code used to assume it always worked and crashed on failure.
1403 *) Fix potential buffer overrun problem in BIO_printf().
1404 [Ulf Möller, using public domain code by Patrick Powell; problem
1405 pointed out by David Sacerdote <das33@cornell.edu>]
1407 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1408 RAND_egd() and RAND_status(). In the command line application,
1409 the EGD socket can be specified like a seed file using RANDFILE
1413 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1414 Some CAs (e.g. Verisign) distribute certificates in this form.
1417 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1418 list to exclude them. This means that no special compilation option
1419 is needed to use anonymous DH: it just needs to be included in the
1423 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1424 EVP_MD_type. The old functionality is available in a new macro called
1425 EVP_MD_md(). Change code that uses it and update docs.
1428 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1429 where the 'void *' argument is replaced by a function pointer argument.
1430 Previously 'void *' was abused to point to functions, which works on
1431 many platforms, but is not correct. As these functions are usually
1432 called by macros defined in OpenSSL header files, most source code
1433 should work without changes.
1436 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1437 sections with information on -D... compiler switches used for
1438 compiling the library so that applications can see them. To enable
1439 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1440 must be defined. E.g.,
1441 #define OPENSSL_ALGORITHM_DEFINES
1442 #include <openssl/opensslconf.h>
1443 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1444 [Richard Levitte, Ulf and Bodo Möller]
1446 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1450 *) Change the 'other' type in certificate aux info to a STACK_OF
1451 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1452 the required ASN1 format: arbitrary types determined by an OID.
1455 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1456 argument to 'req'. This is not because the function is newer or
1457 better than others it just uses the work 'NEW' in the certificate
1458 request header lines. Some software needs this.
1461 *) Reorganise password command line arguments: now passwords can be
1462 obtained from various sources. Delete the PEM_cb function and make
1463 it the default behaviour: i.e. if the callback is NULL and the
1464 usrdata argument is not NULL interpret it as a null terminated pass
1465 phrase. If usrdata and the callback are NULL then the pass phrase
1466 is prompted for as usual.
1469 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1470 the support is automatically enabled. The resulting binaries will
1471 autodetect the card and use it if present.
1472 [Ben Laurie and Compaq Inc.]
1474 *) Work around for Netscape hang bug. This sends certificate request
1475 and server done in one record. Since this is perfectly legal in the
1476 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1477 the bugs/SSLv3 entry for more info.
1480 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1483 *) Add -rand argument to smime and pkcs12 applications and read/write
1487 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1490 *) Add command line password options to the remaining applications.
1493 *) Bug fix for BN_div_recp() for numerators with an even number of
1497 *) More tests in bntest.c, and changed test_bn output.
1500 *) ./config recognizes MacOS X now.
1503 *) Bug fix for BN_div() when the first words of num and divsor are
1504 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1507 *) Add support for various broken PKCS#8 formats, and command line
1508 options to produce them.
1511 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1512 get temporary BIGNUMs from a BN_CTX.
1515 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1519 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1520 include a #define from the old name to the new. The original intent
1521 was that statically linked binaries could for example just call
1522 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1523 link with digests. This never worked becayse SSLeay_add_all_digests()
1524 and SSLeay_add_all_ciphers() were in the same source file so calling
1525 one would link with the other. They are now in separate source files.
1528 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1531 *) Use a less unusual form of the Miller-Rabin primality test (it used
1532 a binary algorithm for exponentiation integrated into the Miller-Rabin
1533 loop, our standard modexp algorithms are faster).
1536 *) Support for the EBCDIC character set completed.
1537 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1539 *) Source code cleanups: use const where appropriate, eliminate casts,
1540 use void * instead of char * in lhash.
1543 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1544 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1545 this the server could overwrite ephemeral keys that the client
1549 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1550 using 50 iterations of the Rabin-Miller test.
1552 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1553 iterations of the Rabin-Miller test as required by the appendix
1554 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1555 As BN_is_prime_fasttest includes trial division, DSA parameter
1556 generation becomes much faster.
1558 This implies a change for the callback functions in DSA_is_prime
1559 and DSA_generate_parameters: The callback function is called once
1560 for each positive witness in the Rabin-Miller test, not just
1561 occasionally in the inner loop; and the parameters to the
1562 callback function now provide an iteration count for the outer
1563 loop rather than for the current invocation of the inner loop.
1564 DSA_generate_parameters additionally can call the callback
1565 function with an 'iteration count' of -1, meaning that a
1566 candidate has passed the trial division test (when q is generated
1567 from an application-provided seed, trial division is skipped).
1570 *) New function BN_is_prime_fasttest that optionally does trial
1571 division before starting the Rabin-Miller test and has
1572 an additional BN_CTX * argument (whereas BN_is_prime always
1573 has to allocate at least one BN_CTX).
1574 'callback(1, -1, cb_arg)' is called when a number has passed the
1575 trial division stage.
1578 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1582 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1585 *) New function BN_pseudo_rand().
1588 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1589 bignum version of BN_from_montgomery() with the working code from
1590 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1594 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1595 made it impossible to use the same SSL_SESSION data structure in
1596 SSL2 clients in multiple threads.
1599 *) The return value of RAND_load_file() no longer counts bytes obtained
1600 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1601 to seed the PRNG (previously an explicit byte count was required).
1602 [Ulf Möller, Bodo Möller]
1604 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1605 used (char *) instead of (void *) and had casts all over the place.
1608 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1611 *) Retain source code compatibility for BN_prime_checks macro:
1612 BN_is_prime(..., BN_prime_checks, ...) now uses
1613 BN_prime_checks_for_size to determine the appropriate number of
1614 Rabin-Miller iterations.
1617 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1618 DH_CHECK_P_NOT_SAFE_PRIME.
1619 (Check if this is true? OpenPGP calls them "strong".)
1622 *) Merge the functionality of "dh" and "gendh" programs into a new program
1623 "dhparam". The old programs are retained for now but will handle DH keys
1624 (instead of parameters) in future.
1627 *) Make the ciphers, s_server and s_client programs check the return values
1628 when a new cipher list is set.
1631 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1632 ciphers. Before when the 56bit ciphers were enabled the sorting was
1635 The syntax for the cipher sorting has been extended to support sorting by
1636 cipher-strength (using the strength_bits hard coded in the tables).
1637 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1639 Fix a bug in the cipher-command parser: when supplying a cipher command
1640 string with an "undefined" symbol (neither command nor alphanumeric
1641 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1642 an error is flagged.
1644 Due to the strength-sorting extension, the code of the
1645 ssl_create_cipher_list() function was completely rearranged. I hope that
1646 the readability was also increased :-)
1647 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1649 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1650 for the first serial number and places 2 in the serial number file. This
1651 avoids problems when the root CA is created with serial number zero and
1652 the first user certificate has the same issuer name and serial number
1656 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1657 the new code. Add documentation for this stuff.
1660 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1661 X509_*() to X509at_*() on the grounds that they don't handle X509
1662 structures and behave in an analagous way to the X509v3 functions:
1663 they shouldn't be called directly but wrapper functions should be used
1666 So we also now have some wrapper functions that call the X509at functions
1667 when passed certificate requests. (TO DO: similar things can be done with
1668 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1669 things. Some of these need some d2i or i2d and print functionality
1670 because they handle more complex structures.)
1673 *) Add missing #ifndefs that caused missing symbols when building libssl
1674 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1675 NO_RSA in ssl/s2*.c.
1676 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1678 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1679 has a return value which indicates the quality of the random data
1680 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1681 error queue. New function RAND_pseudo_bytes() generates output that is
1682 guaranteed to be unique but not unpredictable. RAND_add is like
1683 RAND_seed, but takes an extra argument for an entropy estimate
1684 (RAND_seed always assumes full entropy).
1687 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1688 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1689 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1690 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1691 false-positive rate of at most 2^-80 for random input.
1694 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1697 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1698 in the 0.9.5 release), this returns the chain
1699 from an X509_CTX structure with a dup of the stack and all
1700 the X509 reference counts upped: so the stack will exist
1701 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1704 Also make SSL_SESSION_print() print out the verify return
1708 *) Add manpage for the pkcs12 command. Also change the default
1709 behaviour so MAC iteration counts are used unless the new
1710 -nomaciter option is used. This improves file security and
1711 only older versions of MSIE (4.0 for example) need it.
1714 *) Honor the no-xxx Configure options when creating .DEF files.
1717 *) Add PKCS#10 attributes to field table: challengePassword,
1718 unstructuredName and unstructuredAddress. These are taken from
1719 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1720 international characters are used.
1722 More changes to X509_ATTRIBUTE code: allow the setting of types
1723 based on strings. Remove the 'loc' parameter when adding
1724 attributes because these will be a SET OF encoding which is sorted
1728 *) Initial changes to the 'req' utility to allow request generation
1729 automation. This will allow an application to just generate a template
1730 file containing all the field values and have req construct the
1733 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1734 used all over the place including certificate requests and PKCS#7
1735 structures. They are currently handled manually where necessary with
1736 some primitive wrappers for PKCS#7. The new functions behave in a
1737 manner analogous to the X509 extension functions: they allow
1738 attributes to be looked up by NID and added.
1740 Later something similar to the X509V3 code would be desirable to
1741 automatically handle the encoding, decoding and printing of the
1742 more complex types. The string types like challengePassword can
1743 be handled by the string table functions.
1745 Also modified the multi byte string table handling. Now there is
1746 a 'global mask' which masks out certain types. The table itself
1747 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1748 is useful when for example there is only one permissible type
1749 (as in countryName) and using the mask might result in no valid
1753 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1754 SSL_get_peer_finished to allow applications to obtain the latest
1755 Finished messages sent to the peer or expected from the peer,
1756 respectively. (SSL_get_peer_finished is usually the Finished message
1757 actually received from the peer, otherwise the protocol will be aborted.)
1759 As the Finished message are message digests of the complete handshake
1760 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1761 be used for external authentication procedures when the authentication
1762 provided by SSL/TLS is not desired or is not enough.
1765 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1766 the host supports BWX extension and if Compaq C is present on the
1767 $PATH. Just exploiting of the BWX extension results in 20-30%
1768 performance kick for some algorithms, e.g. DES and RC4 to mention
1769 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1773 *) Add support for MS "fast SGC". This is arguably a violation of the
1774 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1775 weak crypto and after checking the certificate is SGC a second one
1776 with strong crypto. MS SGC stops the first handshake after receiving
1777 the server certificate message and sends a second client hello. Since
1778 a server will typically do all the time consuming operations before
1779 expecting any further messages from the client (server key exchange
1780 is the most expensive) there is little difference between the two.
1782 To get OpenSSL to support MS SGC we have to permit a second client
1783 hello message after we have sent server done. In addition we have to
1784 reset the MAC if we do get this second client hello.
1787 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1788 if a DER encoded private key is RSA or DSA traditional format. Changed
1789 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1790 format DER encoded private key. Newer code should use PKCS#8 format which
1791 has the key type encoded in the ASN1 structure. Added DER private key
1792 support to pkcs8 application.
1795 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1796 ciphersuites has been selected (as required by the SSL 3/TLS 1
1797 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1798 is set, we interpret this as a request to violate the specification
1799 (the worst that can happen is a handshake failure, and 'correct'
1800 behaviour would result in a handshake failure anyway).
1803 *) In SSL_CTX_add_session, take into account that there might be multiple
1804 SSL_SESSION structures with the same session ID (e.g. when two threads
1805 concurrently obtain them from an external cache).
1806 The internal cache can handle only one SSL_SESSION with a given ID,
1807 so if there's a conflict, we now throw out the old one to achieve
1811 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1812 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1813 some routines that use cipher OIDs: some ciphers do not have OIDs
1814 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1818 *) Simplify the trust setting structure and code. Now we just have
1819 two sequences of OIDs for trusted and rejected settings. These will
1820 typically have values the same as the extended key usage extension
1821 and any application specific purposes.
1823 The trust checking code now has a default behaviour: it will just
1824 check for an object with the same NID as the passed id. Functions can
1825 be provided to override either the default behaviour or the behaviour
1826 for a given id. SSL client, server and email already have functions
1827 in place for compatibility: they check the NID and also return "trusted"
1828 if the certificate is self signed.
1831 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1832 traditional format into an EVP_PKEY structure.
1835 *) Add a password callback function PEM_cb() which either prompts for
1836 a password if usr_data is NULL or otherwise assumes it is a null
1837 terminated password. Allow passwords to be passed on command line
1838 environment or config files in a few more utilities.
1841 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1842 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1843 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1844 Update documentation.
1847 *) Support for ASN1 "NULL" type. This could be handled before by using
1848 ASN1_TYPE but there wasn't any function that would try to read a NULL
1849 and produce an error if it couldn't. For compatibility we also have
1850 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1851 don't allocate anything because they don't need to.
1854 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1856 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1858 *) Rebuild of the memory allocation routines used by OpenSSL code and
1859 possibly others as well. The purpose is to make an interface that
1860 provide hooks so anyone can build a separate set of allocation and
1861 deallocation routines to be used by OpenSSL, for example memory
1862 pool implementations, or something else, which was previously hard
1863 since Malloc(), Realloc() and Free() were defined as macros having
1864 the values malloc, realloc and free, respectively (except for Win32
1865 compilations). The same is provided for memory debugging code.
1866 OpenSSL already comes with functionality to find memory leaks, but
1867 this gives people a chance to debug other memory problems.
1869 With these changes, a new set of functions and macros have appeared:
1871 CRYPTO_set_mem_debug_functions() [F]
1872 CRYPTO_get_mem_debug_functions() [F]
1873 CRYPTO_dbg_set_options() [F]
1874 CRYPTO_dbg_get_options() [F]
1875 CRYPTO_malloc_debug_init() [M]
1877 The memory debug functions are NULL by default, unless the library
1878 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1879 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1880 gives the standard debugging functions that come with OpenSSL) or
1881 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1882 provided by the library user) must be used. When the standard
1883 debugging functions are used, CRYPTO_dbg_set_options can be used to
1884 request additional information:
1885 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1886 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1888 Also, things like CRYPTO_set_mem_functions will always give the
1889 expected result (the new set of functions is used for allocation
1890 and deallocation) at all times, regardless of platform and compiler
1893 To finish it up, some functions that were never use in any other
1894 way than through macros have a new API and new semantic:
1897 CRYPTO_dbg_realloc()
1900 All macros of value have retained their old syntax.
1901 [Richard Levitte and Bodo Moeller]
1903 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1904 ordering of SMIMECapabilities wasn't in "strength order" and there
1905 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1909 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1910 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1911 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1913 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1914 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1915 functionality to handle multipart/signed properly) and a utility
1916 called 'smime' to call all this stuff. This is based on code I
1917 originally wrote for Celo who have kindly allowed it to be
1918 included in OpenSSL.
1921 *) Add variants des_set_key_checked and des_set_key_unchecked of
1922 des_set_key (aka des_key_sched). Global variable des_check_key
1923 decides which of these is called by des_set_key; this way
1924 des_check_key behaves as it always did, but applications and
1925 the library itself, which was buggy for des_check_key == 1,
1926 have a cleaner way to pick the version they need.
1929 *) New function PKCS12_newpass() which changes the password of a
1933 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1934 dynamic mix. In both cases the ids can be used as an index into the
1935 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
1936 functions so they accept a list of the field values and the
1937 application doesn't need to directly manipulate the X509_TRUST
1941 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
1945 *) Modify the way the V3 extension code looks up extensions. This now
1946 works in a similar way to the object code: we have some "standard"
1947 extensions in a static table which is searched with OBJ_bsearch()
1948 and the application can add dynamic ones if needed. The file
1949 crypto/x509v3/ext_dat.h now has the info: this file needs to be
1950 updated whenever a new extension is added to the core code and kept
1951 in ext_nid order. There is a simple program 'tabtest.c' which checks
1952 this. New extensions are not added too often so this file can readily
1953 be maintained manually.
1955 There are two big advantages in doing things this way. The extensions
1956 can be looked up immediately and no longer need to be "added" using
1957 X509V3_add_standard_extensions(): this function now does nothing.
1958 [Side note: I get *lots* of email saying the extension code doesn't
1959 work because people forget to call this function]
1960 Also no dynamic allocation is done unless new extensions are added:
1961 so if we don't add custom extensions there is no need to call
1962 X509V3_EXT_cleanup().
1965 *) Modify enc utility's salting as follows: make salting the default. Add a
1966 magic header, so unsalted files fail gracefully instead of just decrypting
1967 to garbage. This is because not salting is a big security hole, so people
1968 should be discouraged from doing it.
1971 *) Fixes and enhancements to the 'x509' utility. It allowed a message
1972 digest to be passed on the command line but it only used this
1973 parameter when signing a certificate. Modified so all relevant
1974 operations are affected by the digest parameter including the
1975 -fingerprint and -x509toreq options. Also -x509toreq choked if a
1976 DSA key was used because it didn't fix the digest.
1979 *) Initial certificate chain verify code. Currently tests the untrusted
1980 certificates for consistency with the verify purpose (which is set
1981 when the X509_STORE_CTX structure is set up) and checks the pathlength.
1983 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
1984 this is because it will reject chains with invalid extensions whereas
1985 every previous version of OpenSSL and SSLeay made no checks at all.
1987 Trust code: checks the root CA for the relevant trust settings. Trust
1988 settings have an initial value consistent with the verify purpose: e.g.
1989 if the verify purpose is for SSL client use it expects the CA to be
1990 trusted for SSL client use. However the default value can be changed to
1991 permit custom trust settings: one example of this would be to only trust
1992 certificates from a specific "secure" set of CAs.
1994 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
1995 which should be used for version portability: especially since the
1996 verify structure is likely to change more often now.
1998 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
1999 to set them. If not set then assume SSL clients will verify SSL servers
2002 Two new options to the verify program: -untrusted allows a set of
2003 untrusted certificates to be passed in and -purpose which sets the
2004 intended purpose of the certificate. If a purpose is set then the
2005 new chain verify code is used to check extension consistency.
2008 *) Support for the authority information access extension.
2011 *) Modify RSA and DSA PEM read routines to transparently handle
2012 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2013 public keys in a format compatible with certificate
2014 SubjectPublicKeyInfo structures. Unfortunately there were already
2015 functions called *_PublicKey_* which used various odd formats so
2016 these are retained for compatibility: however the DSA variants were
2017 never in a public release so they have been deleted. Changed dsa/rsa
2018 utilities to handle the new format: note no releases ever handled public
2019 keys so we should be OK.
2021 The primary motivation for this change is to avoid the same fiasco
2022 that dogs private keys: there are several incompatible private key
2023 formats some of which are standard and some OpenSSL specific and
2024 require various evil hacks to allow partial transparent handling and
2025 even then it doesn't work with DER formats. Given the option anything
2026 other than PKCS#8 should be dumped: but the other formats have to
2027 stay in the name of compatibility.
2029 With public keys and the benefit of hindsight one standard format
2030 is used which works with EVP_PKEY, RSA or DSA structures: though
2031 it clearly returns an error if you try to read the wrong kind of key.
2033 Added a -pubkey option to the 'x509' utility to output the public key.
2034 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2035 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2036 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2037 that do the same as the EVP_PKEY_assign_*() except they up the
2038 reference count of the added key (they don't "swallow" the
2042 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2043 CRLs would fail if the file contained no certificates or no CRLs:
2044 added a new function to read in both types and return the number
2045 read: this means that if none are read it will be an error. The
2046 DER versions of the certificate and CRL reader would always fail
2047 because it isn't possible to mix certificates and CRLs in DER format
2048 without choking one or the other routine. Changed this to just read
2049 a certificate: this is the best we can do. Also modified the code
2050 in apps/verify.c to take notice of return codes: it was previously
2051 attempting to read in certificates from NULL pointers and ignoring
2052 any errors: this is one reason why the cert and CRL reader seemed
2053 to work. It doesn't check return codes from the default certificate
2054 routines: these may well fail if the certificates aren't installed.
2057 *) Code to support otherName option in GeneralName.
2060 *) First update to verify code. Change the verify utility
2061 so it warns if it is passed a self signed certificate:
2062 for consistency with the normal behaviour. X509_verify
2063 has been modified to it will now verify a self signed
2064 certificate if *exactly* the same certificate appears
2065 in the store: it was previously impossible to trust a
2066 single self signed certificate. This means that:
2067 openssl verify ss.pem
2068 now gives a warning about a self signed certificate but
2069 openssl verify -CAfile ss.pem ss.pem
2073 *) For servers, store verify_result in SSL_SESSION data structure
2074 (and add it to external session representation).
2075 This is needed when client certificate verifications fails,
2076 but an application-provided verification callback (set by
2077 SSL_CTX_set_cert_verify_callback) allows accepting the session
2078 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2079 but returns 1): When the session is reused, we have to set
2080 ssl->verify_result to the appropriate error code to avoid
2082 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2084 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2085 case in PKCS7_dataInit() where the signed PKCS7 structure
2086 didn't contain any existing data because it was being created.
2087 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2089 *) Add a salt to the key derivation routines in enc.c. This
2090 forms the first 8 bytes of the encrypted file. Also add a
2091 -S option to allow a salt to be input on the command line.
2094 *) New function X509_cmp(). Oddly enough there wasn't a function
2095 to compare two certificates. We do this by working out the SHA1
2096 hash and comparing that. X509_cmp() will be needed by the trust
2100 *) SSL_get1_session() is like SSL_get_session(), but increments
2101 the reference count in the SSL_SESSION returned.
2102 [Geoff Thorpe <geoff@eu.c2.net>]
2104 *) Fix for 'req': it was adding a null to request attributes.
2105 Also change the X509_LOOKUP and X509_INFO code to handle
2106 certificate auxiliary information.
2109 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2113 *) Add the possibility to add extra information to the memory leak
2114 detecting output, to form tracebacks, showing from where each
2115 allocation was originated: CRYPTO_push_info("constant string") adds
2116 the string plus current file name and line number to a per-thread
2117 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2118 is like calling CYRPTO_pop_info() until the stack is empty.
2119 Also updated memory leak detection code to be multi-thread-safe.
2122 *) Add options -text and -noout to pkcs7 utility and delete the
2123 encryption options which never did anything. Update docs.
2126 *) Add options to some of the utilities to allow the pass phrase
2127 to be included on either the command line (not recommended on
2128 OSes like Unix) or read from the environment. Update the
2129 manpages and fix a few bugs.
2132 *) Add a few manpages for some of the openssl commands.
2135 *) Fix the -revoke option in ca. It was freeing up memory twice,
2136 leaking and not finding already revoked certificates.
2139 *) Extensive changes to support certificate auxiliary information.
2140 This involves the use of X509_CERT_AUX structure and X509_AUX
2141 functions. An X509_AUX function such as PEM_read_X509_AUX()
2142 can still read in a certificate file in the usual way but it
2143 will also read in any additional "auxiliary information". By
2144 doing things this way a fair degree of compatibility can be
2145 retained: existing certificates can have this information added
2146 using the new 'x509' options.
2148 Current auxiliary information includes an "alias" and some trust
2149 settings. The trust settings will ultimately be used in enhanced
2150 certificate chain verification routines: currently a certificate
2151 can only be trusted if it is self signed and then it is trusted
2155 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2156 The problem was that one of the replacement routines had not been working
2157 since SSLeay releases. For now the offending routine has been replaced
2158 with non-optimised assembler. Even so, this now gives around 95%
2159 performance improvement for 1024 bit RSA signs.
2162 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2163 handling. Most clients have the effective key size in bits equal to
2164 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2165 A few however don't do this and instead use the size of the decrypted key
2166 to determine the RC2 key length and the AlgorithmIdentifier to determine
2167 the effective key length. In this case the effective key length can still
2168 be 40 bits but the key length can be 168 bits for example. This is fixed
2169 by manually forcing an RC2 key into the EVP_PKEY structure because the
2170 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2171 the key length and effective key length are equal.
2174 *) Add a bunch of functions that should simplify the creation of
2175 X509_NAME structures. Now you should be able to do:
2176 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2177 and have it automatically work out the correct field type and fill in
2178 the structures. The more adventurous can try:
2179 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2180 and it will (hopefully) work out the correct multibyte encoding.
2183 *) Change the 'req' utility to use the new field handling and multibyte
2184 copy routines. Before the DN field creation was handled in an ad hoc
2185 way in req, ca, and x509 which was rather broken and didn't support
2186 BMPStrings or UTF8Strings. Since some software doesn't implement
2187 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2188 using the dirstring_type option. See the new comment in the default
2189 openssl.cnf for more info.
2192 *) Make crypto/rand/md_rand.c more robust:
2193 - Assure unique random numbers after fork().
2194 - Make sure that concurrent threads access the global counter and
2195 md serializably so that we never lose entropy in them
2196 or use exactly the same state in multiple threads.
2197 Access to the large state is not always serializable because
2198 the additional locking could be a performance killer, and
2199 md should be large enough anyway.
2202 *) New file apps/app_rand.c with commonly needed functionality
2203 for handling the random seed file.
2205 Use the random seed file in some applications that previously did not:
2207 dsaparam -genkey (which also ignored its '-rand' option),
2210 x509 (when signing).
2211 Except on systems with /dev/urandom, it is crucial to have a random
2212 seed file at least for key creation, DSA signing, and for DH exchanges;
2213 for RSA signatures we could do without one.
2215 gendh and gendsa (unlike genrsa) used to read only the first byte
2216 of each file listed in the '-rand' option. The function as previously
2217 found in genrsa is now in app_rand.c and is used by all programs
2218 that support '-rand'.
2221 *) In RAND_write_file, use mode 0600 for creating files;
2222 don't just chmod when it may be too late.
2225 *) Report an error from X509_STORE_load_locations
2226 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2229 *) New function ASN1_mbstring_copy() this copies a string in either
2230 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2231 into an ASN1_STRING type. A mask of permissible types is passed
2232 and it chooses the "minimal" type to use or an error if not type
2236 *) Add function equivalents to the various macros in asn1.h. The old
2237 macros are retained with an M_ prefix. Code inside the library can
2238 use the M_ macros. External code (including the openssl utility)
2239 should *NOT* in order to be "shared library friendly".
2242 *) Add various functions that can check a certificate's extensions
2243 to see if it usable for various purposes such as SSL client,
2244 server or S/MIME and CAs of these types. This is currently
2245 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2246 verification. Also added a -purpose flag to x509 utility to
2247 print out all the purposes.
2250 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2254 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2255 for, obtain and decode and extension and obtain its critical flag.
2256 This allows all the necessary extension code to be handled in a
2257 single function call.
2260 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2261 platforms. See crypto/rc4/rc4_enc.c for further details.
2264 *) New -noout option to asn1parse. This causes no output to be produced
2265 its main use is when combined with -strparse and -out to extract data
2266 from a file (which may not be in ASN.1 format).
2269 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2270 when producing the local key id.
2271 [Richard Levitte <levitte@stacken.kth.se>]
2273 *) New option -dhparam in s_server. This allows a DH parameter file to be
2274 stated explicitly. If it is not stated then it tries the first server
2275 certificate file. The previous behaviour hard coded the filename
2279 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2280 a public key to be input or output. For example:
2281 openssl rsa -in key.pem -pubout -out pubkey.pem
2282 Also added necessary DSA public key functions to handle this.
2285 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2286 in the message. This was handled by allowing
2287 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2288 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2290 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2291 to the end of the strings whereas this didn't. This would cause problems
2292 if strings read with d2i_ASN1_bytes() were later modified.
2293 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2295 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2296 data and it contains EOF it will end up returning an error. This is
2297 caused by input 46 bytes long. The cause is due to the way base64
2298 BIOs find the start of base64 encoded data. They do this by trying a
2299 trial decode on each line until they find one that works. When they
2300 do a flag is set and it starts again knowing it can pass all the
2301 data directly through the decoder. Unfortunately it doesn't reset
2302 the context it uses. This means that if EOF is reached an attempt
2303 is made to pass two EOFs through the context and this causes the
2304 resulting error. This can also cause other problems as well. As is
2305 usual with these problems it takes *ages* to find and the fix is
2306 trivial: move one line.
2307 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2309 *) Ugly workaround to get s_client and s_server working under Windows. The
2310 old code wouldn't work because it needed to select() on sockets and the
2311 tty (for keypresses and to see if data could be written). Win32 only
2312 supports select() on sockets so we select() with a 1s timeout on the
2313 sockets and then see if any characters are waiting to be read, if none
2314 are present then we retry, we also assume we can always write data to
2315 the tty. This isn't nice because the code then blocks until we've
2316 received a complete line of data and it is effectively polling the
2317 keyboard at 1s intervals: however it's quite a bit better than not
2318 working at all :-) A dedicated Windows application might handle this
2319 with an event loop for example.
2322 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2323 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2324 will be called when RSA_sign() and RSA_verify() are used. This is useful
2325 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2326 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2327 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2328 This necessitated the support of an extra signature type NID_md5_sha1
2329 for SSL signatures and modifications to the SSL library to use it instead
2330 of calling RSA_public_decrypt() and RSA_private_encrypt().
2333 *) Add new -verify -CAfile and -CApath options to the crl program, these
2334 will lookup a CRL issuers certificate and verify the signature in a
2335 similar way to the verify program. Tidy up the crl program so it
2336 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2337 less strict. It will now permit CRL extensions even if it is not
2338 a V2 CRL: this will allow it to tolerate some broken CRLs.
2341 *) Initialize all non-automatic variables each time one of the openssl
2342 sub-programs is started (this is necessary as they may be started
2343 multiple times from the "OpenSSL>" prompt).
2344 [Lennart Bang, Bodo Moeller]
2346 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2347 removing all other RSA functionality (this is what NO_RSA does). This
2348 is so (for example) those in the US can disable those operations covered
2349 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2353 *) Non-copying interface to BIO pairs.
2354 (still largely untested)
2357 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2358 ASCII string. This was handled independently in various places before.
2361 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2362 UTF8 strings a character at a time.
2365 *) Use client_version from client hello to select the protocol
2366 (s23_srvr.c) and for RSA client key exchange verification
2367 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2370 *) Add various utility functions to handle SPKACs, these were previously
2371 handled by poking round in the structure internals. Added new function
2372 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2373 print, verify and generate SPKACs. Based on an original idea from
2374 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2377 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2380 *) Allow the config file extension section to be overwritten on the
2381 command line. Based on an original idea from Massimiliano Pala
2382 <madwolf@comune.modena.it>. The new option is called -extensions
2383 and can be applied to ca, req and x509. Also -reqexts to override
2384 the request extensions in req and -crlexts to override the crl extensions
2388 *) Add new feature to the SPKAC handling in ca. Now you can include
2389 the same field multiple times by preceding it by "XXXX." for example:
2392 this is the same syntax as used in the req config file.
2395 *) Allow certificate extensions to be added to certificate requests. These
2396 are specified in a 'req_extensions' option of the req section of the
2397 config file. They can be printed out with the -text option to req but
2398 are otherwise ignored at present.
2401 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2402 data read consists of only the final block it would not decrypted because
2403 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2404 A misplaced 'break' also meant the decrypted final block might not be
2405 copied until the next read.
2408 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2409 a few extra parameters to the DH structure: these will be useful if
2410 for example we want the value of 'q' or implement X9.42 DH.
2413 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2414 provides hooks that allow the default DSA functions or functions on a
2415 "per key" basis to be replaced. This allows hardware acceleration and
2416 hardware key storage to be handled without major modification to the
2417 library. Also added low level modexp hooks and CRYPTO_EX structure and
2418 associated functions.
2421 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2422 as "read only": it can't be written to and the buffer it points to will
2423 not be freed. Reading from a read only BIO is much more efficient than
2424 a normal memory BIO. This was added because there are several times when
2425 an area of memory needs to be read from a BIO. The previous method was
2426 to create a memory BIO and write the data to it, this results in two
2427 copies of the data and an O(n^2) reading algorithm. There is a new
2428 function BIO_new_mem_buf() which creates a read only memory BIO from
2429 an area of memory. Also modified the PKCS#7 routines to use read only
2433 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2434 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2435 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2436 but a retry condition occured while trying to read the rest.
2439 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2440 NID_pkcs7_encrypted by default: this was wrong since this should almost
2441 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2442 the encrypted data type: this is a more sensible place to put it and it
2443 allows the PKCS#12 code to be tidied up that duplicated this
2447 *) Changed obj_dat.pl script so it takes its input and output files on
2448 the command line. This should avoid shell escape redirection problems
2452 *) Initial support for certificate extension requests, these are included
2453 in things like Xenroll certificate requests. Included functions to allow
2454 extensions to be obtained and added.
2457 *) -crlf option to s_client and s_server for sending newlines as
2458 CRLF (as required by many protocols).
2461 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2463 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2464 [Ralf S. Engelschall]
2466 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2467 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2469 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2473 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2474 DH parameters/keys (q is lost during that conversion, but the resulting
2475 DH parameters contain its length).
2477 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2478 much faster than DH_generate_parameters (which creates parameters
2479 where p = 2*q + 1), and also the smaller q makes DH computations
2480 much more efficient (160-bit exponentiation instead of 1024-bit
2481 exponentiation); so this provides a convenient way to support DHE
2482 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2483 utter importance to use
2484 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2486 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2487 when such DH parameters are used, because otherwise small subgroup
2488 attacks may become possible!
2491 *) Avoid memory leak in i2d_DHparams.
2494 *) Allow the -k option to be used more than once in the enc program:
2495 this allows the same encrypted message to be read by multiple recipients.
2498 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2499 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2500 it will always use the numerical form of the OID, even if it has a short
2504 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2505 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2506 otherwise bn_mod_exp was called. In the case of hardware keys for example
2507 no private key components need be present and it might store extra data
2508 in the RSA structure, which cannot be accessed from bn_mod_exp.
2509 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2510 private key operations.
2513 *) Added support for SPARC Linux.
2516 *) pem_password_cb function type incompatibly changed from
2517 typedef int pem_password_cb(char *buf, int size, int rwflag);
2519 ....(char *buf, int size, int rwflag, void *userdata);
2520 so that applications can pass data to their callbacks:
2521 The PEM[_ASN1]_{read,write}... functions and macros now take an
2522 additional void * argument, which is just handed through whenever
2523 the password callback is called.
2524 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2526 New function SSL_CTX_set_default_passwd_cb_userdata.
2528 Compatibility note: As many C implementations push function arguments
2529 onto the stack in reverse order, the new library version is likely to
2530 interoperate with programs that have been compiled with the old
2531 pem_password_cb definition (PEM_whatever takes some data that
2532 happens to be on the stack as its last argument, and the callback
2533 just ignores this garbage); but there is no guarantee whatsoever that
2536 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2537 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2538 problems not only on Windows, but also on some Unix platforms.
2539 To avoid problematic command lines, these definitions are now in an
2540 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2541 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2544 *) MIPS III/IV assembler module is reimplemented.
2547 *) More DES library cleanups: remove references to srand/rand and
2548 delete an unused file.
2551 *) Add support for the the free Netwide assembler (NASM) under Win32,
2552 since not many people have MASM (ml) and it can be hard to obtain.
2553 This is currently experimental but it seems to work OK and pass all
2554 the tests. Check out INSTALL.W32 for info.
2557 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2558 without temporary keys kept an extra copy of the server key,
2559 and connections with temporary keys did not free everything in case
2563 *) New function RSA_check_key and new openssl rsa option -check
2564 for verifying the consistency of RSA keys.
2565 [Ulf Moeller, Bodo Moeller]
2567 *) Various changes to make Win32 compile work:
2568 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2569 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2570 comparison" warnings.
2571 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2574 *) Add a debugging option to PKCS#5 v2 key generation function: when
2575 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2576 derived keys are printed to stderr.
2579 *) Copy the flags in ASN1_STRING_dup().
2580 [Roman E. Pavlov <pre@mo.msk.ru>]
2582 *) The x509 application mishandled signing requests containing DSA
2583 keys when the signing key was also DSA and the parameters didn't match.
2585 It was supposed to omit the parameters when they matched the signing key:
2586 the verifying software was then supposed to automatically use the CA's
2587 parameters if they were absent from the end user certificate.
2589 Omitting parameters is no longer recommended. The test was also
2590 the wrong way round! This was probably due to unusual behaviour in
2591 EVP_cmp_parameters() which returns 1 if the parameters match.
2592 This meant that parameters were omitted when they *didn't* match and
2593 the certificate was useless. Certificates signed with 'ca' didn't have
2595 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2597 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2598 The interface is as follows:
2599 Applications can use
2600 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2601 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2602 "off" is now the default.
2603 The library internally uses
2604 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2605 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2606 to disable memory-checking temporarily.
2608 Some inconsistent states that previously were possible (and were
2609 even the default) are now avoided.
2611 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2612 with each memory chunk allocated; this is occasionally more helpful
2613 than just having a counter.
2615 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2617 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2621 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2622 which largely parallels "options", but is for changing API behaviour,
2623 whereas "options" are about protocol behaviour.
2624 Initial "mode" flags are:
2626 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2627 a single record has been written.
2628 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2629 retries use the same buffer location.
2630 (But all of the contents must be
2634 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2637 *) Fix problems with no-hmac etc.
2638 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2640 *) New functions RSA_get_default_method(), RSA_set_method() and
2641 RSA_get_method(). These allows replacement of RSA_METHODs without having
2642 to mess around with the internals of an RSA structure.
2645 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2646 Also really enable memory leak checks in openssl.c and in some
2648 [Chad C. Mulligan, Bodo Moeller]
2650 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2651 up the length of negative integers. This has now been simplified to just
2652 store the length when it is first determined and use it later, rather
2653 than trying to keep track of where data is copied and updating it to
2655 [Steve Henson, reported by Brien Wheeler
2656 <bwheeler@authentica-security.com>]
2658 *) Add a new function PKCS7_signatureVerify. This allows the verification
2659 of a PKCS#7 signature but with the signing certificate passed to the
2660 function itself. This contrasts with PKCS7_dataVerify which assumes the
2661 certificate is present in the PKCS#7 structure. This isn't always the
2662 case: certificates can be omitted from a PKCS#7 structure and be
2663 distributed by "out of band" means (such as a certificate database).
2666 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2667 function prototypes in pem.h, also change util/mkdef.pl to add the
2668 necessary function names.
2671 *) mk1mf.pl (used by Windows builds) did not properly read the
2672 options set by Configure in the top level Makefile, and Configure
2673 was not even able to write more than one option correctly.
2674 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2677 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2678 file to be loaded from a BIO or FILE pointer. The BIO version will
2679 for example allow memory BIOs to contain config info.
2682 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2683 Whoever hopes to achieve shared-library compatibility across versions
2684 must use this, not the compile-time macro.
2685 (Exercise 0.9.4: Which is the minimum library version required by
2687 Note: All this applies only to multi-threaded programs, others don't
2691 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2692 through a BIO pair triggered the default case, i.e.
2693 SSLerr(...,SSL_R_UNKNOWN_STATE).
2696 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2697 can use the SSL library even if none of the specific BIOs is
2701 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2702 for the encoded length.
2703 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2705 *) Add initial documentation of the X509V3 functions.
2708 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2709 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2710 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2711 secure PKCS#8 private key format with a high iteration count.
2714 *) Fix determination of Perl interpreter: A perl or perl5
2715 _directory_ in $PATH was also accepted as the interpreter.
2716 [Ralf S. Engelschall]
2718 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2719 wrong with it but it was very old and did things like calling
2720 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2724 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2725 to use the new extension code.
2728 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2729 with macros. This should make it easier to change their form, add extra
2730 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2734 *) Add to configuration table a new entry that can specify an alternative
2735 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2736 according to Mark Crispin <MRC@Panda.COM>.
2740 *) DES CBC did not update the IV. Weird.
2743 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2744 Changing the behaviour of the former might break existing programs --
2745 where IV updating is needed, des_ncbc_encrypt can be used.
2748 *) When bntest is run from "make test" it drives bc to check its
2749 calculations, as well as internally checking them. If an internal check
2750 fails, it needs to cause bc to give a non-zero result or make test carries
2751 on without noticing the failure. Fixed.
2754 *) DES library cleanups.
2757 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2758 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2759 ciphers. NOTE: although the key derivation function has been verified
2760 against some published test vectors it has not been extensively tested
2761 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2765 *) Instead of "mkdir -p", which is not fully portable, use new
2766 Perl script "util/mkdir-p.pl".
2769 *) Rewrite the way password based encryption (PBE) is handled. It used to
2770 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2771 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2772 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2773 the 'parameter' field of the AlgorithmIdentifier is passed to the
2774 underlying key generation function so it must do its own ASN1 parsing.
2775 This has also changed the EVP_PBE_CipherInit() function which now has a
2776 'parameter' argument instead of literal salt and iteration count values
2777 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2780 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2781 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2782 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2783 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2784 value was just used as a "magic string" and not used directly its
2785 value doesn't matter.
2788 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2792 *) "linux-sparc64" configuration (ultrapenguin).
2793 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2794 "linux-sparc" configuration.
2795 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2797 *) config now generates no-xxx options for missing ciphers.
2800 *) Support the EBCDIC character set (work in progress).
2801 File ebcdic.c not yet included because it has a different license.
2802 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2804 *) Support BS2000/OSD-POSIX.
2805 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2807 *) Make callbacks for key generation use void * instead of char *.
2810 *) Make S/MIME samples compile (not yet tested).
2813 *) Additional typesafe stacks.
2816 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2820 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2822 *) New configuration variant "sco5-gcc".
2824 *) Updated some demos.
2825 [Sean O Riordain, Wade Scholine]
2827 *) Add missing BIO_free at exit of pkcs12 application.
2830 *) Fix memory leak in conf.c.
2833 *) Updates for Win32 to assembler version of MD5.
2836 *) Set #! path to perl in apps/der_chop to where we found it
2837 instead of using a fixed path.
2840 *) SHA library changes for irix64-mips4-cc.
2843 *) Improvements for VMS support.
2847 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2849 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2850 This also avoids the problems with SC4.2 and unpatched SC5.
2851 [Andy Polyakov <appro@fy.chalmers.se>]
2853 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2854 These are required because of the typesafe stack would otherwise break
2855 existing code. If old code used a structure member which used to be STACK
2856 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2857 sk_num or sk_value it would produce an error because the num, data members
2858 are not present in STACK_OF. Now it just produces a warning. sk_set
2859 replaces the old method of assigning a value to sk_value
2860 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2861 that does this will no longer work (and should use sk_set instead) but
2862 this could be regarded as a "questionable" behaviour anyway.
2865 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2866 correctly handle encrypted S/MIME data.
2869 *) Change type of various DES function arguments from des_cblock
2870 (which means, in function argument declarations, pointer to char)
2871 to des_cblock * (meaning pointer to array with 8 char elements),
2872 which allows the compiler to do more typechecking; it was like
2873 that back in SSLeay, but with lots of ugly casts.
2875 Introduce new type const_des_cblock.
2878 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2879 problems: find RecipientInfo structure that matches recipient certificate
2880 and initialise the ASN1 structures properly based on passed cipher.
2883 *) Belatedly make the BN tests actually check the results.
2886 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2887 to and from BNs: it was completely broken. New compilation option
2888 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2889 key elements as negative integers.
2892 *) Reorganize and speed up MD5.
2893 [Andy Polyakov <appro@fy.chalmers.se>]
2896 [Richard Levitte <richard@levitte.org>]
2898 *) New option -out to asn1parse to allow the parsed structure to be
2899 output to a file. This is most useful when combined with the -strparse
2900 option to examine the output of things like OCTET STRINGS.
2903 *) Make SSL library a little more fool-proof by not requiring any longer
2904 that SSL_set_{accept,connect}_state be called before
2905 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2906 in many applications because usually everything *appeared* to work as
2907 intended anyway -- now it really works as intended).
2910 *) Move openssl.cnf out of lib/.
2913 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2914 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2915 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2916 [Ralf S. Engelschall]
2918 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2919 handle PKCS#7 enveloped data properly.
2920 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2922 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2923 copying pointers. The cert_st handling is changed by this in
2924 various ways (and thus what used to be known as ctx->default_cert
2925 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2926 any longer when s->cert does not give us what we need).
2927 ssl_cert_instantiate becomes obsolete by this change.
2928 As soon as we've got the new code right (possibly it already is?),
2929 we have solved a couple of bugs of the earlier code where s->cert
2930 was used as if it could not have been shared with other SSL structures.
2932 Note that using the SSL API in certain dirty ways now will result
2933 in different behaviour than observed with earlier library versions:
2934 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
2935 does not influence s as it used to.
2937 In order to clean up things more thoroughly, inside SSL_SESSION
2938 we don't use CERT any longer, but a new structure SESS_CERT
2939 that holds per-session data (if available); currently, this is
2940 the peer's certificate chain and, for clients, the server's certificate
2941 and temporary key. CERT holds only those values that can have
2942 meaningful defaults in an SSL_CTX.
2945 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
2946 from the internal representation. Various PKCS#7 fixes: remove some
2947 evil casts and set the enc_dig_alg field properly based on the signing
2951 *) Allow PKCS#12 password to be set from the command line or the
2952 environment. Let 'ca' get its config file name from the environment
2953 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
2957 *) Allow certificate policies extension to use an IA5STRING for the
2958 organization field. This is contrary to the PKIX definition but
2959 VeriSign uses it and IE5 only recognises this form. Document 'x509'
2963 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
2964 without disallowing inline assembler and the like for non-pedantic builds.
2967 *) Support Borland C++ builder.
2968 [Janez Jere <jj@void.si>, modified by Ulf Möller]
2973 *) SHA-1 cleanups and performance enhancements.
2974 [Andy Polyakov <appro@fy.chalmers.se>]
2976 *) Sparc v8plus assembler for the bignum library.
2977 [Andy Polyakov <appro@fy.chalmers.se>]
2979 *) Accept any -xxx and +xxx compiler options in Configure.
2982 *) Update HPUX configuration.
2985 *) Add missing sk_<type>_unshift() function to safestack.h
2986 [Ralf S. Engelschall]
2988 *) New function SSL_CTX_use_certificate_chain_file that sets the
2989 "extra_cert"s in addition to the certificate. (This makes sense
2990 only for "PEM" format files, as chains as a whole are not
2994 *) Support verify_depth from the SSL API.
2995 x509_vfy.c had what can be considered an off-by-one-error:
2996 Its depth (which was not part of the external interface)
2997 was actually counting the number of certificates in a chain;
2998 now it really counts the depth.
3001 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3002 instead of X509err, which often resulted in confusing error
3003 messages since the error codes are not globally unique
3004 (e.g. an alleged error in ssl3_accept when a certificate
3005 didn't match the private key).
3007 *) New function SSL_CTX_set_session_id_context that allows to set a default
3008 value (so that you don't need SSL_set_session_id_context for each
3009 connection using the SSL_CTX).
3012 *) OAEP decoding bug fix.
3015 *) Support INSTALL_PREFIX for package builders, as proposed by
3019 *) New Configure options "threads" and "no-threads". For systems
3020 where the proper compiler options are known (currently Solaris
3021 and Linux), "threads" is the default.
3024 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3027 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3028 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3029 such as /usr/local/bin.
3032 *) "make linux-shared" to build shared libraries.
3033 [Niels Poppe <niels@netbox.org>]
3035 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3038 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3039 extension adding in x509 utility.
3042 *) Remove NOPROTO sections and error code comments.
3045 *) Partial rewrite of the DEF file generator to now parse the ANSI
3049 *) New Configure options --prefix=DIR and --openssldir=DIR.
3052 *) Complete rewrite of the error code script(s). It is all now handled
3053 by one script at the top level which handles error code gathering,
3054 header rewriting and C source file generation. It should be much better
3055 than the old method: it now uses a modified version of Ulf's parser to
3056 read the ANSI prototypes in all header files (thus the old K&R definitions
3057 aren't needed for error creation any more) and do a better job of
3058 translating function codes into names. The old 'ASN1 error code imbedded
3059 in a comment' is no longer necessary and it doesn't use .err files which
3060 have now been deleted. Also the error code call doesn't have to appear all
3061 on one line (which resulted in some large lines...).
3064 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3067 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3068 0 (which usually indicates a closed connection), but continue reading.
3071 *) Fix some race conditions.
3074 *) Add support for CRL distribution points extension. Add Certificate
3075 Policies and CRL distribution points documentation.
3078 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3081 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3082 8 of keying material. Merlin has also confirmed interop with this fix
3083 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3084 [Merlin Hughes <merlin@baltimore.ie>]
3086 *) Fix lots of warnings.
3087 [Richard Levitte <levitte@stacken.kth.se>]
3089 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3090 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3091 [Richard Levitte <levitte@stacken.kth.se>]
3093 *) Fix problems with sizeof(long) == 8.
3094 [Andy Polyakov <appro@fy.chalmers.se>]
3096 *) Change functions to ANSI C.
3099 *) Fix typos in error codes.
3100 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3102 *) Remove defunct assembler files from Configure.
3105 *) SPARC v8 assembler BIGNUM implementation.
3106 [Andy Polyakov <appro@fy.chalmers.se>]
3108 *) Support for Certificate Policies extension: both print and set.
3109 Various additions to support the r2i method this uses.
3112 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3113 return a const string when you are expecting an allocated buffer.
3116 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3117 types DirectoryString and DisplayText.
3120 *) Add code to allow r2i extensions to access the configuration database,
3121 add an LHASH database driver and add several ctx helper functions.
3124 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3125 fail when they extended the size of a BIGNUM.
3128 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3129 support typesafe stack.
3132 *) Fix typo in SSL_[gs]et_options().
3133 [Nils Frostberg <nils@medcom.se>]
3135 *) Delete various functions and files that belonged to the (now obsolete)
3136 old X509V3 handling code.
3139 *) New Configure option "rsaref".
3142 *) Don't auto-generate pem.h.
3145 *) Introduce type-safe ASN.1 SETs.
3148 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3149 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3151 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3152 that links with OpenSSL (well at least cause lots of warnings), but fear
3153 not: the conversion is trivial, and it eliminates loads of evil casts. A
3154 few STACKed things have been converted already. Feel free to convert more.
3155 In the fullness of time, I'll do away with the STACK type altogether.
3158 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3159 specified in <certfile> by updating the entry in the index.txt file.
3160 This way one no longer has to edit the index.txt file manually for
3161 revoking a certificate. The -revoke option does the gory details now.
3162 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3164 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3165 `-text' option at all and this way the `-noout -text' combination was
3166 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3167 [Ralf S. Engelschall]
3169 *) Make sure a corresponding plain text error message exists for the
3170 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3171 verify callback function determined that a certificate was revoked.
3172 [Ralf S. Engelschall]
3174 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3175 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3176 all available cipers including rc5, which was forgotten until now.
3177 In order to let the testing shell script know which algorithms
3178 are available, a new (up to now undocumented) command
3179 "openssl list-cipher-commands" is used.
3182 *) Bugfix: s_client occasionally would sleep in select() when
3183 it should have checked SSL_pending() first.
3186 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3187 the raw DSA values prior to ASN.1 encoding.
3190 *) Tweaks to Configure
3191 [Niels Poppe <niels@netbox.org>]
3193 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3197 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3200 *) New config option to avoid instructions that are illegal on the 80386.
3201 The default code is faster, but requires at least a 486.
3204 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3205 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3206 same as SSL2_VERSION anyway.
3209 *) New "-showcerts" option for s_client.
3212 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3213 application. Various cleanups and fixes.
3216 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3217 modify error routines to work internally. Add error codes and PBE init
3218 to library startup routines.
3221 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3222 packing functions to asn1 and evp. Changed function names and error
3223 codes along the way.
3226 *) PKCS12 integration: and so it begins... First of several patches to
3227 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3228 objects to objects.h
3231 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3232 and display support for Thawte strong extranet extension.
3235 *) Add LinuxPPC support.
3236 [Jeff Dubrule <igor@pobox.org>]
3238 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3239 bn_div_words in alpha.s.
3240 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3242 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3243 OAEP isn't supported when OpenSSL is built with RSAref.
3244 [Ulf Moeller <ulf@fitug.de>]
3246 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3247 so they no longer are missing under -DNOPROTO.
3248 [Soren S. Jorvang <soren@t.dk>]
3251 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3253 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3254 doesn't work when the session is reused. Coming soon!
3257 *) Fix a security hole, that allows sessions to be reused in the wrong
3258 context thus bypassing client cert protection! All software that uses
3259 client certs and session caches in multiple contexts NEEDS PATCHING to
3260 allow session reuse! A fuller solution is in the works.
3261 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3263 *) Some more source tree cleanups (removed obsolete files
3264 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3265 permission on "config" script to be executable) and a fix for the INSTALL
3267 [Ulf Moeller <ulf@fitug.de>]
3269 *) Remove some legacy and erroneous uses of malloc, free instead of
3271 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3273 *) Make rsa_oaep_test return non-zero on error.
3274 [Ulf Moeller <ulf@fitug.de>]
3276 *) Add support for native Solaris shared libraries. Configure
3277 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3278 if someone would make that last step automatic.
3279 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3281 *) ctx_size was not built with the right compiler during "make links". Fixed.
3284 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3285 except NULL ciphers". This means the default cipher list will no longer
3286 enable NULL ciphers. They need to be specifically enabled e.g. with
3287 the string "DEFAULT:eNULL".
3290 *) Fix to RSA private encryption routines: if p < q then it would
3291 occasionally produce an invalid result. This will only happen with
3292 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3295 *) Be less restrictive and allow also `perl util/perlpath.pl
3296 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3297 because this way one can also use an interpreter named `perl5' (which is
3298 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3299 installed as `perl').
3300 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3302 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3303 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3305 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3306 advapi32.lib to Win32 build and change the pem test comparision
3307 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3308 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3309 and crypto/des/ede_cbcm_enc.c.
3312 *) DES quad checksum was broken on big-endian architectures. Fixed.
3315 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3316 Win32 test batch file so it (might) work again. The Win32 test batch file
3317 is horrible: I feel ill....
3320 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3321 in e_os.h. Audit of header files to check ANSI and non ANSI
3322 sections: 10 functions were absent from non ANSI section and not exported
3323 from Windows DLLs. Fixed up libeay.num for new functions.
3326 *) Make `openssl version' output lines consistent.
3327 [Ralf S. Engelschall]
3329 *) Fix Win32 symbol export lists for BIO functions: Added
3330 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3331 to ms/libeay{16,32}.def.
3332 [Ralf S. Engelschall]
3334 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3335 fine under Unix and passes some trivial tests I've now added. But the
3336 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3337 added to make sure no one expects that this stuff really works in the
3338 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3339 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3341 [Ralf S. Engelschall]
3343 *) Fix the generation of two part addresses in perl.
3344 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3346 *) Add config entry for Linux on MIPS.
3347 [John Tobey <jtobey@channel1.com>]
3349 *) Make links whenever Configure is run, unless we are on Windoze.
3352 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3353 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3357 *) Add a useful kludge to allow package maintainers to specify compiler and
3358 other platforms details on the command line without having to patch the
3359 Configure script everytime: One now can use ``perl Configure
3360 <id>:<details>'', i.e. platform ids are allowed to have details appended
3361 to them (seperated by colons). This is treated as there would be a static
3362 pre-configured entry in Configure's %table under key <id> with value
3363 <details> and ``perl Configure <id>'' is called. So, when you want to
3364 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3365 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3366 now, which overrides the FreeBSD-elf entry on-the-fly.
3367 [Ralf S. Engelschall]
3369 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3372 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3373 on the `perl Configure ...' command line. This way one can compile
3374 OpenSSL libraries with Position Independent Code (PIC) which is needed
3375 for linking it into DSOs.
3376 [Ralf S. Engelschall]
3378 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3382 *) Cleaned up the LICENSE document: The official contact for any license
3383 questions now is the OpenSSL core team under openssl-core@openssl.org.
3384 And add a paragraph about the dual-license situation to make sure people
3385 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3386 to the OpenSSL toolkit.
3387 [Ralf S. Engelschall]
3389 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3390 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3391 Additonally cleaned up the `make links' target: Remove unnecessary
3392 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3393 to speed processing and no longer clutter the display with confusing
3394 stuff. Instead only the actually done links are displayed.
3395 [Ralf S. Engelschall]
3397 *) Permit null encryption ciphersuites, used for authentication only. It used
3398 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3399 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3403 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3404 signed attributes when verifying signatures (this would break them),
3405 the detached data encoding was wrong and public keys obtained using
3406 X509_get_pubkey() weren't freed.
3409 *) Add text documentation for the BUFFER functions. Also added a work around
3410 to a Win95 console bug. This was triggered by the password read stuff: the
3411 last character typed gets carried over to the next fread(). If you were
3412 generating a new cert request using 'req' for example then the last
3413 character of the passphrase would be CR which would then enter the first
3417 *) Added the new `Includes OpenSSL Cryptography Software' button as
3418 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3419 button and can be used by applications based on OpenSSL to show the
3420 relationship to the OpenSSL project.
3421 [Ralf S. Engelschall]
3423 *) Remove confusing variables in function signatures in files
3424 ssl/ssl_lib.c and ssl/ssl.h.
3425 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3427 *) Don't install bss_file.c under PREFIX/include/
3428 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3430 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3431 functions that return function pointers and has support for NT specific
3432 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3433 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3434 unsigned to signed types: this was killing the Win32 compile.
3437 *) Add new certificate file to stack functions,
3438 SSL_add_dir_cert_subjects_to_stack() and
3439 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3440 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3441 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3442 This means that Apache-SSL and similar packages don't have to mess around
3443 to add as many CAs as they want to the preferred list.
3446 *) Experiment with doxygen documentation. Currently only partially applied to
3448 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3449 openssl.doxy as the configuration file.
3452 *) Get rid of remaining C++-style comments which strict C compilers hate.
3453 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3455 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3456 compiled in by default: it has problems with large keys.
3459 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3460 DH private keys and/or callback functions which directly correspond to
3461 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3462 is needed for applications which have to configure certificates on a
3463 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3465 For the RSA certificate situation is makes no difference, but
3466 for the DSA certificate situation this fixes the "no shared cipher"
3467 problem where the OpenSSL cipher selection procedure failed because the
3468 temporary keys were not overtaken from the context and the API provided
3469 no way to reconfigure them.
3470 The new functions now let applications reconfigure the stuff and they
3471 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3472 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3473 non-public-API function ssl_cert_instantiate() is used as a helper
3474 function and also to reduce code redundancy inside ssl_rsa.c.
3475 [Ralf S. Engelschall]
3477 *) Move s_server -dcert and -dkey options out of the undocumented feature
3478 area because they are useful for the DSA situation and should be
3479 recognized by the users.
3480 [Ralf S. Engelschall]
3482 *) Fix the cipher decision scheme for export ciphers: the export bits are
3483 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3484 SSL_EXP_MASK. So, the original variable has to be used instead of the
3485 already masked variable.
3486 [Richard Levitte <levitte@stacken.kth.se>]
3488 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3489 [Richard Levitte <levitte@stacken.kth.se>]
3491 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3492 from `int' to `unsigned int' because it's a length and initialized by
3493 EVP_DigestFinal() which expects an `unsigned int *'.
3494 [Richard Levitte <levitte@stacken.kth.se>]
3496 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3497 script. Instead use the usual Shell->Perl transition trick.
3498 [Ralf S. Engelschall]
3500 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3501 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3502 -noout -modulus' as it's already the case for `openssl rsa -noout
3503 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3504 currently the public key is printed (a decision which was already done by
3505 `openssl dsa -modulus' in the past) which serves a similar purpose.
3506 Additionally the NO_RSA no longer completely removes the whole -modulus
3507 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3509 [Ralf S. Engelschall]
3511 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3512 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3513 [Arne Ansper <arne@ats.cyber.ee>]
3515 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3516 to be added. Now both 'req' and 'ca' can use new objects defined in the
3520 *) Add cool BIO that does syslog (or event log on NT).
3521 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3523 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3524 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3525 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3526 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3529 *) Add preliminary config info for new extension code.
3532 *) Make RSA_NO_PADDING really use no padding.
3533 [Ulf Moeller <ulf@fitug.de>]
3535 *) Generate errors when private/public key check is done.
3538 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3539 for some CRL extensions and new objects added.
3542 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3543 key usage extension and fuller support for authority key id.
3546 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3547 padding method for RSA, which is recommended for new applications in PKCS
3548 #1 v2.0 (RFC 2437, October 1998).
3549 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3550 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3551 against Bleichbacher's attack on RSA.
3552 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3555 *) Updates to the new SSL compression code
3556 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3558 *) Fix so that the version number in the master secret, when passed
3559 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3560 (because the server will not accept higher), that the version number
3561 is 0x03,0x01, not 0x03,0x00
3562 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3564 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3565 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3566 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3569 *) Support for RAW extensions where an arbitrary extension can be
3570 created by including its DER encoding. See apps/openssl.cnf for
3574 *) Make sure latest Perl versions don't interpret some generated C array
3575 code as Perl array code in the crypto/err/err_genc.pl script.
3576 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3578 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3579 not many people have the assembler. Various Win32 compilation fixes and
3580 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3584 *) Modify configure script 'Configure' to automatically create crypto/date.h
3585 file under Win32 and also build pem.h from pem.org. New script
projects / openssl.git / blob