4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Add additional OCSP certificate checks. These are those specified
7 in RFC2560. This consists of two separate checks: the CA of the
8 certificate being checked must either be the OCSP signer certificate
9 or the issuer of the OCSP signer certificate. In the latter case the
10 OCSP signer certificate must contain the OCSP signing extended key
11 usage. This check is performed by attempting to match the OCSP
12 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13 in the OCSP_CERTID structures of the response.
16 *) Initial OCSP certificate verification added to OCSP_basic_verify()
17 and related routines. This uses the standard OpenSSL certificate
18 verify routines to perform initial checks (just CA validity) and
19 to obtain the certificate chain. Then additional checks will be
20 performed on the chain. Currently the root CA is checked to see
21 if it is explicitly trusted for OCSP signing. This is used to set
22 a root CA as a global signing root: that is any certificate that
23 chains to that CA is an acceptable OCSP signing certificate.
26 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
27 extensions from a separate configuration file.
28 As when reading extensions from the main configuration file,
29 the '-extensions ...' option may be used for specifying the
31 [Massimiliano Pala <madwolf@comune.modena.it>]
33 *) Change PKCS12_key_gen_asc() so it can cope with non null
34 terminated strings whose length is passed in the passlen
35 parameter, for example from PEM callbacks. This was done
36 by adding an extra length parameter to asc2uni().
37 [Steve Henson, reported by <oddissey@samsung.co.kr>]
39 *) New OCSP utility. Allows OCSP requests to be generated or
40 read. The request can be sent to a responder and the output
41 parsed, outputed or printed in text form. Not complete yet:
42 still needs to check the OCSP response validity.
45 *) New subcommands for 'openssl ca':
46 'openssl ca -status <serial>' prints the status of the cert with
47 the given serial number (according to the index file).
48 'openssl ca -updatedb' updates the expiry status of certificates
50 [Massimiliano Pala <madwolf@comune.modena.it>]
52 *) New '-newreq-nodes' command option to CA.pl. This is like
53 '-newreq', but calls 'openssl req' with the '-nodes' option
54 so that the resulting key is not encrypted.
55 [Damien Miller <djm@mindrot.org>]
57 *) New configuration for the GNU Hurd.
58 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
60 *) Initial code to implement OCSP basic response verify. This
61 is currently incomplete. Currently just finds the signer's
62 certificate and verifies the signature on the response.
65 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
66 value of OPENSSLDIR. This is available via the new '-d' option
67 to 'openssl version', and is also included in 'openssl version -a'.
70 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
71 call failed, free the DSA structure.
74 *) Allowing defining memory allocation callbacks that will be given
75 file name and line number information in additional arguments
76 (a const char* and an int). The basic functionality remains, as
77 well as the original possibility to just replace malloc(),
78 realloc() and free() by functions that do not know about these
79 additional arguments. To register and find out the current
80 settings for extended allocation functions, the following
81 functions are provided:
83 CRYPTO_set_mem_ex_functions
84 CRYPTO_set_locked_mem_ex_functions
85 CRYPTO_get_mem_ex_functions
86 CRYPTO_get_locked_mem_ex_functions
88 These work the same way as CRYPTO_set_mem_functions and friends.
89 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
90 extended allocation function is enabled.
91 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
92 a conventional allocation function is enabled.
93 [Richard Levitte, Bodo Moeller]
95 *) Fix to uni2asc() to cope with zero length Unicode strings.
96 These are present in some PKCS#12 files.
99 *) Finish off removing the remaining LHASH function pointer casts.
100 There should no longer be any prototype-casting required when using
101 the LHASH abstraction, and any casts that remain are "bugs". See
102 the callback types and macros at the head of lhash.h for details
103 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
106 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
107 If an EGD or PRNGD is running and enough entropy is returned, automatic
108 seeding like with /dev/[u]random will be performed.
109 Positions tried are: /etc/entropy, /var/run/egd-pool.
112 *) Change the Unix RAND_poll() variant to be able to poll several
113 random devices, as specified by DEVRANDOM, until a sufficient amount
114 of data has been collected. We spend at most 10 ms on each file
115 (select timeout) and read in non-blocking mode. DEVRANDOM now
116 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
117 (previously it was just the string "/dev/urandom"), so on typical
118 platforms the 10 ms delay will never occur.
119 Also separate out the Unix variant to its own file, rand_unix.c.
120 For VMS, there's a currently-empty rand_vms.c.
123 *) Move OCSP client related routines to ocsp_cl.c. These
124 provide utility functions which an application needing
125 to issue a request to an OCSP responder and analyse the
126 response will typically need: as opposed to those which an
127 OCSP responder itself would need which will be added later.
129 OCSP_request_sign() signs an OCSP request with an API similar
130 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
131 response. OCSP_response_get1_basic() extracts basic response
132 from response. OCSP_resp_find_status(): finds and extracts status
133 information from an OCSP_CERTID structure (which will be created
134 when the request structure is built). These are built from lower
135 level functions which work on OCSP_SINGLERESP structures but
136 wont normally be used unless the application wishes to examine
137 extensions in the OCSP response for example.
139 Replace nonce routines with a pair of functions.
140 OCSP_request_add1_nonce() adds a nonce value and optionally
141 generates a random value. OCSP_check_nonce() checks the
142 validity of the nonce in an OCSP response.
145 *) Change function OCSP_request_add() to OCSP_request_add0_id().
146 This doesn't copy the supplied OCSP_CERTID and avoids the
147 need to free up the newly created id. Change return type
148 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
149 This can then be used to add extensions to the request.
150 Deleted OCSP_request_new(), since most of its functionality
151 is now in OCSP_REQUEST_new() (and the case insensitive name
152 clash) apart from the ability to set the request name which
153 will be added elsewhere.
156 *) Update OCSP API. Remove obsolete extensions argument from
157 various functions. Extensions are now handled using the new
158 OCSP extension code. New simple OCSP HTTP function which
159 can be used to send requests and parse the response.
162 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
163 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
164 uses the special reorder version of SET OF to sort the attributes
165 and reorder them to match the encoded order. This resolves a long
166 standing problem: a verify on a PKCS7 structure just after signing
167 it used to fail because the attribute order did not match the
168 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
169 it uses the received order. This is necessary to tolerate some broken
170 software that does not order SET OF. This is handled by encoding
171 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
172 to produce the required SET OF.
175 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
176 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
177 files to get correct declarations of the ASN.1 item variables.
180 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
181 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
182 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
183 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
184 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
185 ASN1_ITEM and no wrapper functions.
188 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
189 replace the old function pointer based I/O routines. Change most of
190 the *_d2i_bio() and *_d2i_fp() functions to use these.
193 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
194 lines, recognice more "algorithms" that can be deselected, and make
195 it complain about algorithm deselection that isn't recognised.
198 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
199 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
200 to use new functions. Add NO_ASN1_OLD which can be set to remove
201 some old style ASN1 functions: this can be used to determine if old
202 code will still work when these eventually go away.
205 *) New extension functions for OCSP structures, these follow the
206 same conventions as certificates and CRLs.
209 *) New function X509V3_add1_i2d(). This automatically encodes and
210 adds an extension. Its behaviour can be customised with various
211 flags to append, replace or delete. Various wrappers added for
212 certifcates and CRLs.
215 *) Fix to avoid calling the underlying ASN1 print routine when
216 an extension cannot be parsed. Correct a typo in the
217 OCSP_SERVICELOC extension. Tidy up print OCSP format.
220 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
221 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
222 when writing a 32767 byte record.
223 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
225 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
226 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
228 (RSA objects have a reference count access to which is protected
229 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
230 so they are meant to be shared between threads.)
231 [Bodo Moeller, Geoff Thorpe; original patch submitted by
232 "Reddie, Steven" <Steven.Reddie@ca.com>]
234 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
235 entries for variables.
238 *) Fix a deadlock in CRYPTO_mem_leaks().
241 *) Add functionality to apps/openssl.c for detecting locking
242 problems: As the program is single-threaded, all we have
243 to do is register a locking callback using an array for
244 storing which locks are currently held by the program.
247 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
248 SSL_get_ex_data_X509_STORE_idx(), which is used in
249 ssl_verify_cert_chain() and thus can be called at any time
250 during TLS/SSL handshakes so that thread-safety is essential.
251 Unfortunately, the ex_data design is not at all suited
252 for multi-threaded use, so it probably should be abolished.
255 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
256 [Broadcom, tweaked and integrated by Geoff Thorpe]
258 *) Move common extension printing code to new function
259 X509V3_print_extensions(). Reorganise OCSP print routines and
260 implement some needed OCSP ASN1 functions. Add OCSP extensions.
263 *) New function X509_signature_print() to remove duplication in some
267 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
268 set (this was treated exactly the same as SET OF previously). This
269 is used to reorder the STACK representing the structure to match the
270 encoding. This will be used to get round a problem where a PKCS7
271 structure which was signed could not be verified because the STACK
272 order did not reflect the encoded order.
275 *) Reimplement the OCSP ASN1 module using the new code.
278 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
279 for its ASN1 operations. The old style function pointers still exist
280 for now but they will eventually go away.
283 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
284 completely replaces the old ASN1 functionality with a table driven
285 encoder and decoder which interprets an ASN1_ITEM structure describing
286 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
287 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
288 has also been converted to the new form.
291 *) Change BN_mod_exp_recp so that negative moduli are tolerated
292 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
293 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
297 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
298 of not touching the result's sign bit.
301 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
305 *) Changed the LHASH code to use prototypes for callbacks, and created
306 macros to declare and implement thin (optionally static) functions
307 that provide type-safety and avoid function pointer casting for the
308 type-specific callbacks.
311 *) Use better test patterns in bntest.
314 *) Added Kerberos Cipher Suites to be used with TLS, as written in
316 [Veers Staats <staatsvr@asc.hpc.mil>,
317 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
319 *) rand_win.c fix for Borland C.
322 *) BN_rshift bugfix for n == 0.
325 *) Reformat the FAQ so the different questions and answers can be divided
326 in sections depending on the subject.
329 *) Have the zlib compression code load ZLIB.DLL dynamically under
333 *) New function BN_mod_sqrt for computing square roots modulo a prime
334 (using the probabilistic Tonelli-Shanks algorithm unless
335 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
336 be handled deterministically).
337 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
339 *) Store verify_result within SSL_SESSION also for client side to
340 avoid potential security hole. (Re-used sessions on the client side
341 always resulted in verify_result==X509_V_OK, not using the original
342 result of the server certificate verification.)
345 *) Make BN_mod_inverse faster by explicitly handling small quotients
346 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
347 512 bits], about 30% for larger ones [1024 or 2048 bits].)
350 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
351 SSL3_RT_APPLICATION_DATA, return 0.
352 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
356 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
357 releases, have been re-implemented by renaming the previous
358 implementations of ssl2_read and ssl3_read to ssl2_read_internal
359 and ssl3_read_internal, respectively, and adding 'peek' parameters
360 to them. The new ssl[23]_{read,peek} functions are calls to
361 ssl[23]_read_internal with the 'peek' flag set appropriately.
362 A 'peek' parameter has also been added to ssl3_read_bytes, which
363 does the actual work for ssl3_read_internal.
366 *) New function BN_kronecker.
369 *) Fix BN_gcd so that it works on negative inputs; the result is
370 positive unless both parameters are zero.
371 Previously something reasonably close to an infinite loop was
372 possible because numbers could be growing instead of shrinking
373 in the implementation of Euclid's algorithm.
376 *) Fix BN_is_word() and BN_is_one() macros to take into account the
377 sign of the number in question.
379 Fix BN_is_word(a,w) to work correctly for w == 0.
381 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
382 because its test if the absolute value of 'a' equals 'w'.
383 Note that BN_abs_is_word does *not* handle w == 0 reliably;
384 it exists mostly for use in the implementations of BN_is_zero(),
385 BN_is_one(), and BN_is_word().
388 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
389 the method-specific "init()" handler. Also clean up ex_data after
390 calling the method-specific "finish()" handler. Previously, this was
391 happening the other way round.
394 *) New function BN_swap.
397 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
398 the exponentiation functions are more likely to produce reasonable
399 results on negative inputs.
402 *) Change BN_mod_mul so that the result is always non-negative.
403 Previously, it could be negative if one of the factors was negative;
404 I don't think anyone really wanted that behaviour.
407 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
408 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
409 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
410 and add new functions:
423 These functions always generate non-negative results.
425 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
426 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
428 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
429 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
431 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
433 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
434 was actually never needed) and in BN_mul(). The removal in BN_mul()
435 required a small change in bn_mul_part_recursive() and the addition
436 of the functions bn_cmp_part_words(), bn_sub_part_words() and
437 bn_add_part_words(), which do the same thing as bn_cmp_words(),
438 bn_sub_words() and bn_add_words() except they take arrays with
442 *) In 'openssl passwd', verify passwords read from the terminal
443 unless the '-salt' option is used (which usually means that
444 verification would just waste user's time since the resulting
445 hash is going to be compared with some given password hash)
446 or the new '-noverify' option is used.
448 This is an incompatible change, but it does not affect
449 non-interactive use of 'openssl passwd' (passwords on the command
450 line, '-stdin' option, '-in ...' option) and thus should not
454 *) Remove all references to RSAref, since there's no more need for it.
457 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
458 The previous value, 12, was not always sufficient for BN_mod_exp().
461 *) Make DSO load along a path given through an environment variable
462 (SHLIB_PATH) with shl_load().
465 *) Constify the ENGINE code as a result of BIGNUM constification.
466 Also constify the RSA code and most things related to it. In a
467 few places, most notable in the depth of the ASN.1 code, ugly
468 casts back to non-const were required (to be solved at a later
472 *) Make it so the openssl application has all engines loaded by default.
475 *) Constify the BIGNUM routines a little more.
478 *) Make sure that shared libraries get the internal name engine with
479 the full version number and not just 0. This should mark the
480 shared libraries as not backward compatible. Of course, this should
481 be changed again when we can guarantee backward binary compatibility.
484 *) Add the following functions:
490 ENGINE_load_builtin_engines()
492 That way, an application can itself choose if external engines that
493 are built-in in OpenSSL shall ever be used or not. The benefit is
494 that applications won't have to be linked with libdl or other dso
495 libraries unless it's really needed.
497 Changed 'openssl engine' to load all engines on demand.
498 Changed the engine header files to avoid the duplication of some
499 declarations (they differed!).
502 *) 'openssl engine' can now list capabilities.
505 *) Better error reporting in 'openssl engine'.
508 *) Never call load_dh_param(NULL) in s_server.
511 *) Add engine application. It can currently list engines by name and
512 identity, and test if they are actually available.
515 *) Add support for shared libraries under Irix.
516 [Albert Chin-A-Young <china@thewrittenword.com>]
518 *) Improve RPM specification file by forcing symbolic linking and making
519 sure the installed documentation is also owned by root.root.
520 [Damien Miller <djm@mindrot.org>]
522 *) Add configuration option to build on Linux on both big-endian and
524 [Ralf Baechle <ralf@uni-koblenz.de>]
526 *) Give the OpenSSL applications more possibilities to make use of
527 keys (public as well as private) handled by engines.
530 *) Add OCSP code that comes from CertCo.
533 *) Add VMS support for the Rijndael code.
536 *) Added untested support for Nuron crypto accelerator.
539 *) Add support for external cryptographic devices. This code was
540 previously distributed separately as the "engine" branch.
541 [Geoff Thorpe, Richard Levitte]
543 *) Rework the filename-translation in the DSO code. It is now possible to
544 have far greater control over how a "name" is turned into a filename
545 depending on the operating environment and any oddities about the
546 different shared library filenames on each system.
549 *) Support threads on FreeBSD-elf in Configure.
552 *) Add the possibility to create shared libraries on HP-UX
555 *) Fix for SHA1 assembly problem with MASM: it produces
556 warnings about corrupt line number information when assembling
557 with debugging information. This is caused by the overlapping
559 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
562 NCONF_get_number() has no error checking at all. As a replacement,
563 NCONF_get_number_e() is defined (_e for "error checking") and is
564 promoted strongly. The old NCONF_get_number is kept around for
565 binary backward compatibility.
566 Make it possible for methods to load from something other than a BIO,
567 by providing a function pointer that is given a name instead of a BIO.
568 For example, this could be used to load configuration data from an
572 *) Fix typo in get_cert_by_subject() in by_dir.c
573 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
575 *) Rework the system to generate shared libraries:
577 - Make note of the expected extension for the shared libraries and
578 if there is a need for symbolic links from for example libcrypto.so.0
579 to libcrypto.so.0.9.7. There is extended info in Configure for
582 - Make as few rebuilds of the shared libraries as possible.
584 - Still avoid linking the OpenSSL programs with the shared libraries.
586 - When installing, install the shared libraries separately from the
590 *) Fix for non blocking accept BIOs. Added new I/O special reason
591 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
592 with non blocking I/O was not possible because no retry code was
593 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
597 *) Added the beginnings of Rijndael support.
600 *) Fix for bug in DirectoryString mask setting. Add support for
601 X509_NAME_print_ex() in 'req' and X509_print_ex() function
602 to allow certificate printing to more controllable, additional
603 'certopt' option to 'x509' to allow new printing options to be
607 *) Clean old EAY MD5 hack from e_os.h.
610 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
612 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
613 and not in SSL_clear because the latter is also used by the
614 accept/connect functions; previously, the settings made by
615 SSL_set_read_ahead would be lost during the handshake.
616 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
618 *) Correct util/mkdef.pl to be selective about disabled algorithms.
619 Previously, it would create entries for disableed algorithms no
623 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
625 *) In ssl23_get_client_hello, generate an error message when faced
626 with an initial SSL 3.0/TLS record that is too small to contain the
627 first two bytes of the ClientHello message, i.e. client_version.
628 (Note that this is a pathologic case that probably has never happened
629 in real life.) The previous approach was to use the version number
630 from the record header as a substitute; but our protocol choice
631 should not depend on that one because it is not authenticated
632 by the Finished messages.
635 *) More robust randomness gathering functions for Windows.
636 [Jeffrey Altman <jaltman@columbia.edu>]
638 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
639 not set then we don't setup the error code for issuer check errors
640 to avoid possibly overwriting other errors which the callback does
641 handle. If an application does set the flag then we assume it knows
642 what it is doing and can handle the new informational codes
646 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
647 a general "ANY" type, as such it should be able to decode anything
648 including tagged types. However it didn't check the class so it would
649 wrongly interpret tagged types in the same way as their universal
650 counterpart and unknown types were just rejected. Changed so that the
651 tagged and unknown types are handled in the same way as a SEQUENCE:
652 that is the encoding is stored intact. There is also a new type
653 "V_ASN1_OTHER" which is used when the class is not universal, in this
654 case we have no idea what the actual type is so we just lump them all
658 *) On VMS, stdout may very well lead to a file that is written to
659 in a record-oriented fashion. That means that every write() will
660 write a separate record, which will be read separately by the
661 programs trying to read from it. This can be very confusing.
663 The solution is to put a BIO filter in the way that will buffer
664 text until a linefeed is reached, and then write everything a
665 line at a time, so every record written will be an actual line,
666 not chunks of lines and not (usually doesn't happen, but I've
667 seen it once) several lines in one record. BIO_f_linebuffer() is
670 Currently, it's a VMS-only method, because that's where it has
671 been tested well enough.
674 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
675 it can return incorrect results.
676 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
677 but it was in 0.9.6-beta[12].)
680 *) Disable the check for content being present when verifying detached
681 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
682 include zero length content when signing messages.
685 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
686 BIO_ctrl (for BIO pairs).
689 *) Add DSO method for VMS.
692 *) Bug fix: Montgomery multiplication could produce results with the
696 *) Add RPM specification openssl.spec and modify it to build three
697 packages. The default package contains applications, application
698 documentation and run-time libraries. The devel package contains
699 include files, static libraries and function documentation. The
700 doc package contains the contents of the doc directory. The original
701 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
704 *) Add a large number of documentation files for many SSL routines.
705 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
707 *) Add a configuration entry for Sony News 4.
708 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
710 *) Don't set the two most significant bits to one when generating a
711 random number < q in the DSA library.
714 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
715 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
716 the underlying transport is blocking) if a handshake took place.
717 (The default behaviour is needed by applications such as s_client
718 and s_server that use select() to determine when to use SSL_read;
719 but for applications that know in advance when to expect data, it
720 just makes things more complicated.)
723 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
727 *) Add a few more EBCDIC conditionals that make `req' and `x509'
728 work better on such systems.
729 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
731 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
732 Update PKCS12_parse() so it copies the friendlyName and the
733 keyid to the certificates aux info.
736 *) Fix bug in PKCS7_verify() which caused an infinite loop
737 if there was more than one signature.
738 [Sven Uszpelkat <su@celocom.de>]
740 *) Major change in util/mkdef.pl to include extra information
741 about each symbol, as well as presentig variables as well
742 as functions. This change means that there's n more need
743 to rebuild the .num files when some algorithms are excluded.
746 *) Allow the verify time to be set by an application,
747 rather than always using the current time.
750 *) Phase 2 verify code reorganisation. The certificate
751 verify code now looks up an issuer certificate by a
752 number of criteria: subject name, authority key id
753 and key usage. It also verifies self signed certificates
754 by the same criteria. The main comparison function is
755 X509_check_issued() which performs these checks.
757 Lot of changes were necessary in order to support this
758 without completely rewriting the lookup code.
760 Authority and subject key identifier are now cached.
762 The LHASH 'certs' is X509_STORE has now been replaced
763 by a STACK_OF(X509_OBJECT). This is mainly because an
764 LHASH can't store or retrieve multiple objects with
767 As a result various functions (which were all internal
768 use only) have changed to handle the new X509_STORE
769 structure. This will break anything that messed round
770 with X509_STORE internally.
772 The functions X509_STORE_add_cert() now checks for an
773 exact match, rather than just subject name.
775 The X509_STORE API doesn't directly support the retrieval
776 of multiple certificates matching a given criteria, however
777 this can be worked round by performing a lookup first
778 (which will fill the cache with candidate certificates)
779 and then examining the cache for matches. This is probably
780 the best we can do without throwing out X509_LOOKUP
781 entirely (maybe later...).
783 The X509_VERIFY_CTX structure has been enhanced considerably.
785 All certificate lookup operations now go via a get_issuer()
786 callback. Although this currently uses an X509_STORE it
787 can be replaced by custom lookups. This is a simple way
788 to bypass the X509_STORE hackery necessary to make this
789 work and makes it possible to use more efficient techniques
790 in future. A very simple version which uses a simple
791 STACK for its trusted certificate store is also provided
792 using X509_STORE_CTX_trusted_stack().
794 The verify_cb() and verify() callbacks now have equivalents
795 in the X509_STORE_CTX structure.
797 X509_STORE_CTX also has a 'flags' field which can be used
798 to customise the verify behaviour.
801 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
802 excludes S/MIME capabilities.
805 *) When a certificate request is read in keep a copy of the
806 original encoding of the signed data and use it when outputing
807 again. Signatures then use the original encoding rather than
808 a decoded, encoded version which may cause problems if the
809 request is improperly encoded.
812 *) For consistency with other BIO_puts implementations, call
813 buffer_write(b, ...) directly in buffer_puts instead of calling
816 In BIO_puts, increment b->num_write as in BIO_write.
817 [Peter.Sylvester@EdelWeb.fr]
819 *) Fix BN_mul_word for the case where the word is 0. (We have to use
820 BN_zero, we may not return a BIGNUM with an array consisting of
824 *) Avoid calling abort() from within the library when problems are
825 detected, except if preprocessor symbols have been defined
826 (such as REF_CHECK, BN_DEBUG etc.).
829 *) New openssl application 'rsautl'. This utility can be
830 used for low level RSA operations. DER public key
831 BIO/fp routines also added.
834 *) New Configure entry and patches for compiling on QNX 4.
835 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
837 *) A demo state-machine implementation was sponsored by
838 Nuron (http://www.nuron.com/) and is now available in
842 *) New options added to the 'dgst' utility for signature
843 generation and verification.
846 *) Unrecognized PKCS#7 content types are now handled via a
847 catch all ASN1_TYPE structure. This allows unsupported
848 types to be stored as a "blob" and an application can
849 encode and decode it manually.
852 *) Fix various signed/unsigned issues to make a_strex.c
854 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
856 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
857 length if passed a buffer. ASN1_INTEGER_to_BN failed
858 if passed a NULL BN and its argument was negative.
859 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
861 *) Modification to PKCS#7 encoding routines to output definite
862 length encoding. Since currently the whole structures are in
863 memory there's not real point in using indefinite length
864 constructed encoding. However if OpenSSL is compiled with
865 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
868 *) Added BIO_vprintf() and BIO_vsnprintf().
871 *) Added more prefixes to parse for in the the strings written
872 through a logging bio, to cover all the levels that are available
873 through syslog. The prefixes are now:
875 PANIC, EMERG, EMR => LOG_EMERG
876 ALERT, ALR => LOG_ALERT
877 CRIT, CRI => LOG_CRIT
878 ERROR, ERR => LOG_ERR
879 WARNING, WARN, WAR => LOG_WARNING
880 NOTICE, NOTE, NOT => LOG_NOTICE
881 INFO, INF => LOG_INFO
882 DEBUG, DBG => LOG_DEBUG
884 and as before, if none of those prefixes are present at the
885 beginning of the string, LOG_ERR is chosen.
887 On Win32, the LOG_* levels are mapped according to this:
889 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
890 LOG_WARNING => EVENTLOG_WARNING_TYPE
891 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
895 *) Made it possible to reconfigure with just the configuration
896 argument "reconf" or "reconfigure". The command line arguments
897 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
898 and are retrieved from there when reconfiguring.
902 [Assar Westerlund <assar@sics.se>, Richard Levitte]
904 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
907 *) The obj_dat.pl script was messing up the sorting of object
908 names. The reason was that it compared the quoted version
909 of strings as a result "OCSP" > "OCSP Signing" because
910 " > SPACE. Changed script to store unquoted versions of
911 names and add quotes on output. It was also omitting some
912 names from the lookup table if they were given a default
913 value (that is if SN is missing it is given the same
914 value as LN and vice versa), these are now added on the
915 grounds that if an object has a name we should be able to
916 look it up. Finally added warning output when duplicate
917 short or long names are found.
920 *) Changes needed for Tandem NSK.
921 [Scott Uroff <scott@xypro.com>]
923 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
924 RSA_padding_check_SSLv23(), special padding was never detected
925 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
926 version rollback attacks was not effective.
928 In s23_clnt.c, don't use special rollback-attack detection padding
929 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
930 client; similarly, in s23_srvr.c, don't do the rollback check if
931 SSL 2.0 is the only protocol enabled in the server.
934 *) Make it possible to get hexdumps of unprintable data with 'openssl
935 asn1parse'. By implication, the functions ASN1_parse_dump() and
936 BIO_dump_indent() are added.
939 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
940 these print out strings and name structures based on various
941 flags including RFC2253 support and proper handling of
942 multibyte characters. Added options to the 'x509' utility
943 to allow the various flags to be set.
946 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
947 Also change the functions X509_cmp_current_time() and
948 X509_gmtime_adj() work with an ASN1_TIME structure,
949 this will enable certificates using GeneralizedTime in validity
953 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
954 negative public key encodings) on by default,
955 NO_NEG_PUBKEY_BUG can be set to disable it.
958 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
959 content octets. An i2c_ASN1_OBJECT is unnecessary because
960 the encoding can be trivially obtained from the structure.
963 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
964 not read locks (CRYPTO_r_[un]lock).
967 *) A first attempt at creating official support for shared
968 libraries through configuration. I've kept it so the
969 default is static libraries only, and the OpenSSL programs
970 are always statically linked for now, but there are
971 preparations for dynamic linking in place.
972 This has been tested on Linux and True64.
975 *) Randomness polling function for Win9x, as described in:
976 Peter Gutmann, Software Generation of Practically Strong
980 *) Fix so PRNG is seeded in req if using an already existing
984 *) New options to smime application. -inform and -outform
985 allow alternative formats for the S/MIME message including
986 PEM and DER. The -content option allows the content to be
987 specified separately. This should allow things like Netscape
988 form signing output easier to verify.
991 *) Fix the ASN1 encoding of tags using the 'long form'.
994 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
995 STRING types. These convert content octets to and from the
996 underlying type. The actual tag and length octets are
997 already assumed to have been read in and checked. These
998 are needed because all other string types have virtually
999 identical handling apart from the tag. By having versions
1000 of the ASN1 functions that just operate on content octets
1001 IMPLICIT tagging can be handled properly. It also allows
1002 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1003 and ASN1_INTEGER are identical apart from the tag.
1006 *) Change the handling of OID objects as follows:
1008 - New object identifiers are inserted in objects.txt, following
1009 the syntax given in objects.README.
1010 - objects.pl is used to process obj_mac.num and create a new
1012 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1015 This is currently kind of a hack, and the perl code in objects.pl
1016 isn't very elegant, but it works as I intended. The simplest way
1017 to check that it worked correctly is to look in obj_dat.h and
1018 check the array nid_objs and make sure the objects haven't moved
1019 around (this is important!). Additions are OK, as well as
1020 consistent name changes.
1023 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1026 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1027 The given file adds to whatever has already been seeded into the
1028 random pool through the RANDFILE configuration file option or
1029 environment variable, or the default random state file.
1032 *) mkstack.pl now sorts each macro group into lexical order.
1033 Previously the output order depended on the order the files
1034 appeared in the directory, resulting in needless rewriting
1038 *) Patches to make OpenSSL compile under Win32 again. Mostly
1039 work arounds for the VC++ problem that it treats func() as
1040 func(void). Also stripped out the parts of mkdef.pl that
1041 added extra typesafe functions: these no longer exist.
1044 *) Reorganisation of the stack code. The macros are now all
1045 collected in safestack.h . Each macro is defined in terms of
1046 a "stack macro" of the form SKM_<name>(type, a, b). The
1047 DEBUG_SAFESTACK is now handled in terms of function casts,
1048 this has the advantage of retaining type safety without the
1049 use of additional functions. If DEBUG_SAFESTACK is not defined
1050 then the non typesafe macros are used instead. Also modified the
1051 mkstack.pl script to handle the new form. Needs testing to see
1052 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1053 the default if no major problems. Similar behaviour for ASN1_SET_OF
1054 and PKCS12_STACK_OF.
1057 *) When some versions of IIS use the 'NET' form of private key the
1058 key derivation algorithm is different. Normally MD5(password) is
1059 used as a 128 bit RC4 key. In the modified case
1060 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1061 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1062 as the old Netscape_RSA functions except they have an additional
1063 'sgckey' parameter which uses the modified algorithm. Also added
1064 an -sgckey command line option to the rsa utility. Thanks to
1065 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1066 algorithm to openssl-dev.
1069 *) The evp_local.h macros were using 'c.##kname' which resulted in
1070 invalid expansion on some systems (SCO 5.0.5 for example).
1071 Corrected to 'c.kname'.
1072 [Phillip Porch <root@theporch.com>]
1074 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1075 a STACK of email addresses from a certificate or request, these look
1076 in the subject name and the subject alternative name extensions and
1077 omit any duplicate addresses.
1080 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1081 This makes DSA verification about 2 % faster.
1084 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1085 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1086 plus overhead for 1024 bit moduli).
1087 This makes exponentiations about 0.5 % faster for 1024 bit
1088 exponents (as measured by "openssl speed rsa2048").
1091 *) Rename memory handling macros to avoid conflicts with other
1093 Malloc => OPENSSL_malloc
1094 Malloc_locked => OPENSSL_malloc_locked
1095 Realloc => OPENSSL_realloc
1096 Free => OPENSSL_free
1099 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1100 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1103 *) CygWin32 support.
1104 [John Jarvie <jjarvie@newsguy.com>]
1106 *) The type-safe stack code has been rejigged. It is now only compiled
1107 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1108 by default all type-specific stack functions are "#define"d back to
1109 standard stack functions. This results in more streamlined output
1110 but retains the type-safety checking possibilities of the original
1114 *) The STACK code has been cleaned up, and certain type declarations
1115 that didn't make a lot of sense have been brought in line. This has
1116 also involved a cleanup of sorts in safestack.h to more correctly
1117 map type-safe stack functions onto their plain stack counterparts.
1118 This work has also resulted in a variety of "const"ifications of
1119 lots of the code, especially "_cmp" operations which should normally
1120 be prototyped with "const" parameters anyway.
1123 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1124 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1125 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1126 where all of 'md' is used each time the PRNG is used, but 'state'
1127 is used only indexed by a cyclic counter. As entropy may not be
1128 well distributed from the beginning, 'md' is important as a
1129 chaining variable. However, the output function chains only half
1130 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1131 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1132 in all of 'state' being rewritten, with the new values depending
1133 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1136 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1137 the handshake is continued after ssl_verify_cert_chain();
1138 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1139 can lead to 'unexplainable' connection aborts later.
1140 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1142 *) Major EVP API cipher revision.
1143 Add hooks for extra EVP features. This allows various cipher
1144 parameters to be set in the EVP interface. Support added for variable
1145 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1146 setting of RC2 and RC5 parameters.
1148 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1151 Remove lots of duplicated code from the EVP library. For example *every*
1152 cipher init() function handles the 'iv' in the same way according to the
1153 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1154 for CFB and OFB modes they zero ctx->num.
1156 New functionality allows removal of S/MIME code RC2 hack.
1158 Most of the routines have the same form and so can be declared in terms
1161 By shifting this to the top level EVP_CipherInit() it can be removed from
1162 all individual ciphers. If the cipher wants to handle IVs or keys
1163 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1166 Change lots of functions like EVP_EncryptUpdate() to now return a
1167 value: although software versions of the algorithms cannot fail
1168 any installed hardware versions can.
1171 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1172 this option is set, tolerate broken clients that send the negotiated
1173 protocol version number instead of the requested protocol version
1177 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1178 i.e. non-zero for export ciphersuites, zero otherwise.
1179 Previous versions had this flag inverted, inconsistent with
1180 rsa_tmp_cb (..._TMP_RSA_CB).
1181 [Bodo Moeller; problem reported by Amit Chopra]
1183 *) Add missing DSA library text string. Work around for some IIS
1184 key files with invalid SEQUENCE encoding.
1187 *) Add a document (doc/standards.txt) that list all kinds of standards
1188 and so on that are implemented in OpenSSL.
1191 *) Enhance c_rehash script. Old version would mishandle certificates
1192 with the same subject name hash and wouldn't handle CRLs at all.
1193 Added -fingerprint option to crl utility, to support new c_rehash
1197 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1200 *) Fix for SSL server purpose checking. Server checking was
1201 rejecting certificates which had extended key usage present
1202 but no ssl client purpose.
1203 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1205 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1206 is a little unclear about how a blank password is handled.
1207 Since the password in encoded as a BMPString with terminating
1208 double NULL a zero length password would end up as just the
1209 double NULL. However no password at all is different and is
1210 handled differently in the PKCS#12 key generation code. NS
1211 treats a blank password as zero length. MSIE treats it as no
1212 password on export: but it will try both on import. We now do
1213 the same: PKCS12_parse() tries zero length and no password if
1214 the password is set to "" or NULL (NULL is now a valid password:
1215 it wasn't before) as does the pkcs12 application.
1218 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1219 perror when PEM_read_bio_X509_REQ fails, the error message must
1220 be obtained from the error queue.
1223 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1224 it in ERR_remove_state if appropriate, and change ERR_get_state
1225 accordingly to avoid race conditions (this is necessary because
1226 thread_hash is no longer constant once set).
1229 *) Bugfix for linux-elf makefile.one.
1232 *) RSA_get_default_method() will now cause a default
1233 RSA_METHOD to be chosen if one doesn't exist already.
1234 Previously this was only set during a call to RSA_new()
1235 or RSA_new_method(NULL) meaning it was possible for
1236 RSA_get_default_method() to return NULL.
1239 *) Added native name translation to the existing DSO code
1240 that will convert (if the flag to do so is set) filenames
1241 that are sufficiently small and have no path information
1242 into a canonical native form. Eg. "blah" converted to
1243 "libblah.so" or "blah.dll" etc.
1246 *) New function ERR_error_string_n(e, buf, len) which is like
1247 ERR_error_string(e, buf), but writes at most 'len' bytes
1248 including the 0 terminator. For ERR_error_string_n, 'buf'
1250 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1252 *) CONF library reworked to become more general. A new CONF
1253 configuration file reader "class" is implemented as well as a
1254 new functions (NCONF_*, for "New CONF") to handle it. The now
1255 old CONF_* functions are still there, but are reimplemented to
1256 work in terms of the new functions. Also, a set of functions
1257 to handle the internal storage of the configuration data is
1258 provided to make it easier to write new configuration file
1259 reader "classes" (I can definitely see something reading a
1260 configuration file in XML format, for example), called _CONF_*,
1261 or "the configuration storage API"...
1263 The new configuration file reading functions are:
1265 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1266 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1268 NCONF_default, NCONF_WIN32
1270 NCONF_dump_fp, NCONF_dump_bio
1272 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1273 NCONF_new creates a new CONF object. This works in the same way
1274 as other interfaces in OpenSSL, like the BIO interface.
1275 NCONF_dump_* dump the internal storage of the configuration file,
1276 which is useful for debugging. All other functions take the same
1277 arguments as the old CONF_* functions wth the exception of the
1278 first that must be a `CONF *' instead of a `LHASH *'.
1280 To make it easer to use the new classes with the old CONF_* functions,
1281 the function CONF_set_default_method is provided.
1284 *) Add '-tls1' option to 'openssl ciphers', which was already
1285 mentioned in the documentation but had not been implemented.
1286 (This option is not yet really useful because even the additional
1287 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1290 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1291 OpenSSL-based applications) load shared libraries and bind to
1292 them in a portable way.
1293 [Geoff Thorpe, with contributions from Richard Levitte]
1295 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1297 *) Make sure _lrotl and _lrotr are only used with MSVC.
1299 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1300 (the default implementation of RAND_status).
1302 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1303 to '-clrext' (= clear extensions), as intended and documented.
1304 [Bodo Moeller; inconsistency pointed out by Michael Attili
1307 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1308 was larger than the MD block size.
1309 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1311 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1312 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1313 using the passed key: if the passed key was a private key the result
1314 of X509_print(), for example, would be to print out all the private key
1318 *) des_quad_cksum() byte order bug fix.
1319 [Ulf Möller, using the problem description in krb4-0.9.7, where
1320 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1322 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1324 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1326 *) For easily testing in shell scripts whether some command
1327 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1328 returns with exit code 0 iff no command of the given name is available.
1329 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1330 the output goes to stdout and nothing is printed to stderr.
1331 Additional arguments are always ignored.
1333 Since for each cipher there is a command of the same name,
1334 the 'no-cipher' compilation switches can be tested this way.
1336 ('openssl no-XXX' is not able to detect pseudo-commands such
1337 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1340 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1343 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1344 is set; it will be thrown away anyway because each handshake creates
1346 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1347 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1348 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1349 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1352 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1353 'Q' and 'R' lose their special meanings (quit/renegotiate).
1354 This is part of what -quiet does; unlike -quiet, -ign_eof
1355 does not suppress any output.
1358 *) Add compatibility options to the purpose and trust code. The
1359 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1360 accepts a certificate or CA, this was the previous behaviour,
1361 with all the associated security issues.
1363 X509_TRUST_COMPAT is the old trust behaviour: only and
1364 automatically trust self signed roots in certificate store. A
1365 new trust setting X509_TRUST_DEFAULT is used to specify that
1366 a purpose has no associated trust setting and it should instead
1367 use the value in the default purpose.
1370 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1371 and fix a memory leak.
1374 *) In util/mkerr.pl (which implements 'make errors'), preserve
1375 reason strings from the previous version of the .c file, as
1376 the default to have only downcase letters (and digits) in
1377 automatically generated reasons codes is not always appropriate.
1380 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1381 using strerror. Previously, ERR_reason_error_string() returned
1382 library names as reason strings for SYSerr; but SYSerr is a special
1383 case where small numbers are errno values, not library numbers.
1386 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1387 converts DSA parameters into DH parameters. (When creating parameters,
1388 DSA_generate_parameters is used.)
1391 *) Include 'length' (recommended exponent length) in C code generated
1392 by 'openssl dhparam -C'.
1395 *) The second argument to set_label in perlasm was already being used
1396 so couldn't be used as a "file scope" flag. Moved to third argument
1400 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1401 instead of RAND_bytes for encryption IVs and salts.
1404 *) Include RAND_status() into RAND_METHOD instead of implementing
1405 it only for md_rand.c Otherwise replacing the PRNG by calling
1406 RAND_set_rand_method would be impossible.
1409 *) Don't let DSA_generate_key() enter an infinite loop if the random
1410 number generation fails.
1413 *) New 'rand' application for creating pseudo-random output.
1416 *) Added configuration support for Linux/IA64
1417 [Rolf Haberrecker <rolf@suse.de>]
1419 *) Assembler module support for Mingw32.
1422 *) Shared library support for HPUX (in shlib/).
1423 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1425 *) Shared library support for Solaris gcc.
1426 [Lutz Behnke <behnke@trustcenter.de>]
1428 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1430 *) PKCS7_encrypt() was adding text MIME headers twice because they
1431 were added manually and by SMIME_crlf_copy().
1434 *) In bntest.c don't call BN_rand with zero bits argument.
1435 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1437 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1438 case was implemented. This caused BN_div_recp() to fail occasionally.
1441 *) Add an optional second argument to the set_label() in the perl
1442 assembly language builder. If this argument exists and is set
1443 to 1 it signals that the assembler should use a symbol whose
1444 scope is the entire file, not just the current function. This
1445 is needed with MASM which uses the format label:: for this scope.
1446 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1448 *) Change the ASN1 types so they are typedefs by default. Before
1449 almost all types were #define'd to ASN1_STRING which was causing
1450 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1454 *) Change names of new functions to the new get1/get0 naming
1455 convention: After 'get1', the caller owns a reference count
1456 and has to call ..._free; 'get0' returns a pointer to some
1457 data structure without incrementing reference counters.
1458 (Some of the existing 'get' functions increment a reference
1459 counter, some don't.)
1460 Similarly, 'set1' and 'add1' functions increase reference
1461 counters or duplicate objects.
1464 *) Allow for the possibility of temp RSA key generation failure:
1465 the code used to assume it always worked and crashed on failure.
1468 *) Fix potential buffer overrun problem in BIO_printf().
1469 [Ulf Möller, using public domain code by Patrick Powell; problem
1470 pointed out by David Sacerdote <das33@cornell.edu>]
1472 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1473 RAND_egd() and RAND_status(). In the command line application,
1474 the EGD socket can be specified like a seed file using RANDFILE
1478 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1479 Some CAs (e.g. Verisign) distribute certificates in this form.
1482 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1483 list to exclude them. This means that no special compilation option
1484 is needed to use anonymous DH: it just needs to be included in the
1488 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1489 EVP_MD_type. The old functionality is available in a new macro called
1490 EVP_MD_md(). Change code that uses it and update docs.
1493 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1494 where the 'void *' argument is replaced by a function pointer argument.
1495 Previously 'void *' was abused to point to functions, which works on
1496 many platforms, but is not correct. As these functions are usually
1497 called by macros defined in OpenSSL header files, most source code
1498 should work without changes.
1501 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1502 sections with information on -D... compiler switches used for
1503 compiling the library so that applications can see them. To enable
1504 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1505 must be defined. E.g.,
1506 #define OPENSSL_ALGORITHM_DEFINES
1507 #include <openssl/opensslconf.h>
1508 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1509 [Richard Levitte, Ulf and Bodo Möller]
1511 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1515 *) Change the 'other' type in certificate aux info to a STACK_OF
1516 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1517 the required ASN1 format: arbitrary types determined by an OID.
1520 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1521 argument to 'req'. This is not because the function is newer or
1522 better than others it just uses the work 'NEW' in the certificate
1523 request header lines. Some software needs this.
1526 *) Reorganise password command line arguments: now passwords can be
1527 obtained from various sources. Delete the PEM_cb function and make
1528 it the default behaviour: i.e. if the callback is NULL and the
1529 usrdata argument is not NULL interpret it as a null terminated pass
1530 phrase. If usrdata and the callback are NULL then the pass phrase
1531 is prompted for as usual.
1534 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1535 the support is automatically enabled. The resulting binaries will
1536 autodetect the card and use it if present.
1537 [Ben Laurie and Compaq Inc.]
1539 *) Work around for Netscape hang bug. This sends certificate request
1540 and server done in one record. Since this is perfectly legal in the
1541 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1542 the bugs/SSLv3 entry for more info.
1545 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1548 *) Add -rand argument to smime and pkcs12 applications and read/write
1552 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1555 *) Add command line password options to the remaining applications.
1558 *) Bug fix for BN_div_recp() for numerators with an even number of
1562 *) More tests in bntest.c, and changed test_bn output.
1565 *) ./config recognizes MacOS X now.
1568 *) Bug fix for BN_div() when the first words of num and divsor are
1569 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1572 *) Add support for various broken PKCS#8 formats, and command line
1573 options to produce them.
1576 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1577 get temporary BIGNUMs from a BN_CTX.
1580 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1584 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1585 include a #define from the old name to the new. The original intent
1586 was that statically linked binaries could for example just call
1587 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1588 link with digests. This never worked becayse SSLeay_add_all_digests()
1589 and SSLeay_add_all_ciphers() were in the same source file so calling
1590 one would link with the other. They are now in separate source files.
1593 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1596 *) Use a less unusual form of the Miller-Rabin primality test (it used
1597 a binary algorithm for exponentiation integrated into the Miller-Rabin
1598 loop, our standard modexp algorithms are faster).
1601 *) Support for the EBCDIC character set completed.
1602 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1604 *) Source code cleanups: use const where appropriate, eliminate casts,
1605 use void * instead of char * in lhash.
1608 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1609 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1610 this the server could overwrite ephemeral keys that the client
1614 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1615 using 50 iterations of the Rabin-Miller test.
1617 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1618 iterations of the Rabin-Miller test as required by the appendix
1619 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1620 As BN_is_prime_fasttest includes trial division, DSA parameter
1621 generation becomes much faster.
1623 This implies a change for the callback functions in DSA_is_prime
1624 and DSA_generate_parameters: The callback function is called once
1625 for each positive witness in the Rabin-Miller test, not just
1626 occasionally in the inner loop; and the parameters to the
1627 callback function now provide an iteration count for the outer
1628 loop rather than for the current invocation of the inner loop.
1629 DSA_generate_parameters additionally can call the callback
1630 function with an 'iteration count' of -1, meaning that a
1631 candidate has passed the trial division test (when q is generated
1632 from an application-provided seed, trial division is skipped).
1635 *) New function BN_is_prime_fasttest that optionally does trial
1636 division before starting the Rabin-Miller test and has
1637 an additional BN_CTX * argument (whereas BN_is_prime always
1638 has to allocate at least one BN_CTX).
1639 'callback(1, -1, cb_arg)' is called when a number has passed the
1640 trial division stage.
1643 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1647 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1650 *) New function BN_pseudo_rand().
1653 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1654 bignum version of BN_from_montgomery() with the working code from
1655 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1659 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1660 made it impossible to use the same SSL_SESSION data structure in
1661 SSL2 clients in multiple threads.
1664 *) The return value of RAND_load_file() no longer counts bytes obtained
1665 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1666 to seed the PRNG (previously an explicit byte count was required).
1667 [Ulf Möller, Bodo Möller]
1669 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1670 used (char *) instead of (void *) and had casts all over the place.
1673 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1676 *) Retain source code compatibility for BN_prime_checks macro:
1677 BN_is_prime(..., BN_prime_checks, ...) now uses
1678 BN_prime_checks_for_size to determine the appropriate number of
1679 Rabin-Miller iterations.
1682 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1683 DH_CHECK_P_NOT_SAFE_PRIME.
1684 (Check if this is true? OpenPGP calls them "strong".)
1687 *) Merge the functionality of "dh" and "gendh" programs into a new program
1688 "dhparam". The old programs are retained for now but will handle DH keys
1689 (instead of parameters) in future.
1692 *) Make the ciphers, s_server and s_client programs check the return values
1693 when a new cipher list is set.
1696 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1697 ciphers. Before when the 56bit ciphers were enabled the sorting was
1700 The syntax for the cipher sorting has been extended to support sorting by
1701 cipher-strength (using the strength_bits hard coded in the tables).
1702 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1704 Fix a bug in the cipher-command parser: when supplying a cipher command
1705 string with an "undefined" symbol (neither command nor alphanumeric
1706 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1707 an error is flagged.
1709 Due to the strength-sorting extension, the code of the
1710 ssl_create_cipher_list() function was completely rearranged. I hope that
1711 the readability was also increased :-)
1712 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1714 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1715 for the first serial number and places 2 in the serial number file. This
1716 avoids problems when the root CA is created with serial number zero and
1717 the first user certificate has the same issuer name and serial number
1721 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1722 the new code. Add documentation for this stuff.
1725 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1726 X509_*() to X509at_*() on the grounds that they don't handle X509
1727 structures and behave in an analagous way to the X509v3 functions:
1728 they shouldn't be called directly but wrapper functions should be used
1731 So we also now have some wrapper functions that call the X509at functions
1732 when passed certificate requests. (TO DO: similar things can be done with
1733 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1734 things. Some of these need some d2i or i2d and print functionality
1735 because they handle more complex structures.)
1738 *) Add missing #ifndefs that caused missing symbols when building libssl
1739 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1740 NO_RSA in ssl/s2*.c.
1741 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1743 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1744 has a return value which indicates the quality of the random data
1745 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1746 error queue. New function RAND_pseudo_bytes() generates output that is
1747 guaranteed to be unique but not unpredictable. RAND_add is like
1748 RAND_seed, but takes an extra argument for an entropy estimate
1749 (RAND_seed always assumes full entropy).
1752 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1753 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1754 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1755 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1756 false-positive rate of at most 2^-80 for random input.
1759 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1762 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1763 in the 0.9.5 release), this returns the chain
1764 from an X509_CTX structure with a dup of the stack and all
1765 the X509 reference counts upped: so the stack will exist
1766 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1769 Also make SSL_SESSION_print() print out the verify return
1773 *) Add manpage for the pkcs12 command. Also change the default
1774 behaviour so MAC iteration counts are used unless the new
1775 -nomaciter option is used. This improves file security and
1776 only older versions of MSIE (4.0 for example) need it.
1779 *) Honor the no-xxx Configure options when creating .DEF files.
1782 *) Add PKCS#10 attributes to field table: challengePassword,
1783 unstructuredName and unstructuredAddress. These are taken from
1784 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1785 international characters are used.
1787 More changes to X509_ATTRIBUTE code: allow the setting of types
1788 based on strings. Remove the 'loc' parameter when adding
1789 attributes because these will be a SET OF encoding which is sorted
1793 *) Initial changes to the 'req' utility to allow request generation
1794 automation. This will allow an application to just generate a template
1795 file containing all the field values and have req construct the
1798 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1799 used all over the place including certificate requests and PKCS#7
1800 structures. They are currently handled manually where necessary with
1801 some primitive wrappers for PKCS#7. The new functions behave in a
1802 manner analogous to the X509 extension functions: they allow
1803 attributes to be looked up by NID and added.
1805 Later something similar to the X509V3 code would be desirable to
1806 automatically handle the encoding, decoding and printing of the
1807 more complex types. The string types like challengePassword can
1808 be handled by the string table functions.
1810 Also modified the multi byte string table handling. Now there is
1811 a 'global mask' which masks out certain types. The table itself
1812 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1813 is useful when for example there is only one permissible type
1814 (as in countryName) and using the mask might result in no valid
1818 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1819 SSL_get_peer_finished to allow applications to obtain the latest
1820 Finished messages sent to the peer or expected from the peer,
1821 respectively. (SSL_get_peer_finished is usually the Finished message
1822 actually received from the peer, otherwise the protocol will be aborted.)
1824 As the Finished message are message digests of the complete handshake
1825 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1826 be used for external authentication procedures when the authentication
1827 provided by SSL/TLS is not desired or is not enough.
1830 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1831 the host supports BWX extension and if Compaq C is present on the
1832 $PATH. Just exploiting of the BWX extension results in 20-30%
1833 performance kick for some algorithms, e.g. DES and RC4 to mention
1834 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1838 *) Add support for MS "fast SGC". This is arguably a violation of the
1839 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1840 weak crypto and after checking the certificate is SGC a second one
1841 with strong crypto. MS SGC stops the first handshake after receiving
1842 the server certificate message and sends a second client hello. Since
1843 a server will typically do all the time consuming operations before
1844 expecting any further messages from the client (server key exchange
1845 is the most expensive) there is little difference between the two.
1847 To get OpenSSL to support MS SGC we have to permit a second client
1848 hello message after we have sent server done. In addition we have to
1849 reset the MAC if we do get this second client hello.
1852 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1853 if a DER encoded private key is RSA or DSA traditional format. Changed
1854 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1855 format DER encoded private key. Newer code should use PKCS#8 format which
1856 has the key type encoded in the ASN1 structure. Added DER private key
1857 support to pkcs8 application.
1860 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1861 ciphersuites has been selected (as required by the SSL 3/TLS 1
1862 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1863 is set, we interpret this as a request to violate the specification
1864 (the worst that can happen is a handshake failure, and 'correct'
1865 behaviour would result in a handshake failure anyway).
1868 *) In SSL_CTX_add_session, take into account that there might be multiple
1869 SSL_SESSION structures with the same session ID (e.g. when two threads
1870 concurrently obtain them from an external cache).
1871 The internal cache can handle only one SSL_SESSION with a given ID,
1872 so if there's a conflict, we now throw out the old one to achieve
1876 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1877 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1878 some routines that use cipher OIDs: some ciphers do not have OIDs
1879 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1883 *) Simplify the trust setting structure and code. Now we just have
1884 two sequences of OIDs for trusted and rejected settings. These will
1885 typically have values the same as the extended key usage extension
1886 and any application specific purposes.
1888 The trust checking code now has a default behaviour: it will just
1889 check for an object with the same NID as the passed id. Functions can
1890 be provided to override either the default behaviour or the behaviour
1891 for a given id. SSL client, server and email already have functions
1892 in place for compatibility: they check the NID and also return "trusted"
1893 if the certificate is self signed.
1896 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1897 traditional format into an EVP_PKEY structure.
1900 *) Add a password callback function PEM_cb() which either prompts for
1901 a password if usr_data is NULL or otherwise assumes it is a null
1902 terminated password. Allow passwords to be passed on command line
1903 environment or config files in a few more utilities.
1906 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1907 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1908 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1909 Update documentation.
1912 *) Support for ASN1 "NULL" type. This could be handled before by using
1913 ASN1_TYPE but there wasn't any function that would try to read a NULL
1914 and produce an error if it couldn't. For compatibility we also have
1915 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1916 don't allocate anything because they don't need to.
1919 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1921 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1923 *) Rebuild of the memory allocation routines used by OpenSSL code and
1924 possibly others as well. The purpose is to make an interface that
1925 provide hooks so anyone can build a separate set of allocation and
1926 deallocation routines to be used by OpenSSL, for example memory
1927 pool implementations, or something else, which was previously hard
1928 since Malloc(), Realloc() and Free() were defined as macros having
1929 the values malloc, realloc and free, respectively (except for Win32
1930 compilations). The same is provided for memory debugging code.
1931 OpenSSL already comes with functionality to find memory leaks, but
1932 this gives people a chance to debug other memory problems.
1934 With these changes, a new set of functions and macros have appeared:
1936 CRYPTO_set_mem_debug_functions() [F]
1937 CRYPTO_get_mem_debug_functions() [F]
1938 CRYPTO_dbg_set_options() [F]
1939 CRYPTO_dbg_get_options() [F]
1940 CRYPTO_malloc_debug_init() [M]
1942 The memory debug functions are NULL by default, unless the library
1943 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1944 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1945 gives the standard debugging functions that come with OpenSSL) or
1946 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1947 provided by the library user) must be used. When the standard
1948 debugging functions are used, CRYPTO_dbg_set_options can be used to
1949 request additional information:
1950 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1951 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1953 Also, things like CRYPTO_set_mem_functions will always give the
1954 expected result (the new set of functions is used for allocation
1955 and deallocation) at all times, regardless of platform and compiler
1958 To finish it up, some functions that were never use in any other
1959 way than through macros have a new API and new semantic:
1962 CRYPTO_dbg_realloc()
1965 All macros of value have retained their old syntax.
1966 [Richard Levitte and Bodo Moeller]
1968 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1969 ordering of SMIMECapabilities wasn't in "strength order" and there
1970 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1974 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1975 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1976 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1978 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1979 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1980 functionality to handle multipart/signed properly) and a utility
1981 called 'smime' to call all this stuff. This is based on code I
1982 originally wrote for Celo who have kindly allowed it to be
1983 included in OpenSSL.
1986 *) Add variants des_set_key_checked and des_set_key_unchecked of
1987 des_set_key (aka des_key_sched). Global variable des_check_key
1988 decides which of these is called by des_set_key; this way
1989 des_check_key behaves as it always did, but applications and
1990 the library itself, which was buggy for des_check_key == 1,
1991 have a cleaner way to pick the version they need.
1994 *) New function PKCS12_newpass() which changes the password of a
1998 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
1999 dynamic mix. In both cases the ids can be used as an index into the
2000 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2001 functions so they accept a list of the field values and the
2002 application doesn't need to directly manipulate the X509_TRUST
2006 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2010 *) Modify the way the V3 extension code looks up extensions. This now
2011 works in a similar way to the object code: we have some "standard"
2012 extensions in a static table which is searched with OBJ_bsearch()
2013 and the application can add dynamic ones if needed. The file
2014 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2015 updated whenever a new extension is added to the core code and kept
2016 in ext_nid order. There is a simple program 'tabtest.c' which checks
2017 this. New extensions are not added too often so this file can readily
2018 be maintained manually.
2020 There are two big advantages in doing things this way. The extensions
2021 can be looked up immediately and no longer need to be "added" using
2022 X509V3_add_standard_extensions(): this function now does nothing.
2023 [Side note: I get *lots* of email saying the extension code doesn't
2024 work because people forget to call this function]
2025 Also no dynamic allocation is done unless new extensions are added:
2026 so if we don't add custom extensions there is no need to call
2027 X509V3_EXT_cleanup().
2030 *) Modify enc utility's salting as follows: make salting the default. Add a
2031 magic header, so unsalted files fail gracefully instead of just decrypting
2032 to garbage. This is because not salting is a big security hole, so people
2033 should be discouraged from doing it.
2036 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2037 digest to be passed on the command line but it only used this
2038 parameter when signing a certificate. Modified so all relevant
2039 operations are affected by the digest parameter including the
2040 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2041 DSA key was used because it didn't fix the digest.
2044 *) Initial certificate chain verify code. Currently tests the untrusted
2045 certificates for consistency with the verify purpose (which is set
2046 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2048 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2049 this is because it will reject chains with invalid extensions whereas
2050 every previous version of OpenSSL and SSLeay made no checks at all.
2052 Trust code: checks the root CA for the relevant trust settings. Trust
2053 settings have an initial value consistent with the verify purpose: e.g.
2054 if the verify purpose is for SSL client use it expects the CA to be
2055 trusted for SSL client use. However the default value can be changed to
2056 permit custom trust settings: one example of this would be to only trust
2057 certificates from a specific "secure" set of CAs.
2059 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2060 which should be used for version portability: especially since the
2061 verify structure is likely to change more often now.
2063 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2064 to set them. If not set then assume SSL clients will verify SSL servers
2067 Two new options to the verify program: -untrusted allows a set of
2068 untrusted certificates to be passed in and -purpose which sets the
2069 intended purpose of the certificate. If a purpose is set then the
2070 new chain verify code is used to check extension consistency.
2073 *) Support for the authority information access extension.
2076 *) Modify RSA and DSA PEM read routines to transparently handle
2077 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2078 public keys in a format compatible with certificate
2079 SubjectPublicKeyInfo structures. Unfortunately there were already
2080 functions called *_PublicKey_* which used various odd formats so
2081 these are retained for compatibility: however the DSA variants were
2082 never in a public release so they have been deleted. Changed dsa/rsa
2083 utilities to handle the new format: note no releases ever handled public
2084 keys so we should be OK.
2086 The primary motivation for this change is to avoid the same fiasco
2087 that dogs private keys: there are several incompatible private key
2088 formats some of which are standard and some OpenSSL specific and
2089 require various evil hacks to allow partial transparent handling and
2090 even then it doesn't work with DER formats. Given the option anything
2091 other than PKCS#8 should be dumped: but the other formats have to
2092 stay in the name of compatibility.
2094 With public keys and the benefit of hindsight one standard format
2095 is used which works with EVP_PKEY, RSA or DSA structures: though
2096 it clearly returns an error if you try to read the wrong kind of key.
2098 Added a -pubkey option to the 'x509' utility to output the public key.
2099 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2100 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2101 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2102 that do the same as the EVP_PKEY_assign_*() except they up the
2103 reference count of the added key (they don't "swallow" the
2107 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2108 CRLs would fail if the file contained no certificates or no CRLs:
2109 added a new function to read in both types and return the number
2110 read: this means that if none are read it will be an error. The
2111 DER versions of the certificate and CRL reader would always fail
2112 because it isn't possible to mix certificates and CRLs in DER format
2113 without choking one or the other routine. Changed this to just read
2114 a certificate: this is the best we can do. Also modified the code
2115 in apps/verify.c to take notice of return codes: it was previously
2116 attempting to read in certificates from NULL pointers and ignoring
2117 any errors: this is one reason why the cert and CRL reader seemed
2118 to work. It doesn't check return codes from the default certificate
2119 routines: these may well fail if the certificates aren't installed.
2122 *) Code to support otherName option in GeneralName.
2125 *) First update to verify code. Change the verify utility
2126 so it warns if it is passed a self signed certificate:
2127 for consistency with the normal behaviour. X509_verify
2128 has been modified to it will now verify a self signed
2129 certificate if *exactly* the same certificate appears
2130 in the store: it was previously impossible to trust a
2131 single self signed certificate. This means that:
2132 openssl verify ss.pem
2133 now gives a warning about a self signed certificate but
2134 openssl verify -CAfile ss.pem ss.pem
2138 *) For servers, store verify_result in SSL_SESSION data structure
2139 (and add it to external session representation).
2140 This is needed when client certificate verifications fails,
2141 but an application-provided verification callback (set by
2142 SSL_CTX_set_cert_verify_callback) allows accepting the session
2143 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2144 but returns 1): When the session is reused, we have to set
2145 ssl->verify_result to the appropriate error code to avoid
2147 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2149 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2150 case in PKCS7_dataInit() where the signed PKCS7 structure
2151 didn't contain any existing data because it was being created.
2152 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2154 *) Add a salt to the key derivation routines in enc.c. This
2155 forms the first 8 bytes of the encrypted file. Also add a
2156 -S option to allow a salt to be input on the command line.
2159 *) New function X509_cmp(). Oddly enough there wasn't a function
2160 to compare two certificates. We do this by working out the SHA1
2161 hash and comparing that. X509_cmp() will be needed by the trust
2165 *) SSL_get1_session() is like SSL_get_session(), but increments
2166 the reference count in the SSL_SESSION returned.
2167 [Geoff Thorpe <geoff@eu.c2.net>]
2169 *) Fix for 'req': it was adding a null to request attributes.
2170 Also change the X509_LOOKUP and X509_INFO code to handle
2171 certificate auxiliary information.
2174 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2178 *) Add the possibility to add extra information to the memory leak
2179 detecting output, to form tracebacks, showing from where each
2180 allocation was originated: CRYPTO_push_info("constant string") adds
2181 the string plus current file name and line number to a per-thread
2182 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2183 is like calling CYRPTO_pop_info() until the stack is empty.
2184 Also updated memory leak detection code to be multi-thread-safe.
2187 *) Add options -text and -noout to pkcs7 utility and delete the
2188 encryption options which never did anything. Update docs.
2191 *) Add options to some of the utilities to allow the pass phrase
2192 to be included on either the command line (not recommended on
2193 OSes like Unix) or read from the environment. Update the
2194 manpages and fix a few bugs.
2197 *) Add a few manpages for some of the openssl commands.
2200 *) Fix the -revoke option in ca. It was freeing up memory twice,
2201 leaking and not finding already revoked certificates.
2204 *) Extensive changes to support certificate auxiliary information.
2205 This involves the use of X509_CERT_AUX structure and X509_AUX
2206 functions. An X509_AUX function such as PEM_read_X509_AUX()
2207 can still read in a certificate file in the usual way but it
2208 will also read in any additional "auxiliary information". By
2209 doing things this way a fair degree of compatibility can be
2210 retained: existing certificates can have this information added
2211 using the new 'x509' options.
2213 Current auxiliary information includes an "alias" and some trust
2214 settings. The trust settings will ultimately be used in enhanced
2215 certificate chain verification routines: currently a certificate
2216 can only be trusted if it is self signed and then it is trusted
2220 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2221 The problem was that one of the replacement routines had not been working
2222 since SSLeay releases. For now the offending routine has been replaced
2223 with non-optimised assembler. Even so, this now gives around 95%
2224 performance improvement for 1024 bit RSA signs.
2227 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2228 handling. Most clients have the effective key size in bits equal to
2229 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2230 A few however don't do this and instead use the size of the decrypted key
2231 to determine the RC2 key length and the AlgorithmIdentifier to determine
2232 the effective key length. In this case the effective key length can still
2233 be 40 bits but the key length can be 168 bits for example. This is fixed
2234 by manually forcing an RC2 key into the EVP_PKEY structure because the
2235 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2236 the key length and effective key length are equal.
2239 *) Add a bunch of functions that should simplify the creation of
2240 X509_NAME structures. Now you should be able to do:
2241 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2242 and have it automatically work out the correct field type and fill in
2243 the structures. The more adventurous can try:
2244 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2245 and it will (hopefully) work out the correct multibyte encoding.
2248 *) Change the 'req' utility to use the new field handling and multibyte
2249 copy routines. Before the DN field creation was handled in an ad hoc
2250 way in req, ca, and x509 which was rather broken and didn't support
2251 BMPStrings or UTF8Strings. Since some software doesn't implement
2252 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2253 using the dirstring_type option. See the new comment in the default
2254 openssl.cnf for more info.
2257 *) Make crypto/rand/md_rand.c more robust:
2258 - Assure unique random numbers after fork().
2259 - Make sure that concurrent threads access the global counter and
2260 md serializably so that we never lose entropy in them
2261 or use exactly the same state in multiple threads.
2262 Access to the large state is not always serializable because
2263 the additional locking could be a performance killer, and
2264 md should be large enough anyway.
2267 *) New file apps/app_rand.c with commonly needed functionality
2268 for handling the random seed file.
2270 Use the random seed file in some applications that previously did not:
2272 dsaparam -genkey (which also ignored its '-rand' option),
2275 x509 (when signing).
2276 Except on systems with /dev/urandom, it is crucial to have a random
2277 seed file at least for key creation, DSA signing, and for DH exchanges;
2278 for RSA signatures we could do without one.
2280 gendh and gendsa (unlike genrsa) used to read only the first byte
2281 of each file listed in the '-rand' option. The function as previously
2282 found in genrsa is now in app_rand.c and is used by all programs
2283 that support '-rand'.
2286 *) In RAND_write_file, use mode 0600 for creating files;
2287 don't just chmod when it may be too late.
2290 *) Report an error from X509_STORE_load_locations
2291 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2294 *) New function ASN1_mbstring_copy() this copies a string in either
2295 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2296 into an ASN1_STRING type. A mask of permissible types is passed
2297 and it chooses the "minimal" type to use or an error if not type
2301 *) Add function equivalents to the various macros in asn1.h. The old
2302 macros are retained with an M_ prefix. Code inside the library can
2303 use the M_ macros. External code (including the openssl utility)
2304 should *NOT* in order to be "shared library friendly".
2307 *) Add various functions that can check a certificate's extensions
2308 to see if it usable for various purposes such as SSL client,
2309 server or S/MIME and CAs of these types. This is currently
2310 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2311 verification. Also added a -purpose flag to x509 utility to
2312 print out all the purposes.
2315 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2319 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2320 for, obtain and decode and extension and obtain its critical flag.
2321 This allows all the necessary extension code to be handled in a
2322 single function call.
2325 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2326 platforms. See crypto/rc4/rc4_enc.c for further details.
2329 *) New -noout option to asn1parse. This causes no output to be produced
2330 its main use is when combined with -strparse and -out to extract data
2331 from a file (which may not be in ASN.1 format).
2334 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2335 when producing the local key id.
2336 [Richard Levitte <levitte@stacken.kth.se>]
2338 *) New option -dhparam in s_server. This allows a DH parameter file to be
2339 stated explicitly. If it is not stated then it tries the first server
2340 certificate file. The previous behaviour hard coded the filename
2344 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2345 a public key to be input or output. For example:
2346 openssl rsa -in key.pem -pubout -out pubkey.pem
2347 Also added necessary DSA public key functions to handle this.
2350 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2351 in the message. This was handled by allowing
2352 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2353 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2355 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2356 to the end of the strings whereas this didn't. This would cause problems
2357 if strings read with d2i_ASN1_bytes() were later modified.
2358 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2360 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2361 data and it contains EOF it will end up returning an error. This is
2362 caused by input 46 bytes long. The cause is due to the way base64
2363 BIOs find the start of base64 encoded data. They do this by trying a
2364 trial decode on each line until they find one that works. When they
2365 do a flag is set and it starts again knowing it can pass all the
2366 data directly through the decoder. Unfortunately it doesn't reset
2367 the context it uses. This means that if EOF is reached an attempt
2368 is made to pass two EOFs through the context and this causes the
2369 resulting error. This can also cause other problems as well. As is
2370 usual with these problems it takes *ages* to find and the fix is
2371 trivial: move one line.
2372 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2374 *) Ugly workaround to get s_client and s_server working under Windows. The
2375 old code wouldn't work because it needed to select() on sockets and the
2376 tty (for keypresses and to see if data could be written). Win32 only
2377 supports select() on sockets so we select() with a 1s timeout on the
2378 sockets and then see if any characters are waiting to be read, if none
2379 are present then we retry, we also assume we can always write data to
2380 the tty. This isn't nice because the code then blocks until we've
2381 received a complete line of data and it is effectively polling the
2382 keyboard at 1s intervals: however it's quite a bit better than not
2383 working at all :-) A dedicated Windows application might handle this
2384 with an event loop for example.
2387 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2388 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2389 will be called when RSA_sign() and RSA_verify() are used. This is useful
2390 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2391 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2392 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2393 This necessitated the support of an extra signature type NID_md5_sha1
2394 for SSL signatures and modifications to the SSL library to use it instead
2395 of calling RSA_public_decrypt() and RSA_private_encrypt().
2398 *) Add new -verify -CAfile and -CApath options to the crl program, these
2399 will lookup a CRL issuers certificate and verify the signature in a
2400 similar way to the verify program. Tidy up the crl program so it
2401 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2402 less strict. It will now permit CRL extensions even if it is not
2403 a V2 CRL: this will allow it to tolerate some broken CRLs.
2406 *) Initialize all non-automatic variables each time one of the openssl
2407 sub-programs is started (this is necessary as they may be started
2408 multiple times from the "OpenSSL>" prompt).
2409 [Lennart Bang, Bodo Moeller]
2411 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2412 removing all other RSA functionality (this is what NO_RSA does). This
2413 is so (for example) those in the US can disable those operations covered
2414 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2418 *) Non-copying interface to BIO pairs.
2419 (still largely untested)
2422 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2423 ASCII string. This was handled independently in various places before.
2426 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2427 UTF8 strings a character at a time.
2430 *) Use client_version from client hello to select the protocol
2431 (s23_srvr.c) and for RSA client key exchange verification
2432 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2435 *) Add various utility functions to handle SPKACs, these were previously
2436 handled by poking round in the structure internals. Added new function
2437 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2438 print, verify and generate SPKACs. Based on an original idea from
2439 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2442 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2445 *) Allow the config file extension section to be overwritten on the
2446 command line. Based on an original idea from Massimiliano Pala
2447 <madwolf@comune.modena.it>. The new option is called -extensions
2448 and can be applied to ca, req and x509. Also -reqexts to override
2449 the request extensions in req and -crlexts to override the crl extensions
2453 *) Add new feature to the SPKAC handling in ca. Now you can include
2454 the same field multiple times by preceding it by "XXXX." for example:
2457 this is the same syntax as used in the req config file.
2460 *) Allow certificate extensions to be added to certificate requests. These
2461 are specified in a 'req_extensions' option of the req section of the
2462 config file. They can be printed out with the -text option to req but
2463 are otherwise ignored at present.
2466 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2467 data read consists of only the final block it would not decrypted because
2468 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2469 A misplaced 'break' also meant the decrypted final block might not be
2470 copied until the next read.
2473 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2474 a few extra parameters to the DH structure: these will be useful if
2475 for example we want the value of 'q' or implement X9.42 DH.
2478 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2479 provides hooks that allow the default DSA functions or functions on a
2480 "per key" basis to be replaced. This allows hardware acceleration and
2481 hardware key storage to be handled without major modification to the
2482 library. Also added low level modexp hooks and CRYPTO_EX structure and
2483 associated functions.
2486 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2487 as "read only": it can't be written to and the buffer it points to will
2488 not be freed. Reading from a read only BIO is much more efficient than
2489 a normal memory BIO. This was added because there are several times when
2490 an area of memory needs to be read from a BIO. The previous method was
2491 to create a memory BIO and write the data to it, this results in two
2492 copies of the data and an O(n^2) reading algorithm. There is a new
2493 function BIO_new_mem_buf() which creates a read only memory BIO from
2494 an area of memory. Also modified the PKCS#7 routines to use read only
2498 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2499 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2500 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2501 but a retry condition occured while trying to read the rest.
2504 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2505 NID_pkcs7_encrypted by default: this was wrong since this should almost
2506 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2507 the encrypted data type: this is a more sensible place to put it and it
2508 allows the PKCS#12 code to be tidied up that duplicated this
2512 *) Changed obj_dat.pl script so it takes its input and output files on
2513 the command line. This should avoid shell escape redirection problems
2517 *) Initial support for certificate extension requests, these are included
2518 in things like Xenroll certificate requests. Included functions to allow
2519 extensions to be obtained and added.
2522 *) -crlf option to s_client and s_server for sending newlines as
2523 CRLF (as required by many protocols).
2526 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2528 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2529 [Ralf S. Engelschall]
2531 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2532 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2534 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2538 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2539 DH parameters/keys (q is lost during that conversion, but the resulting
2540 DH parameters contain its length).
2542 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2543 much faster than DH_generate_parameters (which creates parameters
2544 where p = 2*q + 1), and also the smaller q makes DH computations
2545 much more efficient (160-bit exponentiation instead of 1024-bit
2546 exponentiation); so this provides a convenient way to support DHE
2547 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2548 utter importance to use
2549 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2551 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2552 when such DH parameters are used, because otherwise small subgroup
2553 attacks may become possible!
2556 *) Avoid memory leak in i2d_DHparams.
2559 *) Allow the -k option to be used more than once in the enc program:
2560 this allows the same encrypted message to be read by multiple recipients.
2563 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2564 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2565 it will always use the numerical form of the OID, even if it has a short
2569 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2570 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2571 otherwise bn_mod_exp was called. In the case of hardware keys for example
2572 no private key components need be present and it might store extra data
2573 in the RSA structure, which cannot be accessed from bn_mod_exp.
2574 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2575 private key operations.
2578 *) Added support for SPARC Linux.
2581 *) pem_password_cb function type incompatibly changed from
2582 typedef int pem_password_cb(char *buf, int size, int rwflag);
2584 ....(char *buf, int size, int rwflag, void *userdata);
2585 so that applications can pass data to their callbacks:
2586 The PEM[_ASN1]_{read,write}... functions and macros now take an
2587 additional void * argument, which is just handed through whenever
2588 the password callback is called.
2589 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2591 New function SSL_CTX_set_default_passwd_cb_userdata.
2593 Compatibility note: As many C implementations push function arguments
2594 onto the stack in reverse order, the new library version is likely to
2595 interoperate with programs that have been compiled with the old
2596 pem_password_cb definition (PEM_whatever takes some data that
2597 happens to be on the stack as its last argument, and the callback
2598 just ignores this garbage); but there is no guarantee whatsoever that
2601 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2602 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2603 problems not only on Windows, but also on some Unix platforms.
2604 To avoid problematic command lines, these definitions are now in an
2605 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2606 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2609 *) MIPS III/IV assembler module is reimplemented.
2612 *) More DES library cleanups: remove references to srand/rand and
2613 delete an unused file.
2616 *) Add support for the the free Netwide assembler (NASM) under Win32,
2617 since not many people have MASM (ml) and it can be hard to obtain.
2618 This is currently experimental but it seems to work OK and pass all
2619 the tests. Check out INSTALL.W32 for info.
2622 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2623 without temporary keys kept an extra copy of the server key,
2624 and connections with temporary keys did not free everything in case
2628 *) New function RSA_check_key and new openssl rsa option -check
2629 for verifying the consistency of RSA keys.
2630 [Ulf Moeller, Bodo Moeller]
2632 *) Various changes to make Win32 compile work:
2633 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2634 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2635 comparison" warnings.
2636 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2639 *) Add a debugging option to PKCS#5 v2 key generation function: when
2640 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2641 derived keys are printed to stderr.
2644 *) Copy the flags in ASN1_STRING_dup().
2645 [Roman E. Pavlov <pre@mo.msk.ru>]
2647 *) The x509 application mishandled signing requests containing DSA
2648 keys when the signing key was also DSA and the parameters didn't match.
2650 It was supposed to omit the parameters when they matched the signing key:
2651 the verifying software was then supposed to automatically use the CA's
2652 parameters if they were absent from the end user certificate.
2654 Omitting parameters is no longer recommended. The test was also
2655 the wrong way round! This was probably due to unusual behaviour in
2656 EVP_cmp_parameters() which returns 1 if the parameters match.
2657 This meant that parameters were omitted when they *didn't* match and
2658 the certificate was useless. Certificates signed with 'ca' didn't have
2660 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2662 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2663 The interface is as follows:
2664 Applications can use
2665 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2666 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2667 "off" is now the default.
2668 The library internally uses
2669 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2670 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2671 to disable memory-checking temporarily.
2673 Some inconsistent states that previously were possible (and were
2674 even the default) are now avoided.
2676 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2677 with each memory chunk allocated; this is occasionally more helpful
2678 than just having a counter.
2680 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2682 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2686 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2687 which largely parallels "options", but is for changing API behaviour,
2688 whereas "options" are about protocol behaviour.
2689 Initial "mode" flags are:
2691 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2692 a single record has been written.
2693 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2694 retries use the same buffer location.
2695 (But all of the contents must be
2699 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2702 *) Fix problems with no-hmac etc.
2703 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2705 *) New functions RSA_get_default_method(), RSA_set_method() and
2706 RSA_get_method(). These allows replacement of RSA_METHODs without having
2707 to mess around with the internals of an RSA structure.
2710 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2711 Also really enable memory leak checks in openssl.c and in some
2713 [Chad C. Mulligan, Bodo Moeller]
2715 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2716 up the length of negative integers. This has now been simplified to just
2717 store the length when it is first determined and use it later, rather
2718 than trying to keep track of where data is copied and updating it to
2720 [Steve Henson, reported by Brien Wheeler
2721 <bwheeler@authentica-security.com>]
2723 *) Add a new function PKCS7_signatureVerify. This allows the verification
2724 of a PKCS#7 signature but with the signing certificate passed to the
2725 function itself. This contrasts with PKCS7_dataVerify which assumes the
2726 certificate is present in the PKCS#7 structure. This isn't always the
2727 case: certificates can be omitted from a PKCS#7 structure and be
2728 distributed by "out of band" means (such as a certificate database).
2731 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2732 function prototypes in pem.h, also change util/mkdef.pl to add the
2733 necessary function names.
2736 *) mk1mf.pl (used by Windows builds) did not properly read the
2737 options set by Configure in the top level Makefile, and Configure
2738 was not even able to write more than one option correctly.
2739 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2742 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2743 file to be loaded from a BIO or FILE pointer. The BIO version will
2744 for example allow memory BIOs to contain config info.
2747 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2748 Whoever hopes to achieve shared-library compatibility across versions
2749 must use this, not the compile-time macro.
2750 (Exercise 0.9.4: Which is the minimum library version required by
2752 Note: All this applies only to multi-threaded programs, others don't
2756 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2757 through a BIO pair triggered the default case, i.e.
2758 SSLerr(...,SSL_R_UNKNOWN_STATE).
2761 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2762 can use the SSL library even if none of the specific BIOs is
2766 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2767 for the encoded length.
2768 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2770 *) Add initial documentation of the X509V3 functions.
2773 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2774 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2775 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2776 secure PKCS#8 private key format with a high iteration count.
2779 *) Fix determination of Perl interpreter: A perl or perl5
2780 _directory_ in $PATH was also accepted as the interpreter.
2781 [Ralf S. Engelschall]
2783 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2784 wrong with it but it was very old and did things like calling
2785 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2789 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2790 to use the new extension code.
2793 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2794 with macros. This should make it easier to change their form, add extra
2795 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2799 *) Add to configuration table a new entry that can specify an alternative
2800 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2801 according to Mark Crispin <MRC@Panda.COM>.
2805 *) DES CBC did not update the IV. Weird.
2808 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2809 Changing the behaviour of the former might break existing programs --
2810 where IV updating is needed, des_ncbc_encrypt can be used.
2813 *) When bntest is run from "make test" it drives bc to check its
2814 calculations, as well as internally checking them. If an internal check
2815 fails, it needs to cause bc to give a non-zero result or make test carries
2816 on without noticing the failure. Fixed.
2819 *) DES library cleanups.
2822 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2823 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2824 ciphers. NOTE: although the key derivation function has been verified
2825 against some published test vectors it has not been extensively tested
2826 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2830 *) Instead of "mkdir -p", which is not fully portable, use new
2831 Perl script "util/mkdir-p.pl".
2834 *) Rewrite the way password based encryption (PBE) is handled. It used to
2835 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2836 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2837 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2838 the 'parameter' field of the AlgorithmIdentifier is passed to the
2839 underlying key generation function so it must do its own ASN1 parsing.
2840 This has also changed the EVP_PBE_CipherInit() function which now has a
2841 'parameter' argument instead of literal salt and iteration count values
2842 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2845 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2846 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2847 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2848 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2849 value was just used as a "magic string" and not used directly its
2850 value doesn't matter.
2853 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2857 *) "linux-sparc64" configuration (ultrapenguin).
2858 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2859 "linux-sparc" configuration.
2860 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2862 *) config now generates no-xxx options for missing ciphers.
2865 *) Support the EBCDIC character set (work in progress).
2866 File ebcdic.c not yet included because it has a different license.
2867 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2869 *) Support BS2000/OSD-POSIX.
2870 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2872 *) Make callbacks for key generation use void * instead of char *.
2875 *) Make S/MIME samples compile (not yet tested).
2878 *) Additional typesafe stacks.
2881 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2885 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2887 *) New configuration variant "sco5-gcc".
2889 *) Updated some demos.
2890 [Sean O Riordain, Wade Scholine]
2892 *) Add missing BIO_free at exit of pkcs12 application.
2895 *) Fix memory leak in conf.c.
2898 *) Updates for Win32 to assembler version of MD5.
2901 *) Set #! path to perl in apps/der_chop to where we found it
2902 instead of using a fixed path.
2905 *) SHA library changes for irix64-mips4-cc.
2908 *) Improvements for VMS support.
2912 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2914 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2915 This also avoids the problems with SC4.2 and unpatched SC5.
2916 [Andy Polyakov <appro@fy.chalmers.se>]
2918 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2919 These are required because of the typesafe stack would otherwise break
2920 existing code. If old code used a structure member which used to be STACK
2921 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2922 sk_num or sk_value it would produce an error because the num, data members
2923 are not present in STACK_OF. Now it just produces a warning. sk_set
2924 replaces the old method of assigning a value to sk_value
2925 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2926 that does this will no longer work (and should use sk_set instead) but
2927 this could be regarded as a "questionable" behaviour anyway.
2930 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2931 correctly handle encrypted S/MIME data.
2934 *) Change type of various DES function arguments from des_cblock
2935 (which means, in function argument declarations, pointer to char)
2936 to des_cblock * (meaning pointer to array with 8 char elements),
2937 which allows the compiler to do more typechecking; it was like
2938 that back in SSLeay, but with lots of ugly casts.
2940 Introduce new type const_des_cblock.
2943 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2944 problems: find RecipientInfo structure that matches recipient certificate
2945 and initialise the ASN1 structures properly based on passed cipher.
2948 *) Belatedly make the BN tests actually check the results.
2951 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2952 to and from BNs: it was completely broken. New compilation option
2953 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2954 key elements as negative integers.
2957 *) Reorganize and speed up MD5.
2958 [Andy Polyakov <appro@fy.chalmers.se>]
2961 [Richard Levitte <richard@levitte.org>]
2963 *) New option -out to asn1parse to allow the parsed structure to be
2964 output to a file. This is most useful when combined with the -strparse
2965 option to examine the output of things like OCTET STRINGS.
2968 *) Make SSL library a little more fool-proof by not requiring any longer
2969 that SSL_set_{accept,connect}_state be called before
2970 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2971 in many applications because usually everything *appeared* to work as
2972 intended anyway -- now it really works as intended).
2975 *) Move openssl.cnf out of lib/.
2978 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2979 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2980 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2981 [Ralf S. Engelschall]
2983 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2984 handle PKCS#7 enveloped data properly.
2985 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2987 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2988 copying pointers. The cert_st handling is changed by this in
2989 various ways (and thus what used to be known as ctx->default_cert
2990 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
2991 any longer when s->cert does not give us what we need).
2992 ssl_cert_instantiate becomes obsolete by this change.
2993 As soon as we've got the new code right (possibly it already is?),
2994 we have solved a couple of bugs of the earlier code where s->cert
2995 was used as if it could not have been shared with other SSL structures.
2997 Note that using the SSL API in certain dirty ways now will result
2998 in different behaviour than observed with earlier library versions:
2999 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3000 does not influence s as it used to.
3002 In order to clean up things more thoroughly, inside SSL_SESSION
3003 we don't use CERT any longer, but a new structure SESS_CERT
3004 that holds per-session data (if available); currently, this is
3005 the peer's certificate chain and, for clients, the server's certificate
3006 and temporary key. CERT holds only those values that can have
3007 meaningful defaults in an SSL_CTX.
3010 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3011 from the internal representation. Various PKCS#7 fixes: remove some
3012 evil casts and set the enc_dig_alg field properly based on the signing
3016 *) Allow PKCS#12 password to be set from the command line or the
3017 environment. Let 'ca' get its config file name from the environment
3018 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3022 *) Allow certificate policies extension to use an IA5STRING for the
3023 organization field. This is contrary to the PKIX definition but
3024 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3028 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3029 without disallowing inline assembler and the like for non-pedantic builds.
3032 *) Support Borland C++ builder.
3033 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3038 *) SHA-1 cleanups and performance enhancements.
3039 [Andy Polyakov <appro@fy.chalmers.se>]
3041 *) Sparc v8plus assembler for the bignum library.
3042 [Andy Polyakov <appro@fy.chalmers.se>]
3044 *) Accept any -xxx and +xxx compiler options in Configure.
3047 *) Update HPUX configuration.
3050 *) Add missing sk_<type>_unshift() function to safestack.h
3051 [Ralf S. Engelschall]
3053 *) New function SSL_CTX_use_certificate_chain_file that sets the
3054 "extra_cert"s in addition to the certificate. (This makes sense
3055 only for "PEM" format files, as chains as a whole are not
3059 *) Support verify_depth from the SSL API.
3060 x509_vfy.c had what can be considered an off-by-one-error:
3061 Its depth (which was not part of the external interface)
3062 was actually counting the number of certificates in a chain;
3063 now it really counts the depth.
3066 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3067 instead of X509err, which often resulted in confusing error
3068 messages since the error codes are not globally unique
3069 (e.g. an alleged error in ssl3_accept when a certificate
3070 didn't match the private key).
3072 *) New function SSL_CTX_set_session_id_context that allows to set a default
3073 value (so that you don't need SSL_set_session_id_context for each
3074 connection using the SSL_CTX).
3077 *) OAEP decoding bug fix.
3080 *) Support INSTALL_PREFIX for package builders, as proposed by
3084 *) New Configure options "threads" and "no-threads". For systems
3085 where the proper compiler options are known (currently Solaris
3086 and Linux), "threads" is the default.
3089 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3092 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3093 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3094 such as /usr/local/bin.
3097 *) "make linux-shared" to build shared libraries.
3098 [Niels Poppe <niels@netbox.org>]
3100 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3103 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3104 extension adding in x509 utility.
3107 *) Remove NOPROTO sections and error code comments.
3110 *) Partial rewrite of the DEF file generator to now parse the ANSI
3114 *) New Configure options --prefix=DIR and --openssldir=DIR.
3117 *) Complete rewrite of the error code script(s). It is all now handled
3118 by one script at the top level which handles error code gathering,
3119 header rewriting and C source file generation. It should be much better
3120 than the old method: it now uses a modified version of Ulf's parser to
3121 read the ANSI prototypes in all header files (thus the old K&R definitions
3122 aren't needed for error creation any more) and do a better job of
3123 translating function codes into names. The old 'ASN1 error code imbedded
3124 in a comment' is no longer necessary and it doesn't use .err files which
3125 have now been deleted. Also the error code call doesn't have to appear all
3126 on one line (which resulted in some large lines...).
3129 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3132 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3133 0 (which usually indicates a closed connection), but continue reading.
3136 *) Fix some race conditions.
3139 *) Add support for CRL distribution points extension. Add Certificate
3140 Policies and CRL distribution points documentation.
3143 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3146 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3147 8 of keying material. Merlin has also confirmed interop with this fix
3148 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3149 [Merlin Hughes <merlin@baltimore.ie>]
3151 *) Fix lots of warnings.
3152 [Richard Levitte <levitte@stacken.kth.se>]
3154 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3155 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3156 [Richard Levitte <levitte@stacken.kth.se>]
3158 *) Fix problems with sizeof(long) == 8.
3159 [Andy Polyakov <appro@fy.chalmers.se>]
3161 *) Change functions to ANSI C.
3164 *) Fix typos in error codes.
3165 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3167 *) Remove defunct assembler files from Configure.
3170 *) SPARC v8 assembler BIGNUM implementation.
3171 [Andy Polyakov <appro@fy.chalmers.se>]
3173 *) Support for Certificate Policies extension: both print and set.
3174 Various additions to support the r2i method this uses.
3177 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3178 return a const string when you are expecting an allocated buffer.
3181 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3182 types DirectoryString and DisplayText.
3185 *) Add code to allow r2i extensions to access the configuration database,
3186 add an LHASH database driver and add several ctx helper functions.
3189 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3190 fail when they extended the size of a BIGNUM.
3193 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3194 support typesafe stack.
3197 *) Fix typo in SSL_[gs]et_options().
3198 [Nils Frostberg <nils@medcom.se>]
3200 *) Delete various functions and files that belonged to the (now obsolete)
3201 old X509V3 handling code.
3204 *) New Configure option "rsaref".
3207 *) Don't auto-generate pem.h.
3210 *) Introduce type-safe ASN.1 SETs.
3213 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3214 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3216 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3217 that links with OpenSSL (well at least cause lots of warnings), but fear
3218 not: the conversion is trivial, and it eliminates loads of evil casts. A
3219 few STACKed things have been converted already. Feel free to convert more.
3220 In the fullness of time, I'll do away with the STACK type altogether.
3223 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3224 specified in <certfile> by updating the entry in the index.txt file.
3225 This way one no longer has to edit the index.txt file manually for
3226 revoking a certificate. The -revoke option does the gory details now.
3227 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3229 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3230 `-text' option at all and this way the `-noout -text' combination was
3231 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3232 [Ralf S. Engelschall]
3234 *) Make sure a corresponding plain text error message exists for the
3235 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3236 verify callback function determined that a certificate was revoked.
3237 [Ralf S. Engelschall]
3239 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3240 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3241 all available cipers including rc5, which was forgotten until now.
3242 In order to let the testing shell script know which algorithms
3243 are available, a new (up to now undocumented) command
3244 "openssl list-cipher-commands" is used.
3247 *) Bugfix: s_client occasionally would sleep in select() when
3248 it should have checked SSL_pending() first.
3251 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3252 the raw DSA values prior to ASN.1 encoding.
3255 *) Tweaks to Configure
3256 [Niels Poppe <niels@netbox.org>]
3258 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3262 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3265 *) New config option to avoid instructions that are illegal on the 80386.
3266 The default code is faster, but requires at least a 486.
3269 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3270 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3271 same as SSL2_VERSION anyway.
3274 *) New "-showcerts" option for s_client.
3277 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3278 application. Various cleanups and fixes.
3281 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3282 modify error routines to work internally. Add error codes and PBE init
3283 to library startup routines.
3286 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3287 packing functions to asn1 and evp. Changed function names and error
3288 codes along the way.
3291 *) PKCS12 integration: and so it begins... First of several patches to
3292 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3293 objects to objects.h
3296 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3297 and display support for Thawte strong extranet extension.
3300 *) Add LinuxPPC support.
3301 [Jeff Dubrule <igor@pobox.org>]
3303 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3304 bn_div_words in alpha.s.
3305 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3307 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3308 OAEP isn't supported when OpenSSL is built with RSAref.
3309 [Ulf Moeller <ulf@fitug.de>]
3311 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3312 so they no longer are missing under -DNOPROTO.
3313 [Soren S. Jorvang <soren@t.dk>]
3316 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3318 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3319 doesn't work when the session is reused. Coming soon!
3322 *) Fix a security hole, that allows sessions to be reused in the wrong
3323 context thus bypassing client cert protection! All software that uses
3324 client certs and session caches in multiple contexts NEEDS PATCHING to
3325 allow session reuse! A fuller solution is in the works.
3326 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3328 *) Some more source tree cleanups (removed obsolete files
3329 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3330 permission on "config" script to be executable) and a fix for the INSTALL
3332 [Ulf Moeller <ulf@fitug.de>]
3334 *) Remove some legacy and erroneous uses of malloc, free instead of
3336 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3338 *) Make rsa_oaep_test return non-zero on error.
3339 [Ulf Moeller <ulf@fitug.de>]
3341 *) Add support for native Solaris shared libraries. Configure
3342 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3343 if someone would make that last step automatic.
3344 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3346 *) ctx_size was not built with the right compiler during "make links". Fixed.
3349 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3350 except NULL ciphers". This means the default cipher list will no longer
3351 enable NULL ciphers. They need to be specifically enabled e.g. with
3352 the string "DEFAULT:eNULL".
3355 *) Fix to RSA private encryption routines: if p < q then it would
3356 occasionally produce an invalid result. This will only happen with
3357 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3360 *) Be less restrictive and allow also `perl util/perlpath.pl
3361 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3362 because this way one can also use an interpreter named `perl5' (which is
3363 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3364 installed as `perl').
3365 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3367 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3368 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3370 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3371 advapi32.lib to Win32 build and change the pem test comparision
3372 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3373 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3374 and crypto/des/ede_cbcm_enc.c.
3377 *) DES quad checksum was broken on big-endian architectures. Fixed.
3380 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3381 Win32 test batch file so it (might) work again. The Win32 test batch file
3382 is horrible: I feel ill....
3385 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3386 in e_os.h. Audit of header files to check ANSI and non ANSI
3387 sections: 10 functions were absent from non ANSI section and not exported
3388 from Windows DLLs. Fixed up libeay.num for new functions.
3391 *) Make `openssl version' output lines consistent.
3392 [Ralf S. Engelschall]
3394 *) Fix Win32 symbol export lists for BIO functions: Added
3395 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3396 to ms/libeay{16,32}.def.
3397 [Ralf S. Engelschall]
3399 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3400 fine under Unix and passes some trivial tests I've now added. But the
3401 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3402 added to make sure no one expects that this stuff really works in the
3403 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3404 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3406 [Ralf S. Engelschall]
3408 *) Fix the generation of two part addresses in perl.
3409 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3411 *) Add config entry for Linux on MIPS.
3412 [John Tobey <jtobey@channel1.com>]
3414 *) Make links whenever Configure is run, unless we are on Windoze.
3417 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3418 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3422 *) Add a useful kludge to allow package maintainers to specify compiler and
3423 other platforms details on the command line without having to patch the
3424 Configure script everytime: One now can use ``perl Configure
3425 <id>:<details>'', i.e. platform ids are allowed to have details appended
3426 to them (seperated by colons). This is treated as there would be a static
3427 pre-configured entry in Configure's %table under key <id> with value
3428 <details> and ``perl Configure <id>'' is called. So, when you want to
3429 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3430 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3431 now, which overrides the FreeBSD-elf entry on-the-fly.
3432 [Ralf S. Engelschall]
3434 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3437 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3438 on the `perl Configure ...' command line. This way one can compile
3439 OpenSSL libraries with Position Independent Code (PIC) which is needed
3440 for linking it into DSOs.
3441 [Ralf S. Engelschall]
3443 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3447 *) Cleaned up the LICENSE document: The official contact for any license
3448 questions now is the OpenSSL core team under openssl-core@openssl.org.
3449 And add a paragraph about the dual-license situation to make sure people
3450 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3451 to the OpenSSL toolkit.
3452 [Ralf S. Engelschall]
3454 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3455 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3456 Additonally cleaned up the `make links' target: Remove unnecessary
3457 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3458 to speed processing and no longer clutter the display with confusing
3459 stuff. Instead only the actually done links are displayed.
3460 [Ralf S. Engelschall]
3462 *) Permit null encryption ciphersuites, used for authentication only. It used
3463 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3464 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3468 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3469 signed attributes when verifying signatures (this would break them),
3470 the detached data encoding was wrong and public keys obtained using
3471 X509_get_pubkey() weren't freed.
3474 *) Add text documentation for the BUFFER functions. Also added a work around
3475 to a Win95 console bug. This was triggered by the password read stuff: the
3476 last character typed gets carried over to the next fread(). If you were
3477 generating a new cert request using 'req' for example then the last
3478 character of the passphrase would be CR which would then enter the first
3482 *) Added the new `Includes OpenSSL Cryptography Software' button as
3483 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3484 button and can be used by applications based on OpenSSL to show the
3485 relationship to the OpenSSL project.
3486 [Ralf S. Engelschall]
3488 *) Remove confusing variables in function signatures in files
3489 ssl/ssl_lib.c and ssl/ssl.h.
3490 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3492 *) Don't install bss_file.c under PREFIX/include/
3493 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3495 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3496 functions that return function pointers and has support for NT specific
3497 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3498 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3499 unsigned to signed types: this was killing the Win32 compile.
3502 *) Add new certificate file to stack functions,
3503 SSL_add_dir_cert_subjects_to_stack() and
3504 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3505 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3506 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3507 This means that Apache-SSL and similar packages don't have to mess around
3508 to add as many CAs as they want to the preferred list.
3511 *) Experiment with doxygen documentation. Currently only partially applied to
3513 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3514 openssl.doxy as the configuration file.
3517 *) Get rid of remaining C++-style comments which strict C compilers hate.
3518 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3520 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3521 compiled in by default: it has problems with large keys.
3524 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3525 DH private keys and/or callback functions which directly correspond to
3526 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3527 is needed for applications which have to configure certificates on a
3528 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3530 For the RSA certificate situation is makes no difference, but
3531 for the DSA certificate situation this fixes the "no shared cipher"
3532 problem where the OpenSSL cipher selection procedure failed because the
3533 temporary keys were not overtaken from the context and the API provided
3534 no way to reconfigure them.
3535 The new functions now let applications reconfigure the stuff and they
3536 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3537 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3538 non-public-API function ssl_cert_instantiate() is used as a helper
3539 function and also to reduce code redundancy inside ssl_rsa.c.
3540 [Ralf S. Engelschall]
3542 *) Move s_server -dcert and -dkey options out of the undocumented feature
3543 area because they are useful for the DSA situation and should be
3544 recognized by the users.
3545 [Ralf S. Engelschall]
3547 *) Fix the cipher decision scheme for export ciphers: the export bits are
3548 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3549 SSL_EXP_MASK. So, the original variable has to be used instead of the
3550 already masked variable.
3551 [Richard Levitte <levitte@stacken.kth.se>]
3553 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3554 [Richard Levitte <levitte@stacken.kth.se>]
3556 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3557 from `int' to `unsigned int' because it's a length and initialized by
3558 EVP_DigestFinal() which expects an `unsigned int *'.
3559 [Richard Levitte <levitte@stacken.kth.se>]
3561 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3562 script. Instead use the usual Shell->Perl transition trick.
3563 [Ralf S. Engelschall]
3565 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3566 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3567 -noout -modulus' as it's already the case for `openssl rsa -noout
3568 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3569 currently the public key is printed (a decision which was already done by
3570 `openssl dsa -modulus' in the past) which serves a similar purpose.
3571 Additionally the NO_RSA no longer completely removes the whole -modulus
3572 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3574 [Ralf S. Engelschall]
3576 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3577 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3578 [Arne Ansper <arne@ats.cyber.ee>]
3580 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3581 to be added. Now both 'req' and 'ca' can use new objects defined in the
3585 *) Add cool BIO that does syslog (or event log on NT).
3586 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3588 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3589 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3590 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3591 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3594 *) Add preliminary config info for new extension code.
3597 *) Make RSA_NO_PADDING really use no padding.
3598 [Ulf Moeller <ulf@fitug.de>]
3600 *) Generate errors when private/public key check is done.
3603 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3604 for some CRL extensions and new objects added.
3607 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3608 key usage extension and fuller support for authority key id.
3611 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3612 padding method for RSA, which is recommended for new applications in PKCS
3613 #1 v2.0 (RFC 2437, October 1998).
3614 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3615 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3616 against Bleichbacher's attack on RSA.
3617 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3620 *) Updates to the new SSL compression code
3621 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3623 *) Fix so that the version number in the master secret, when passed
3624 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3625 (because the server will not accept higher), that the version number
3626 is 0x03,0x01, not 0x03,0x00
3627 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3629 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3630 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3631 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3634 *) Support for RAW extensions where an arbitrary extension can be
3635 created by including its DER encoding. See apps/openssl.cnf for
3639 *) Make sure latest Perl versions don't interpret some generated C array
3640 code as Perl array code in the crypto/err/err_genc.pl script.
3641 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3643 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3644 not many people have the assembler. Various Win32 compilation fixes and
3645 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3649 *) Modify configure script 'Configure' to automatically create crypto/date.h
3650 file under Win32 and also build pem.h from pem.org. New script
3651 util/mkfiles.pl to create the MINFO file on environments that can't do a
3652 'make files': perl util/mkfiles.pl >MINFO should work.
3655 *) Major rework of DES function declarations, in the pursuit of correctness
3656 and purity. As a result, many evil casts evaporated, and some weirdness,
3657 too. You may find this causes warnings in your code. Zapping your evil
3658 casts will probably fix them. Mostly.
3661 *) Fix for a typo in asn1.h. Bug fix to object creation script
3662 obj_dat.pl. It considered a zero in an object definition to mean
3663 "end of object": none of the objects in objects.h have any zeros
3664 so it wasn't spotted.
3665 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3667 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3668 Masking (CBCM). In the absence of test vectors, the best I have been able
3669 to do is check that the decrypt undoes the encrypt, so far. Send me test
3670 vectors if you have them.
3673 *) Correct calculation of key length for export ciphers (too much space was
3674 allocated for null ciphers). This has not been tested!
3677 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3678 message is now correct (it understands "crypto" and "ssl" on its
3679 command line). There is also now an "update" option. This will update
3680 the util/ssleay.num and util/libeay.num files with any new functions.
3682 perl util/mkdef.pl crypto ssl update
3683 it will update them.
3686 *) Overhauled the Perl interface (perl/*):
3687 - ported BN stuff to OpenSSL's different BN library
3688 - made the perl/ source tree CVS-aware
3689 - renamed the package from SSLeay to OpenSSL (the files still contain
3690 their history because I've copied them in the repository)
3691 - removed obsolete files (the test scripts will be replaced
3692 by better Test::Harness variants in the future)
3693 [Ralf S. Engelschall]
3695 *) First cut for a very conservative source tree cleanup:
3696 1. merge various obsolete readme texts into doc/ssleay.txt
3697 where we collect the old documents and readme texts.
3698 2. remove the first part of files where I'm already sure that we no
3699 longer need them because of three reasons: either they are just temporary
3700 files which were left by Eric or they are preserved original files where
3701 I've verified that the diff is also available in the CVS via "cvs diff
3702 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3703 the crypto/md/ stuff).
3704 [Ralf S. Engelschall]
3706 *) More extension code. Incomplete support for subject and issuer alt
3707 name, issuer and authority key id. Change the i2v function parameters
3708 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3709 what that's for :-) Fix to ASN1 macro which messed up
3710 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3713 *) Preliminary support for ENUMERATED type. This is largely copied from the
3717 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3718 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3720 *) Make sure `make rehash' target really finds the `openssl' program.
3721 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3723 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3724 like to hear about it if this slows down other processors.
3727 *) Add CygWin32 platform information to Configure script.
3728 [Alan Batie <batie@aahz.jf.intel.com>]
3730 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3731 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3733 *) New program nseq to manipulate netscape certificate sequences
3736 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3740 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3741 but the BN code had some problems that would cause failures when
3742 doing certificate verification and some other functions.
3743 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3745 *) Add ASN1 and PEM code to support netscape certificate sequences.
3748 *) Add ASN1 and PEM code to support netscape certificate sequences.
3751 *) Add several PKIX and private extended key usage OIDs.
3754 *) Modify the 'ca' program to handle the new extension code. Modify
3755 openssl.cnf for new extension format, add comments.
3758 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3759 and add a sample to openssl.cnf so req -x509 now adds appropriate
3763 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3764 error code, add initial support to X509_print() and x509 application.
3767 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3768 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3769 stuff is currently isolated and isn't even compiled yet.
3772 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3773 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3774 Removed the versions check from X509 routines when loading extensions:
3775 this allows certain broken certificates that don't set the version
3776 properly to be processed.
3779 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3780 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3781 can still be regenerated with "make depend".
3784 *) Spelling mistake in C version of CAST-128.
3785 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3787 *) Changes to the error generation code. The perl script err-code.pl
3788 now reads in the old error codes and retains the old numbers, only
3789 adding new ones if necessary. It also only changes the .err files if new
3790 codes are added. The makefiles have been modified to only insert errors
3791 when needed (to avoid needlessly modifying header files). This is done
3792 by only inserting errors if the .err file is newer than the auto generated
3793 C file. To rebuild all the error codes from scratch (the old behaviour)
3794 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3795 or delete all the .err files.
3798 *) CAST-128 was incorrectly implemented for short keys. The C version has
3799 been fixed, but is untested. The assembler versions are also fixed, but
3800 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3801 to regenerate it if needed.
3802 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3803 Hagino <itojun@kame.net>]
3805 *) File was opened incorrectly in randfile.c.
3806 [Ulf Möller <ulf@fitug.de>]
3808 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3809 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3810 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3811 al: it's just almost always a UTCTime. Note this patch adds new error
3812 codes so do a "make errors" if there are problems.
3815 *) Correct Linux 1 recognition in config.
3816 [Ulf Möller <ulf@fitug.de>]
3818 *) Remove pointless MD5 hash when using DSA keys in ca.
3819 [Anonymous <nobody@replay.com>]
3821 *) Generate an error if given an empty string as a cert directory. Also
3822 generate an error if handed NULL (previously returned 0 to indicate an
3823 error, but didn't set one).
3824 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3826 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3829 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3830 parameters. This was causing a warning which killed off the Win32 compile.
3833 *) Remove C++ style comments from crypto/bn/bn_local.h.
3834 [Neil Costigan <neil.costigan@celocom.com>]
3836 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3837 based on a text string, looking up short and long names and finally
3838 "dot" format. The "dot" format stuff didn't work. Added new function
3839 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3840 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3841 OID is not part of the table.
3844 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3845 X509_LOOKUP_by_alias().
3848 *) Sort openssl functions by name.
3851 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3852 encryption from sample DSA keys (in case anyone is interested the password
3856 *) Make _all_ *_free functions accept a NULL pointer.
3857 [Frans Heymans <fheymans@isaserver.be>]
3859 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3861 [Anonymous <nobody@replay.com>]
3863 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3864 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3866 *) Don't blow it for numeric -newkey arguments to apps/req.
3867 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3869 *) Temp key "for export" tests were wrong in s3_srvr.c.
3870 [Anonymous <nobody@replay.com>]
3872 *) Add prototype for temp key callback functions
3873 SSL_CTX_set_tmp_{rsa,dh}_callback().
3876 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3877 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3880 *) X509_name_add_entry() freed the wrong thing after an error.
3881 [Arne Ansper <arne@ats.cyber.ee>]
3883 *) rsa_eay.c would attempt to free a NULL context.
3884 [Arne Ansper <arne@ats.cyber.ee>]
3886 *) BIO_s_socket() had a broken should_retry() on Windoze.
3887 [Arne Ansper <arne@ats.cyber.ee>]
3889 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3890 [Arne Ansper <arne@ats.cyber.ee>]
3892 *) Make sure the already existing X509_STORE->depth variable is initialized
3893 in X509_STORE_new(), but document the fact that this variable is still
3894 unused in the certificate verification process.
3895 [Ralf S. Engelschall]
3897 *) Fix the various library and apps files to free up pkeys obtained from
3898 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3901 *) Fix reference counting in X509_PUBKEY_get(). This makes
3902 demos/maurice/example2.c work, amongst others, probably.
3903 [Steve Henson and Ben Laurie]
3905 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3906 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3907 are no longer created. This way we have a single and consistent command
3908 line interface `openssl <command>', similar to `cvs <command>'.
3909 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3911 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3912 BIT STRING wrapper always have zero unused bits.
3915 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3918 *) Make the top-level INSTALL documentation easier to understand.
3921 *) Makefiles updated to exit if an error occurs in a sub-directory
3922 make (including if user presses ^C) [Paul Sutton]
3924 *) Make Montgomery context stuff explicit in RSA data structure.
3927 *) Fix build order of pem and err to allow for generated pem.h.
3930 *) Fix renumbering bug in X509_NAME_delete_entry().
3933 *) Enhanced the err-ins.pl script so it makes the error library number
3934 global and can add a library name. This is needed for external ASN1 and
3935 other error libraries.
3938 *) Fixed sk_insert which never worked properly.
3941 *) Fix ASN1 macros so they can handle indefinite length construted
3942 EXPLICIT tags. Some non standard certificates use these: they can now
3946 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3947 into a single doc/ssleay.txt bundle. This way the information is still
3948 preserved but no longer messes up this directory. Now it's new room for
3949 the new set of documenation files.
3950 [Ralf S. Engelschall]
3952 *) SETs were incorrectly DER encoded. This was a major pain, because they
3953 shared code with SEQUENCEs, which aren't coded the same. This means that
3954 almost everything to do with SETs or SEQUENCEs has either changed name or
3955 number of arguments.
3956 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
3958 *) Fix test data to work with the above.
3961 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
3962 was already fixed by Eric for 0.9.1 it seems.
3963 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
3965 *) Autodetect FreeBSD3.
3968 *) Fix various bugs in Configure. This affects the following platforms:
3972 unixware-2.0-pentium
3976 *) Eliminate generated files from CVS. Reorder tests to regenerate files
3977 before they are needed.
3980 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
3984 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
3986 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
3987 changed SSLeay to OpenSSL in version strings.
3988 [Ralf S. Engelschall]
3990 *) Some fixups to the top-level documents.
3993 *) Fixed the nasty bug where rsaref.h was not found under compile-time
3994 because the symlink to include/ was missing.
3995 [Ralf S. Engelschall]
3997 *) Incorporated the popular no-RSA/DSA-only patches
3998 which allow to compile a RSA-free SSLeay.
3999 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4001 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4002 when "ssleay" is still not found.
4003 [Ralf S. Engelschall]
4005 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4006 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4008 *) Updated the README file.
4009 [Ralf S. Engelschall]
4011 *) Added various .cvsignore files in the CVS repository subdirs
4012 to make a "cvs update" really silent.
4013 [Ralf S. Engelschall]
4015 *) Recompiled the error-definition header files and added
4016 missing symbols to the Win32 linker tables.
4017 [Ralf S. Engelschall]
4019 *) Cleaned up the top-level documents;
4020 o new files: CHANGES and LICENSE
4021 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4022 o merged COPYRIGHT into LICENSE
4023 o removed obsolete TODO file
4024 o renamed MICROSOFT to INSTALL.W32
4025 [Ralf S. Engelschall]
4027 *) Removed dummy files from the 0.9.1b source tree:
4028 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4029 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4030 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4031 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4032 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4033 [Ralf S. Engelschall]
4035 *) Added various platform portability fixes.
4038 *) The Genesis of the OpenSSL rpject:
4039 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4040 Young and Tim J. Hudson created while they were working for C2Net until
4042 [The OpenSSL Project]
4045 Changes between 0.9.0b and 0.9.1b [not released]
4047 *) Updated a few CA certificates under certs/
4050 *) Changed some BIGNUM api stuff.
4053 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4054 DGUX x86, Linux Alpha, etc.
4057 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4058 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4062 *) Add -strparse option to asn1pars program which parses nested
4064 [Dr Stephen Henson <shenson@bigfoot.com>]
4066 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4069 *) DSA fix for "ca" program.
4072 *) Added "-genkey" option to "dsaparam" program.
4075 *) Added RIPE MD160 (rmd160) message digest.
4078 *) Added -a (all) option to "ssleay version" command.
4081 *) Added PLATFORM define which is the id given to Configure.
4084 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4087 *) Extended the ASN.1 parser routines.
4090 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4093 *) Added a BN_CTX to the BN library.
4096 *) Fixed the weak key values in DES library
4099 *) Changed API in EVP library for cipher aliases.
4102 *) Added support for RC2/64bit cipher.
4105 *) Converted the lhash library to the crypto/mem.c functions.
4108 *) Added more recognized ASN.1 object ids.
4111 *) Added more RSA padding checks for SSL/TLS.
4114 *) Added BIO proxy/filter functionality.
4117 *) Added extra_certs to SSL_CTX which can be used
4118 send extra CA certificates to the client in the CA cert chain sending
4119 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4122 *) Now Fortezza is denied in the authentication phase because
4123 this is key exchange mechanism is not supported by SSLeay at all.
4126 *) Additional PKCS1 checks.
4129 *) Support the string "TLSv1" for all TLS v1 ciphers.
4132 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4133 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4136 *) Fixed a few memory leaks.
4139 *) Fixed various code and comment typos.
4142 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4143 bytes sent in the client random.
4144 [Edward Bishop <ebishop@spyglass.com>]