4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7 to data. This was previously part of the PKCS7 ASN1 code. This
8 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
9 [Steve Henson, reported by Kenneth R. Robinette
10 <support@securenetterm.com>]
12 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13 routines: without these tracing memory leaks is very painful.
14 Fix leaks in PKCS12 and PKCS7 routines.
17 *) Fix for Irix with NO_ASM.
18 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
20 *) Add some EVP_add_digest_alias registrations (as found in
21 OpenSSL_add_all_digests()), to SSL_library_init()
22 aka OpenSSL_add_ssl_algorithms(). This provides improved
23 compatibility with peers using X.509 certificates
24 with unconventional AlgorithmIdentifier OIDs.
27 *) ./config script fixes.
28 [Ulf Moeller, Richard Levitte]
30 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
31 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
32 effectively meant GeneralizedTime would never be used. Now it
33 is initialised to -1 but X509_time_adj() now has to check the value
34 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
35 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
36 [Steve Henson, reported by Kenneth R. Robinette
37 <support@securenetterm.com>]
39 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
40 result in a zero length in the ASN1_INTEGER structure which was
41 not consistent with the structure when d2i_ASN1_INTEGER() was used
42 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
43 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
44 where it did not print out a minus for negative ASN1_INTEGER.
47 *) Fix 'openssl passwd -1'.
50 *) Add summary printout to ocsp utility. The various functions which
51 convert status values to strings have been renamed to:
52 OCSP_response_status_str(), OCSP_cert_status_str() and
53 OCSP_crl_reason_str() and are no longer static. New options
54 to verify nonce values and to disable verification. OCSP response
55 printout format cleaned up.
58 *) Add additional OCSP certificate checks. These are those specified
59 in RFC2560. This consists of two separate checks: the CA of the
60 certificate being checked must either be the OCSP signer certificate
61 or the issuer of the OCSP signer certificate. In the latter case the
62 OCSP signer certificate must contain the OCSP signing extended key
63 usage. This check is performed by attempting to match the OCSP
64 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
65 in the OCSP_CERTID structures of the response.
68 *) Initial OCSP certificate verification added to OCSP_basic_verify()
69 and related routines. This uses the standard OpenSSL certificate
70 verify routines to perform initial checks (just CA validity) and
71 to obtain the certificate chain. Then additional checks will be
72 performed on the chain. Currently the root CA is checked to see
73 if it is explicitly trusted for OCSP signing. This is used to set
74 a root CA as a global signing root: that is any certificate that
75 chains to that CA is an acceptable OCSP signing certificate.
78 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
79 extensions from a separate configuration file.
80 As when reading extensions from the main configuration file,
81 the '-extensions ...' option may be used for specifying the
83 [Massimiliano Pala <madwolf@comune.modena.it>]
85 *) Change PKCS12_key_gen_asc() so it can cope with non null
86 terminated strings whose length is passed in the passlen
87 parameter, for example from PEM callbacks. This was done
88 by adding an extra length parameter to asc2uni().
89 [Steve Henson, reported by <oddissey@samsung.co.kr>]
91 *) New OCSP utility. Allows OCSP requests to be generated or
92 read. The request can be sent to a responder and the output
93 parsed, outputed or printed in text form. Not complete yet:
94 still needs to check the OCSP response validity.
97 *) New subcommands for 'openssl ca':
98 'openssl ca -status <serial>' prints the status of the cert with
99 the given serial number (according to the index file).
100 'openssl ca -updatedb' updates the expiry status of certificates
102 [Massimiliano Pala <madwolf@comune.modena.it>]
104 *) New '-newreq-nodes' command option to CA.pl. This is like
105 '-newreq', but calls 'openssl req' with the '-nodes' option
106 so that the resulting key is not encrypted.
107 [Damien Miller <djm@mindrot.org>]
109 *) New configuration for the GNU Hurd.
110 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
112 *) Initial code to implement OCSP basic response verify. This
113 is currently incomplete. Currently just finds the signer's
114 certificate and verifies the signature on the response.
117 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
118 value of OPENSSLDIR. This is available via the new '-d' option
119 to 'openssl version', and is also included in 'openssl version -a'.
122 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
123 call failed, free the DSA structure.
126 *) Allowing defining memory allocation callbacks that will be given
127 file name and line number information in additional arguments
128 (a const char* and an int). The basic functionality remains, as
129 well as the original possibility to just replace malloc(),
130 realloc() and free() by functions that do not know about these
131 additional arguments. To register and find out the current
132 settings for extended allocation functions, the following
133 functions are provided:
135 CRYPTO_set_mem_ex_functions
136 CRYPTO_set_locked_mem_ex_functions
137 CRYPTO_get_mem_ex_functions
138 CRYPTO_get_locked_mem_ex_functions
140 These work the same way as CRYPTO_set_mem_functions and friends.
141 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
142 extended allocation function is enabled.
143 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
144 a conventional allocation function is enabled.
145 [Richard Levitte, Bodo Moeller]
147 *) Fix to uni2asc() to cope with zero length Unicode strings.
148 These are present in some PKCS#12 files.
151 *) Finish off removing the remaining LHASH function pointer casts.
152 There should no longer be any prototype-casting required when using
153 the LHASH abstraction, and any casts that remain are "bugs". See
154 the callback types and macros at the head of lhash.h for details
155 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
158 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
159 If an EGD or PRNGD is running and enough entropy is returned, automatic
160 seeding like with /dev/[u]random will be performed.
161 Positions tried are: /etc/entropy, /var/run/egd-pool.
164 *) Change the Unix RAND_poll() variant to be able to poll several
165 random devices, as specified by DEVRANDOM, until a sufficient amount
166 of data has been collected. We spend at most 10 ms on each file
167 (select timeout) and read in non-blocking mode. DEVRANDOM now
168 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
169 (previously it was just the string "/dev/urandom"), so on typical
170 platforms the 10 ms delay will never occur.
171 Also separate out the Unix variant to its own file, rand_unix.c.
172 For VMS, there's a currently-empty rand_vms.c.
175 *) Move OCSP client related routines to ocsp_cl.c. These
176 provide utility functions which an application needing
177 to issue a request to an OCSP responder and analyse the
178 response will typically need: as opposed to those which an
179 OCSP responder itself would need which will be added later.
181 OCSP_request_sign() signs an OCSP request with an API similar
182 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
183 response. OCSP_response_get1_basic() extracts basic response
184 from response. OCSP_resp_find_status(): finds and extracts status
185 information from an OCSP_CERTID structure (which will be created
186 when the request structure is built). These are built from lower
187 level functions which work on OCSP_SINGLERESP structures but
188 wont normally be used unless the application wishes to examine
189 extensions in the OCSP response for example.
191 Replace nonce routines with a pair of functions.
192 OCSP_request_add1_nonce() adds a nonce value and optionally
193 generates a random value. OCSP_check_nonce() checks the
194 validity of the nonce in an OCSP response.
197 *) Change function OCSP_request_add() to OCSP_request_add0_id().
198 This doesn't copy the supplied OCSP_CERTID and avoids the
199 need to free up the newly created id. Change return type
200 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
201 This can then be used to add extensions to the request.
202 Deleted OCSP_request_new(), since most of its functionality
203 is now in OCSP_REQUEST_new() (and the case insensitive name
204 clash) apart from the ability to set the request name which
205 will be added elsewhere.
208 *) Update OCSP API. Remove obsolete extensions argument from
209 various functions. Extensions are now handled using the new
210 OCSP extension code. New simple OCSP HTTP function which
211 can be used to send requests and parse the response.
214 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
215 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
216 uses the special reorder version of SET OF to sort the attributes
217 and reorder them to match the encoded order. This resolves a long
218 standing problem: a verify on a PKCS7 structure just after signing
219 it used to fail because the attribute order did not match the
220 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
221 it uses the received order. This is necessary to tolerate some broken
222 software that does not order SET OF. This is handled by encoding
223 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
224 to produce the required SET OF.
227 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
228 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
229 files to get correct declarations of the ASN.1 item variables.
232 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
233 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
234 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
235 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
236 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
237 ASN1_ITEM and no wrapper functions.
240 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
241 replace the old function pointer based I/O routines. Change most of
242 the *_d2i_bio() and *_d2i_fp() functions to use these.
245 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
246 lines, recognice more "algorithms" that can be deselected, and make
247 it complain about algorithm deselection that isn't recognised.
250 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
251 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
252 to use new functions. Add NO_ASN1_OLD which can be set to remove
253 some old style ASN1 functions: this can be used to determine if old
254 code will still work when these eventually go away.
257 *) New extension functions for OCSP structures, these follow the
258 same conventions as certificates and CRLs.
261 *) New function X509V3_add1_i2d(). This automatically encodes and
262 adds an extension. Its behaviour can be customised with various
263 flags to append, replace or delete. Various wrappers added for
264 certifcates and CRLs.
267 *) Fix to avoid calling the underlying ASN1 print routine when
268 an extension cannot be parsed. Correct a typo in the
269 OCSP_SERVICELOC extension. Tidy up print OCSP format.
272 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
273 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
274 when writing a 32767 byte record.
275 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
277 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
278 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
280 (RSA objects have a reference count access to which is protected
281 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
282 so they are meant to be shared between threads.)
283 [Bodo Moeller, Geoff Thorpe; original patch submitted by
284 "Reddie, Steven" <Steven.Reddie@ca.com>]
286 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
287 entries for variables.
290 *) Fix a deadlock in CRYPTO_mem_leaks().
293 *) Add functionality to apps/openssl.c for detecting locking
294 problems: As the program is single-threaded, all we have
295 to do is register a locking callback using an array for
296 storing which locks are currently held by the program.
299 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
300 SSL_get_ex_data_X509_STORE_idx(), which is used in
301 ssl_verify_cert_chain() and thus can be called at any time
302 during TLS/SSL handshakes so that thread-safety is essential.
303 Unfortunately, the ex_data design is not at all suited
304 for multi-threaded use, so it probably should be abolished.
307 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
308 [Broadcom, tweaked and integrated by Geoff Thorpe]
310 *) Move common extension printing code to new function
311 X509V3_print_extensions(). Reorganise OCSP print routines and
312 implement some needed OCSP ASN1 functions. Add OCSP extensions.
315 *) New function X509_signature_print() to remove duplication in some
319 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
320 set (this was treated exactly the same as SET OF previously). This
321 is used to reorder the STACK representing the structure to match the
322 encoding. This will be used to get round a problem where a PKCS7
323 structure which was signed could not be verified because the STACK
324 order did not reflect the encoded order.
327 *) Reimplement the OCSP ASN1 module using the new code.
330 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
331 for its ASN1 operations. The old style function pointers still exist
332 for now but they will eventually go away.
335 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
336 completely replaces the old ASN1 functionality with a table driven
337 encoder and decoder which interprets an ASN1_ITEM structure describing
338 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
339 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
340 has also been converted to the new form.
343 *) Change BN_mod_exp_recp so that negative moduli are tolerated
344 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
345 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
349 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
350 of not touching the result's sign bit.
353 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
357 *) Changed the LHASH code to use prototypes for callbacks, and created
358 macros to declare and implement thin (optionally static) functions
359 that provide type-safety and avoid function pointer casting for the
360 type-specific callbacks.
363 *) Use better test patterns in bntest.
366 *) Added Kerberos Cipher Suites to be used with TLS, as written in
368 [Veers Staats <staatsvr@asc.hpc.mil>,
369 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
371 *) rand_win.c fix for Borland C.
374 *) BN_rshift bugfix for n == 0.
377 *) Reformat the FAQ so the different questions and answers can be divided
378 in sections depending on the subject.
381 *) Have the zlib compression code load ZLIB.DLL dynamically under
385 *) New function BN_mod_sqrt for computing square roots modulo a prime
386 (using the probabilistic Tonelli-Shanks algorithm unless
387 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
388 be handled deterministically).
389 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
391 *) Store verify_result within SSL_SESSION also for client side to
392 avoid potential security hole. (Re-used sessions on the client side
393 always resulted in verify_result==X509_V_OK, not using the original
394 result of the server certificate verification.)
397 *) Make BN_mod_inverse faster by explicitly handling small quotients
398 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
399 512 bits], about 30% for larger ones [1024 or 2048 bits].)
402 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
403 SSL3_RT_APPLICATION_DATA, return 0.
404 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
408 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
409 releases, have been re-implemented by renaming the previous
410 implementations of ssl2_read and ssl3_read to ssl2_read_internal
411 and ssl3_read_internal, respectively, and adding 'peek' parameters
412 to them. The new ssl[23]_{read,peek} functions are calls to
413 ssl[23]_read_internal with the 'peek' flag set appropriately.
414 A 'peek' parameter has also been added to ssl3_read_bytes, which
415 does the actual work for ssl3_read_internal.
418 *) New function BN_kronecker.
421 *) Fix BN_gcd so that it works on negative inputs; the result is
422 positive unless both parameters are zero.
423 Previously something reasonably close to an infinite loop was
424 possible because numbers could be growing instead of shrinking
425 in the implementation of Euclid's algorithm.
428 *) Fix BN_is_word() and BN_is_one() macros to take into account the
429 sign of the number in question.
431 Fix BN_is_word(a,w) to work correctly for w == 0.
433 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
434 because its test if the absolute value of 'a' equals 'w'.
435 Note that BN_abs_is_word does *not* handle w == 0 reliably;
436 it exists mostly for use in the implementations of BN_is_zero(),
437 BN_is_one(), and BN_is_word().
440 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
441 the method-specific "init()" handler. Also clean up ex_data after
442 calling the method-specific "finish()" handler. Previously, this was
443 happening the other way round.
446 *) New function BN_swap.
449 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
450 the exponentiation functions are more likely to produce reasonable
451 results on negative inputs.
454 *) Change BN_mod_mul so that the result is always non-negative.
455 Previously, it could be negative if one of the factors was negative;
456 I don't think anyone really wanted that behaviour.
459 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
460 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
461 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
462 and add new functions:
475 These functions always generate non-negative results.
477 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
478 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
480 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
481 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
483 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
485 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
486 was actually never needed) and in BN_mul(). The removal in BN_mul()
487 required a small change in bn_mul_part_recursive() and the addition
488 of the functions bn_cmp_part_words(), bn_sub_part_words() and
489 bn_add_part_words(), which do the same thing as bn_cmp_words(),
490 bn_sub_words() and bn_add_words() except they take arrays with
494 *) In 'openssl passwd', verify passwords read from the terminal
495 unless the '-salt' option is used (which usually means that
496 verification would just waste user's time since the resulting
497 hash is going to be compared with some given password hash)
498 or the new '-noverify' option is used.
500 This is an incompatible change, but it does not affect
501 non-interactive use of 'openssl passwd' (passwords on the command
502 line, '-stdin' option, '-in ...' option) and thus should not
506 *) Remove all references to RSAref, since there's no more need for it.
509 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
510 The previous value, 12, was not always sufficient for BN_mod_exp().
513 *) Make DSO load along a path given through an environment variable
514 (SHLIB_PATH) with shl_load().
517 *) Constify the ENGINE code as a result of BIGNUM constification.
518 Also constify the RSA code and most things related to it. In a
519 few places, most notable in the depth of the ASN.1 code, ugly
520 casts back to non-const were required (to be solved at a later
524 *) Make it so the openssl application has all engines loaded by default.
527 *) Constify the BIGNUM routines a little more.
530 *) Make sure that shared libraries get the internal name engine with
531 the full version number and not just 0. This should mark the
532 shared libraries as not backward compatible. Of course, this should
533 be changed again when we can guarantee backward binary compatibility.
536 *) Add the following functions:
542 ENGINE_load_builtin_engines()
544 That way, an application can itself choose if external engines that
545 are built-in in OpenSSL shall ever be used or not. The benefit is
546 that applications won't have to be linked with libdl or other dso
547 libraries unless it's really needed.
549 Changed 'openssl engine' to load all engines on demand.
550 Changed the engine header files to avoid the duplication of some
551 declarations (they differed!).
554 *) 'openssl engine' can now list capabilities.
557 *) Better error reporting in 'openssl engine'.
560 *) Never call load_dh_param(NULL) in s_server.
563 *) Add engine application. It can currently list engines by name and
564 identity, and test if they are actually available.
567 *) Add support for shared libraries under Irix.
568 [Albert Chin-A-Young <china@thewrittenword.com>]
570 *) Improve RPM specification file by forcing symbolic linking and making
571 sure the installed documentation is also owned by root.root.
572 [Damien Miller <djm@mindrot.org>]
574 *) Add configuration option to build on Linux on both big-endian and
576 [Ralf Baechle <ralf@uni-koblenz.de>]
578 *) Give the OpenSSL applications more possibilities to make use of
579 keys (public as well as private) handled by engines.
582 *) Add OCSP code that comes from CertCo.
585 *) Add VMS support for the Rijndael code.
588 *) Added untested support for Nuron crypto accelerator.
591 *) Add support for external cryptographic devices. This code was
592 previously distributed separately as the "engine" branch.
593 [Geoff Thorpe, Richard Levitte]
595 *) Rework the filename-translation in the DSO code. It is now possible to
596 have far greater control over how a "name" is turned into a filename
597 depending on the operating environment and any oddities about the
598 different shared library filenames on each system.
601 *) Support threads on FreeBSD-elf in Configure.
604 *) Add the possibility to create shared libraries on HP-UX
607 *) Fix for SHA1 assembly problem with MASM: it produces
608 warnings about corrupt line number information when assembling
609 with debugging information. This is caused by the overlapping
611 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
614 NCONF_get_number() has no error checking at all. As a replacement,
615 NCONF_get_number_e() is defined (_e for "error checking") and is
616 promoted strongly. The old NCONF_get_number is kept around for
617 binary backward compatibility.
618 Make it possible for methods to load from something other than a BIO,
619 by providing a function pointer that is given a name instead of a BIO.
620 For example, this could be used to load configuration data from an
624 *) Fix typo in get_cert_by_subject() in by_dir.c
625 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
627 *) Rework the system to generate shared libraries:
629 - Make note of the expected extension for the shared libraries and
630 if there is a need for symbolic links from for example libcrypto.so.0
631 to libcrypto.so.0.9.7. There is extended info in Configure for
634 - Make as few rebuilds of the shared libraries as possible.
636 - Still avoid linking the OpenSSL programs with the shared libraries.
638 - When installing, install the shared libraries separately from the
642 *) Fix for non blocking accept BIOs. Added new I/O special reason
643 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
644 with non blocking I/O was not possible because no retry code was
645 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
649 *) Added the beginnings of Rijndael support.
652 *) Fix for bug in DirectoryString mask setting. Add support for
653 X509_NAME_print_ex() in 'req' and X509_print_ex() function
654 to allow certificate printing to more controllable, additional
655 'certopt' option to 'x509' to allow new printing options to be
659 *) Clean old EAY MD5 hack from e_os.h.
662 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
664 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
665 and not in SSL_clear because the latter is also used by the
666 accept/connect functions; previously, the settings made by
667 SSL_set_read_ahead would be lost during the handshake.
668 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
670 *) Correct util/mkdef.pl to be selective about disabled algorithms.
671 Previously, it would create entries for disableed algorithms no
675 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
677 *) In ssl23_get_client_hello, generate an error message when faced
678 with an initial SSL 3.0/TLS record that is too small to contain the
679 first two bytes of the ClientHello message, i.e. client_version.
680 (Note that this is a pathologic case that probably has never happened
681 in real life.) The previous approach was to use the version number
682 from the record header as a substitute; but our protocol choice
683 should not depend on that one because it is not authenticated
684 by the Finished messages.
687 *) More robust randomness gathering functions for Windows.
688 [Jeffrey Altman <jaltman@columbia.edu>]
690 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
691 not set then we don't setup the error code for issuer check errors
692 to avoid possibly overwriting other errors which the callback does
693 handle. If an application does set the flag then we assume it knows
694 what it is doing and can handle the new informational codes
698 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
699 a general "ANY" type, as such it should be able to decode anything
700 including tagged types. However it didn't check the class so it would
701 wrongly interpret tagged types in the same way as their universal
702 counterpart and unknown types were just rejected. Changed so that the
703 tagged and unknown types are handled in the same way as a SEQUENCE:
704 that is the encoding is stored intact. There is also a new type
705 "V_ASN1_OTHER" which is used when the class is not universal, in this
706 case we have no idea what the actual type is so we just lump them all
710 *) On VMS, stdout may very well lead to a file that is written to
711 in a record-oriented fashion. That means that every write() will
712 write a separate record, which will be read separately by the
713 programs trying to read from it. This can be very confusing.
715 The solution is to put a BIO filter in the way that will buffer
716 text until a linefeed is reached, and then write everything a
717 line at a time, so every record written will be an actual line,
718 not chunks of lines and not (usually doesn't happen, but I've
719 seen it once) several lines in one record. BIO_f_linebuffer() is
722 Currently, it's a VMS-only method, because that's where it has
723 been tested well enough.
726 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
727 it can return incorrect results.
728 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
729 but it was in 0.9.6-beta[12].)
732 *) Disable the check for content being present when verifying detached
733 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
734 include zero length content when signing messages.
737 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
738 BIO_ctrl (for BIO pairs).
741 *) Add DSO method for VMS.
744 *) Bug fix: Montgomery multiplication could produce results with the
748 *) Add RPM specification openssl.spec and modify it to build three
749 packages. The default package contains applications, application
750 documentation and run-time libraries. The devel package contains
751 include files, static libraries and function documentation. The
752 doc package contains the contents of the doc directory. The original
753 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
756 *) Add a large number of documentation files for many SSL routines.
757 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
759 *) Add a configuration entry for Sony News 4.
760 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
762 *) Don't set the two most significant bits to one when generating a
763 random number < q in the DSA library.
766 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
767 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
768 the underlying transport is blocking) if a handshake took place.
769 (The default behaviour is needed by applications such as s_client
770 and s_server that use select() to determine when to use SSL_read;
771 but for applications that know in advance when to expect data, it
772 just makes things more complicated.)
775 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
779 *) Add a few more EBCDIC conditionals that make `req' and `x509'
780 work better on such systems.
781 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
783 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
784 Update PKCS12_parse() so it copies the friendlyName and the
785 keyid to the certificates aux info.
788 *) Fix bug in PKCS7_verify() which caused an infinite loop
789 if there was more than one signature.
790 [Sven Uszpelkat <su@celocom.de>]
792 *) Major change in util/mkdef.pl to include extra information
793 about each symbol, as well as presentig variables as well
794 as functions. This change means that there's n more need
795 to rebuild the .num files when some algorithms are excluded.
798 *) Allow the verify time to be set by an application,
799 rather than always using the current time.
802 *) Phase 2 verify code reorganisation. The certificate
803 verify code now looks up an issuer certificate by a
804 number of criteria: subject name, authority key id
805 and key usage. It also verifies self signed certificates
806 by the same criteria. The main comparison function is
807 X509_check_issued() which performs these checks.
809 Lot of changes were necessary in order to support this
810 without completely rewriting the lookup code.
812 Authority and subject key identifier are now cached.
814 The LHASH 'certs' is X509_STORE has now been replaced
815 by a STACK_OF(X509_OBJECT). This is mainly because an
816 LHASH can't store or retrieve multiple objects with
819 As a result various functions (which were all internal
820 use only) have changed to handle the new X509_STORE
821 structure. This will break anything that messed round
822 with X509_STORE internally.
824 The functions X509_STORE_add_cert() now checks for an
825 exact match, rather than just subject name.
827 The X509_STORE API doesn't directly support the retrieval
828 of multiple certificates matching a given criteria, however
829 this can be worked round by performing a lookup first
830 (which will fill the cache with candidate certificates)
831 and then examining the cache for matches. This is probably
832 the best we can do without throwing out X509_LOOKUP
833 entirely (maybe later...).
835 The X509_VERIFY_CTX structure has been enhanced considerably.
837 All certificate lookup operations now go via a get_issuer()
838 callback. Although this currently uses an X509_STORE it
839 can be replaced by custom lookups. This is a simple way
840 to bypass the X509_STORE hackery necessary to make this
841 work and makes it possible to use more efficient techniques
842 in future. A very simple version which uses a simple
843 STACK for its trusted certificate store is also provided
844 using X509_STORE_CTX_trusted_stack().
846 The verify_cb() and verify() callbacks now have equivalents
847 in the X509_STORE_CTX structure.
849 X509_STORE_CTX also has a 'flags' field which can be used
850 to customise the verify behaviour.
853 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
854 excludes S/MIME capabilities.
857 *) When a certificate request is read in keep a copy of the
858 original encoding of the signed data and use it when outputing
859 again. Signatures then use the original encoding rather than
860 a decoded, encoded version which may cause problems if the
861 request is improperly encoded.
864 *) For consistency with other BIO_puts implementations, call
865 buffer_write(b, ...) directly in buffer_puts instead of calling
868 In BIO_puts, increment b->num_write as in BIO_write.
869 [Peter.Sylvester@EdelWeb.fr]
871 *) Fix BN_mul_word for the case where the word is 0. (We have to use
872 BN_zero, we may not return a BIGNUM with an array consisting of
876 *) Avoid calling abort() from within the library when problems are
877 detected, except if preprocessor symbols have been defined
878 (such as REF_CHECK, BN_DEBUG etc.).
881 *) New openssl application 'rsautl'. This utility can be
882 used for low level RSA operations. DER public key
883 BIO/fp routines also added.
886 *) New Configure entry and patches for compiling on QNX 4.
887 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
889 *) A demo state-machine implementation was sponsored by
890 Nuron (http://www.nuron.com/) and is now available in
894 *) New options added to the 'dgst' utility for signature
895 generation and verification.
898 *) Unrecognized PKCS#7 content types are now handled via a
899 catch all ASN1_TYPE structure. This allows unsupported
900 types to be stored as a "blob" and an application can
901 encode and decode it manually.
904 *) Fix various signed/unsigned issues to make a_strex.c
906 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
908 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
909 length if passed a buffer. ASN1_INTEGER_to_BN failed
910 if passed a NULL BN and its argument was negative.
911 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
913 *) Modification to PKCS#7 encoding routines to output definite
914 length encoding. Since currently the whole structures are in
915 memory there's not real point in using indefinite length
916 constructed encoding. However if OpenSSL is compiled with
917 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
920 *) Added BIO_vprintf() and BIO_vsnprintf().
923 *) Added more prefixes to parse for in the the strings written
924 through a logging bio, to cover all the levels that are available
925 through syslog. The prefixes are now:
927 PANIC, EMERG, EMR => LOG_EMERG
928 ALERT, ALR => LOG_ALERT
929 CRIT, CRI => LOG_CRIT
930 ERROR, ERR => LOG_ERR
931 WARNING, WARN, WAR => LOG_WARNING
932 NOTICE, NOTE, NOT => LOG_NOTICE
933 INFO, INF => LOG_INFO
934 DEBUG, DBG => LOG_DEBUG
936 and as before, if none of those prefixes are present at the
937 beginning of the string, LOG_ERR is chosen.
939 On Win32, the LOG_* levels are mapped according to this:
941 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
942 LOG_WARNING => EVENTLOG_WARNING_TYPE
943 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
947 *) Made it possible to reconfigure with just the configuration
948 argument "reconf" or "reconfigure". The command line arguments
949 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
950 and are retrieved from there when reconfiguring.
954 [Assar Westerlund <assar@sics.se>, Richard Levitte]
956 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
959 *) The obj_dat.pl script was messing up the sorting of object
960 names. The reason was that it compared the quoted version
961 of strings as a result "OCSP" > "OCSP Signing" because
962 " > SPACE. Changed script to store unquoted versions of
963 names and add quotes on output. It was also omitting some
964 names from the lookup table if they were given a default
965 value (that is if SN is missing it is given the same
966 value as LN and vice versa), these are now added on the
967 grounds that if an object has a name we should be able to
968 look it up. Finally added warning output when duplicate
969 short or long names are found.
972 *) Changes needed for Tandem NSK.
973 [Scott Uroff <scott@xypro.com>]
975 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
976 RSA_padding_check_SSLv23(), special padding was never detected
977 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
978 version rollback attacks was not effective.
980 In s23_clnt.c, don't use special rollback-attack detection padding
981 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
982 client; similarly, in s23_srvr.c, don't do the rollback check if
983 SSL 2.0 is the only protocol enabled in the server.
986 *) Make it possible to get hexdumps of unprintable data with 'openssl
987 asn1parse'. By implication, the functions ASN1_parse_dump() and
988 BIO_dump_indent() are added.
991 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
992 these print out strings and name structures based on various
993 flags including RFC2253 support and proper handling of
994 multibyte characters. Added options to the 'x509' utility
995 to allow the various flags to be set.
998 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
999 Also change the functions X509_cmp_current_time() and
1000 X509_gmtime_adj() work with an ASN1_TIME structure,
1001 this will enable certificates using GeneralizedTime in validity
1002 dates to be checked.
1005 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1006 negative public key encodings) on by default,
1007 NO_NEG_PUBKEY_BUG can be set to disable it.
1010 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1011 content octets. An i2c_ASN1_OBJECT is unnecessary because
1012 the encoding can be trivially obtained from the structure.
1015 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1016 not read locks (CRYPTO_r_[un]lock).
1019 *) A first attempt at creating official support for shared
1020 libraries through configuration. I've kept it so the
1021 default is static libraries only, and the OpenSSL programs
1022 are always statically linked for now, but there are
1023 preparations for dynamic linking in place.
1024 This has been tested on Linux and True64.
1027 *) Randomness polling function for Win9x, as described in:
1028 Peter Gutmann, Software Generation of Practically Strong
1032 *) Fix so PRNG is seeded in req if using an already existing
1036 *) New options to smime application. -inform and -outform
1037 allow alternative formats for the S/MIME message including
1038 PEM and DER. The -content option allows the content to be
1039 specified separately. This should allow things like Netscape
1040 form signing output easier to verify.
1043 *) Fix the ASN1 encoding of tags using the 'long form'.
1046 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1047 STRING types. These convert content octets to and from the
1048 underlying type. The actual tag and length octets are
1049 already assumed to have been read in and checked. These
1050 are needed because all other string types have virtually
1051 identical handling apart from the tag. By having versions
1052 of the ASN1 functions that just operate on content octets
1053 IMPLICIT tagging can be handled properly. It also allows
1054 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1055 and ASN1_INTEGER are identical apart from the tag.
1058 *) Change the handling of OID objects as follows:
1060 - New object identifiers are inserted in objects.txt, following
1061 the syntax given in objects.README.
1062 - objects.pl is used to process obj_mac.num and create a new
1064 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1067 This is currently kind of a hack, and the perl code in objects.pl
1068 isn't very elegant, but it works as I intended. The simplest way
1069 to check that it worked correctly is to look in obj_dat.h and
1070 check the array nid_objs and make sure the objects haven't moved
1071 around (this is important!). Additions are OK, as well as
1072 consistent name changes.
1075 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1078 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1079 The given file adds to whatever has already been seeded into the
1080 random pool through the RANDFILE configuration file option or
1081 environment variable, or the default random state file.
1084 *) mkstack.pl now sorts each macro group into lexical order.
1085 Previously the output order depended on the order the files
1086 appeared in the directory, resulting in needless rewriting
1090 *) Patches to make OpenSSL compile under Win32 again. Mostly
1091 work arounds for the VC++ problem that it treats func() as
1092 func(void). Also stripped out the parts of mkdef.pl that
1093 added extra typesafe functions: these no longer exist.
1096 *) Reorganisation of the stack code. The macros are now all
1097 collected in safestack.h . Each macro is defined in terms of
1098 a "stack macro" of the form SKM_<name>(type, a, b). The
1099 DEBUG_SAFESTACK is now handled in terms of function casts,
1100 this has the advantage of retaining type safety without the
1101 use of additional functions. If DEBUG_SAFESTACK is not defined
1102 then the non typesafe macros are used instead. Also modified the
1103 mkstack.pl script to handle the new form. Needs testing to see
1104 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1105 the default if no major problems. Similar behaviour for ASN1_SET_OF
1106 and PKCS12_STACK_OF.
1109 *) When some versions of IIS use the 'NET' form of private key the
1110 key derivation algorithm is different. Normally MD5(password) is
1111 used as a 128 bit RC4 key. In the modified case
1112 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1113 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1114 as the old Netscape_RSA functions except they have an additional
1115 'sgckey' parameter which uses the modified algorithm. Also added
1116 an -sgckey command line option to the rsa utility. Thanks to
1117 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1118 algorithm to openssl-dev.
1121 *) The evp_local.h macros were using 'c.##kname' which resulted in
1122 invalid expansion on some systems (SCO 5.0.5 for example).
1123 Corrected to 'c.kname'.
1124 [Phillip Porch <root@theporch.com>]
1126 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1127 a STACK of email addresses from a certificate or request, these look
1128 in the subject name and the subject alternative name extensions and
1129 omit any duplicate addresses.
1132 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1133 This makes DSA verification about 2 % faster.
1136 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1137 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1138 plus overhead for 1024 bit moduli).
1139 This makes exponentiations about 0.5 % faster for 1024 bit
1140 exponents (as measured by "openssl speed rsa2048").
1143 *) Rename memory handling macros to avoid conflicts with other
1145 Malloc => OPENSSL_malloc
1146 Malloc_locked => OPENSSL_malloc_locked
1147 Realloc => OPENSSL_realloc
1148 Free => OPENSSL_free
1151 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1152 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1155 *) CygWin32 support.
1156 [John Jarvie <jjarvie@newsguy.com>]
1158 *) The type-safe stack code has been rejigged. It is now only compiled
1159 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1160 by default all type-specific stack functions are "#define"d back to
1161 standard stack functions. This results in more streamlined output
1162 but retains the type-safety checking possibilities of the original
1166 *) The STACK code has been cleaned up, and certain type declarations
1167 that didn't make a lot of sense have been brought in line. This has
1168 also involved a cleanup of sorts in safestack.h to more correctly
1169 map type-safe stack functions onto their plain stack counterparts.
1170 This work has also resulted in a variety of "const"ifications of
1171 lots of the code, especially "_cmp" operations which should normally
1172 be prototyped with "const" parameters anyway.
1175 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1176 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1177 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1178 where all of 'md' is used each time the PRNG is used, but 'state'
1179 is used only indexed by a cyclic counter. As entropy may not be
1180 well distributed from the beginning, 'md' is important as a
1181 chaining variable. However, the output function chains only half
1182 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1183 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1184 in all of 'state' being rewritten, with the new values depending
1185 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1188 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1189 the handshake is continued after ssl_verify_cert_chain();
1190 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1191 can lead to 'unexplainable' connection aborts later.
1192 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1194 *) Major EVP API cipher revision.
1195 Add hooks for extra EVP features. This allows various cipher
1196 parameters to be set in the EVP interface. Support added for variable
1197 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1198 setting of RC2 and RC5 parameters.
1200 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1203 Remove lots of duplicated code from the EVP library. For example *every*
1204 cipher init() function handles the 'iv' in the same way according to the
1205 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1206 for CFB and OFB modes they zero ctx->num.
1208 New functionality allows removal of S/MIME code RC2 hack.
1210 Most of the routines have the same form and so can be declared in terms
1213 By shifting this to the top level EVP_CipherInit() it can be removed from
1214 all individual ciphers. If the cipher wants to handle IVs or keys
1215 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1218 Change lots of functions like EVP_EncryptUpdate() to now return a
1219 value: although software versions of the algorithms cannot fail
1220 any installed hardware versions can.
1223 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1224 this option is set, tolerate broken clients that send the negotiated
1225 protocol version number instead of the requested protocol version
1229 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1230 i.e. non-zero for export ciphersuites, zero otherwise.
1231 Previous versions had this flag inverted, inconsistent with
1232 rsa_tmp_cb (..._TMP_RSA_CB).
1233 [Bodo Moeller; problem reported by Amit Chopra]
1235 *) Add missing DSA library text string. Work around for some IIS
1236 key files with invalid SEQUENCE encoding.
1239 *) Add a document (doc/standards.txt) that list all kinds of standards
1240 and so on that are implemented in OpenSSL.
1243 *) Enhance c_rehash script. Old version would mishandle certificates
1244 with the same subject name hash and wouldn't handle CRLs at all.
1245 Added -fingerprint option to crl utility, to support new c_rehash
1249 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1252 *) Fix for SSL server purpose checking. Server checking was
1253 rejecting certificates which had extended key usage present
1254 but no ssl client purpose.
1255 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1257 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1258 is a little unclear about how a blank password is handled.
1259 Since the password in encoded as a BMPString with terminating
1260 double NULL a zero length password would end up as just the
1261 double NULL. However no password at all is different and is
1262 handled differently in the PKCS#12 key generation code. NS
1263 treats a blank password as zero length. MSIE treats it as no
1264 password on export: but it will try both on import. We now do
1265 the same: PKCS12_parse() tries zero length and no password if
1266 the password is set to "" or NULL (NULL is now a valid password:
1267 it wasn't before) as does the pkcs12 application.
1270 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1271 perror when PEM_read_bio_X509_REQ fails, the error message must
1272 be obtained from the error queue.
1275 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1276 it in ERR_remove_state if appropriate, and change ERR_get_state
1277 accordingly to avoid race conditions (this is necessary because
1278 thread_hash is no longer constant once set).
1281 *) Bugfix for linux-elf makefile.one.
1284 *) RSA_get_default_method() will now cause a default
1285 RSA_METHOD to be chosen if one doesn't exist already.
1286 Previously this was only set during a call to RSA_new()
1287 or RSA_new_method(NULL) meaning it was possible for
1288 RSA_get_default_method() to return NULL.
1291 *) Added native name translation to the existing DSO code
1292 that will convert (if the flag to do so is set) filenames
1293 that are sufficiently small and have no path information
1294 into a canonical native form. Eg. "blah" converted to
1295 "libblah.so" or "blah.dll" etc.
1298 *) New function ERR_error_string_n(e, buf, len) which is like
1299 ERR_error_string(e, buf), but writes at most 'len' bytes
1300 including the 0 terminator. For ERR_error_string_n, 'buf'
1302 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1304 *) CONF library reworked to become more general. A new CONF
1305 configuration file reader "class" is implemented as well as a
1306 new functions (NCONF_*, for "New CONF") to handle it. The now
1307 old CONF_* functions are still there, but are reimplemented to
1308 work in terms of the new functions. Also, a set of functions
1309 to handle the internal storage of the configuration data is
1310 provided to make it easier to write new configuration file
1311 reader "classes" (I can definitely see something reading a
1312 configuration file in XML format, for example), called _CONF_*,
1313 or "the configuration storage API"...
1315 The new configuration file reading functions are:
1317 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1318 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1320 NCONF_default, NCONF_WIN32
1322 NCONF_dump_fp, NCONF_dump_bio
1324 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1325 NCONF_new creates a new CONF object. This works in the same way
1326 as other interfaces in OpenSSL, like the BIO interface.
1327 NCONF_dump_* dump the internal storage of the configuration file,
1328 which is useful for debugging. All other functions take the same
1329 arguments as the old CONF_* functions wth the exception of the
1330 first that must be a `CONF *' instead of a `LHASH *'.
1332 To make it easer to use the new classes with the old CONF_* functions,
1333 the function CONF_set_default_method is provided.
1336 *) Add '-tls1' option to 'openssl ciphers', which was already
1337 mentioned in the documentation but had not been implemented.
1338 (This option is not yet really useful because even the additional
1339 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1342 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1343 OpenSSL-based applications) load shared libraries and bind to
1344 them in a portable way.
1345 [Geoff Thorpe, with contributions from Richard Levitte]
1347 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1349 *) Make sure _lrotl and _lrotr are only used with MSVC.
1351 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1352 (the default implementation of RAND_status).
1354 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1355 to '-clrext' (= clear extensions), as intended and documented.
1356 [Bodo Moeller; inconsistency pointed out by Michael Attili
1359 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1360 was larger than the MD block size.
1361 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1363 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1364 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1365 using the passed key: if the passed key was a private key the result
1366 of X509_print(), for example, would be to print out all the private key
1370 *) des_quad_cksum() byte order bug fix.
1371 [Ulf Möller, using the problem description in krb4-0.9.7, where
1372 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1374 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1376 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1378 *) For easily testing in shell scripts whether some command
1379 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1380 returns with exit code 0 iff no command of the given name is available.
1381 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1382 the output goes to stdout and nothing is printed to stderr.
1383 Additional arguments are always ignored.
1385 Since for each cipher there is a command of the same name,
1386 the 'no-cipher' compilation switches can be tested this way.
1388 ('openssl no-XXX' is not able to detect pseudo-commands such
1389 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1392 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1395 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1396 is set; it will be thrown away anyway because each handshake creates
1398 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1399 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1400 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1401 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1404 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1405 'Q' and 'R' lose their special meanings (quit/renegotiate).
1406 This is part of what -quiet does; unlike -quiet, -ign_eof
1407 does not suppress any output.
1410 *) Add compatibility options to the purpose and trust code. The
1411 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1412 accepts a certificate or CA, this was the previous behaviour,
1413 with all the associated security issues.
1415 X509_TRUST_COMPAT is the old trust behaviour: only and
1416 automatically trust self signed roots in certificate store. A
1417 new trust setting X509_TRUST_DEFAULT is used to specify that
1418 a purpose has no associated trust setting and it should instead
1419 use the value in the default purpose.
1422 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1423 and fix a memory leak.
1426 *) In util/mkerr.pl (which implements 'make errors'), preserve
1427 reason strings from the previous version of the .c file, as
1428 the default to have only downcase letters (and digits) in
1429 automatically generated reasons codes is not always appropriate.
1432 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1433 using strerror. Previously, ERR_reason_error_string() returned
1434 library names as reason strings for SYSerr; but SYSerr is a special
1435 case where small numbers are errno values, not library numbers.
1438 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1439 converts DSA parameters into DH parameters. (When creating parameters,
1440 DSA_generate_parameters is used.)
1443 *) Include 'length' (recommended exponent length) in C code generated
1444 by 'openssl dhparam -C'.
1447 *) The second argument to set_label in perlasm was already being used
1448 so couldn't be used as a "file scope" flag. Moved to third argument
1452 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1453 instead of RAND_bytes for encryption IVs and salts.
1456 *) Include RAND_status() into RAND_METHOD instead of implementing
1457 it only for md_rand.c Otherwise replacing the PRNG by calling
1458 RAND_set_rand_method would be impossible.
1461 *) Don't let DSA_generate_key() enter an infinite loop if the random
1462 number generation fails.
1465 *) New 'rand' application for creating pseudo-random output.
1468 *) Added configuration support for Linux/IA64
1469 [Rolf Haberrecker <rolf@suse.de>]
1471 *) Assembler module support for Mingw32.
1474 *) Shared library support for HPUX (in shlib/).
1475 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1477 *) Shared library support for Solaris gcc.
1478 [Lutz Behnke <behnke@trustcenter.de>]
1480 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1482 *) PKCS7_encrypt() was adding text MIME headers twice because they
1483 were added manually and by SMIME_crlf_copy().
1486 *) In bntest.c don't call BN_rand with zero bits argument.
1487 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1489 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1490 case was implemented. This caused BN_div_recp() to fail occasionally.
1493 *) Add an optional second argument to the set_label() in the perl
1494 assembly language builder. If this argument exists and is set
1495 to 1 it signals that the assembler should use a symbol whose
1496 scope is the entire file, not just the current function. This
1497 is needed with MASM which uses the format label:: for this scope.
1498 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1500 *) Change the ASN1 types so they are typedefs by default. Before
1501 almost all types were #define'd to ASN1_STRING which was causing
1502 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1506 *) Change names of new functions to the new get1/get0 naming
1507 convention: After 'get1', the caller owns a reference count
1508 and has to call ..._free; 'get0' returns a pointer to some
1509 data structure without incrementing reference counters.
1510 (Some of the existing 'get' functions increment a reference
1511 counter, some don't.)
1512 Similarly, 'set1' and 'add1' functions increase reference
1513 counters or duplicate objects.
1516 *) Allow for the possibility of temp RSA key generation failure:
1517 the code used to assume it always worked and crashed on failure.
1520 *) Fix potential buffer overrun problem in BIO_printf().
1521 [Ulf Möller, using public domain code by Patrick Powell; problem
1522 pointed out by David Sacerdote <das33@cornell.edu>]
1524 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1525 RAND_egd() and RAND_status(). In the command line application,
1526 the EGD socket can be specified like a seed file using RANDFILE
1530 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1531 Some CAs (e.g. Verisign) distribute certificates in this form.
1534 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1535 list to exclude them. This means that no special compilation option
1536 is needed to use anonymous DH: it just needs to be included in the
1540 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1541 EVP_MD_type. The old functionality is available in a new macro called
1542 EVP_MD_md(). Change code that uses it and update docs.
1545 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1546 where the 'void *' argument is replaced by a function pointer argument.
1547 Previously 'void *' was abused to point to functions, which works on
1548 many platforms, but is not correct. As these functions are usually
1549 called by macros defined in OpenSSL header files, most source code
1550 should work without changes.
1553 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1554 sections with information on -D... compiler switches used for
1555 compiling the library so that applications can see them. To enable
1556 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1557 must be defined. E.g.,
1558 #define OPENSSL_ALGORITHM_DEFINES
1559 #include <openssl/opensslconf.h>
1560 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1561 [Richard Levitte, Ulf and Bodo Möller]
1563 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1567 *) Change the 'other' type in certificate aux info to a STACK_OF
1568 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1569 the required ASN1 format: arbitrary types determined by an OID.
1572 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1573 argument to 'req'. This is not because the function is newer or
1574 better than others it just uses the work 'NEW' in the certificate
1575 request header lines. Some software needs this.
1578 *) Reorganise password command line arguments: now passwords can be
1579 obtained from various sources. Delete the PEM_cb function and make
1580 it the default behaviour: i.e. if the callback is NULL and the
1581 usrdata argument is not NULL interpret it as a null terminated pass
1582 phrase. If usrdata and the callback are NULL then the pass phrase
1583 is prompted for as usual.
1586 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1587 the support is automatically enabled. The resulting binaries will
1588 autodetect the card and use it if present.
1589 [Ben Laurie and Compaq Inc.]
1591 *) Work around for Netscape hang bug. This sends certificate request
1592 and server done in one record. Since this is perfectly legal in the
1593 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1594 the bugs/SSLv3 entry for more info.
1597 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1600 *) Add -rand argument to smime and pkcs12 applications and read/write
1604 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1607 *) Add command line password options to the remaining applications.
1610 *) Bug fix for BN_div_recp() for numerators with an even number of
1614 *) More tests in bntest.c, and changed test_bn output.
1617 *) ./config recognizes MacOS X now.
1620 *) Bug fix for BN_div() when the first words of num and divsor are
1621 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1624 *) Add support for various broken PKCS#8 formats, and command line
1625 options to produce them.
1628 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1629 get temporary BIGNUMs from a BN_CTX.
1632 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1636 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1637 include a #define from the old name to the new. The original intent
1638 was that statically linked binaries could for example just call
1639 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1640 link with digests. This never worked becayse SSLeay_add_all_digests()
1641 and SSLeay_add_all_ciphers() were in the same source file so calling
1642 one would link with the other. They are now in separate source files.
1645 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1648 *) Use a less unusual form of the Miller-Rabin primality test (it used
1649 a binary algorithm for exponentiation integrated into the Miller-Rabin
1650 loop, our standard modexp algorithms are faster).
1653 *) Support for the EBCDIC character set completed.
1654 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1656 *) Source code cleanups: use const where appropriate, eliminate casts,
1657 use void * instead of char * in lhash.
1660 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1661 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1662 this the server could overwrite ephemeral keys that the client
1666 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1667 using 50 iterations of the Rabin-Miller test.
1669 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1670 iterations of the Rabin-Miller test as required by the appendix
1671 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1672 As BN_is_prime_fasttest includes trial division, DSA parameter
1673 generation becomes much faster.
1675 This implies a change for the callback functions in DSA_is_prime
1676 and DSA_generate_parameters: The callback function is called once
1677 for each positive witness in the Rabin-Miller test, not just
1678 occasionally in the inner loop; and the parameters to the
1679 callback function now provide an iteration count for the outer
1680 loop rather than for the current invocation of the inner loop.
1681 DSA_generate_parameters additionally can call the callback
1682 function with an 'iteration count' of -1, meaning that a
1683 candidate has passed the trial division test (when q is generated
1684 from an application-provided seed, trial division is skipped).
1687 *) New function BN_is_prime_fasttest that optionally does trial
1688 division before starting the Rabin-Miller test and has
1689 an additional BN_CTX * argument (whereas BN_is_prime always
1690 has to allocate at least one BN_CTX).
1691 'callback(1, -1, cb_arg)' is called when a number has passed the
1692 trial division stage.
1695 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1699 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1702 *) New function BN_pseudo_rand().
1705 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1706 bignum version of BN_from_montgomery() with the working code from
1707 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1711 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1712 made it impossible to use the same SSL_SESSION data structure in
1713 SSL2 clients in multiple threads.
1716 *) The return value of RAND_load_file() no longer counts bytes obtained
1717 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1718 to seed the PRNG (previously an explicit byte count was required).
1719 [Ulf Möller, Bodo Möller]
1721 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1722 used (char *) instead of (void *) and had casts all over the place.
1725 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1728 *) Retain source code compatibility for BN_prime_checks macro:
1729 BN_is_prime(..., BN_prime_checks, ...) now uses
1730 BN_prime_checks_for_size to determine the appropriate number of
1731 Rabin-Miller iterations.
1734 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1735 DH_CHECK_P_NOT_SAFE_PRIME.
1736 (Check if this is true? OpenPGP calls them "strong".)
1739 *) Merge the functionality of "dh" and "gendh" programs into a new program
1740 "dhparam". The old programs are retained for now but will handle DH keys
1741 (instead of parameters) in future.
1744 *) Make the ciphers, s_server and s_client programs check the return values
1745 when a new cipher list is set.
1748 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1749 ciphers. Before when the 56bit ciphers were enabled the sorting was
1752 The syntax for the cipher sorting has been extended to support sorting by
1753 cipher-strength (using the strength_bits hard coded in the tables).
1754 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1756 Fix a bug in the cipher-command parser: when supplying a cipher command
1757 string with an "undefined" symbol (neither command nor alphanumeric
1758 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1759 an error is flagged.
1761 Due to the strength-sorting extension, the code of the
1762 ssl_create_cipher_list() function was completely rearranged. I hope that
1763 the readability was also increased :-)
1764 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1766 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1767 for the first serial number and places 2 in the serial number file. This
1768 avoids problems when the root CA is created with serial number zero and
1769 the first user certificate has the same issuer name and serial number
1773 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1774 the new code. Add documentation for this stuff.
1777 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1778 X509_*() to X509at_*() on the grounds that they don't handle X509
1779 structures and behave in an analagous way to the X509v3 functions:
1780 they shouldn't be called directly but wrapper functions should be used
1783 So we also now have some wrapper functions that call the X509at functions
1784 when passed certificate requests. (TO DO: similar things can be done with
1785 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1786 things. Some of these need some d2i or i2d and print functionality
1787 because they handle more complex structures.)
1790 *) Add missing #ifndefs that caused missing symbols when building libssl
1791 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1792 NO_RSA in ssl/s2*.c.
1793 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1795 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1796 has a return value which indicates the quality of the random data
1797 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1798 error queue. New function RAND_pseudo_bytes() generates output that is
1799 guaranteed to be unique but not unpredictable. RAND_add is like
1800 RAND_seed, but takes an extra argument for an entropy estimate
1801 (RAND_seed always assumes full entropy).
1804 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1805 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1806 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1807 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1808 false-positive rate of at most 2^-80 for random input.
1811 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1814 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1815 in the 0.9.5 release), this returns the chain
1816 from an X509_CTX structure with a dup of the stack and all
1817 the X509 reference counts upped: so the stack will exist
1818 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1821 Also make SSL_SESSION_print() print out the verify return
1825 *) Add manpage for the pkcs12 command. Also change the default
1826 behaviour so MAC iteration counts are used unless the new
1827 -nomaciter option is used. This improves file security and
1828 only older versions of MSIE (4.0 for example) need it.
1831 *) Honor the no-xxx Configure options when creating .DEF files.
1834 *) Add PKCS#10 attributes to field table: challengePassword,
1835 unstructuredName and unstructuredAddress. These are taken from
1836 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1837 international characters are used.
1839 More changes to X509_ATTRIBUTE code: allow the setting of types
1840 based on strings. Remove the 'loc' parameter when adding
1841 attributes because these will be a SET OF encoding which is sorted
1845 *) Initial changes to the 'req' utility to allow request generation
1846 automation. This will allow an application to just generate a template
1847 file containing all the field values and have req construct the
1850 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1851 used all over the place including certificate requests and PKCS#7
1852 structures. They are currently handled manually where necessary with
1853 some primitive wrappers for PKCS#7. The new functions behave in a
1854 manner analogous to the X509 extension functions: they allow
1855 attributes to be looked up by NID and added.
1857 Later something similar to the X509V3 code would be desirable to
1858 automatically handle the encoding, decoding and printing of the
1859 more complex types. The string types like challengePassword can
1860 be handled by the string table functions.
1862 Also modified the multi byte string table handling. Now there is
1863 a 'global mask' which masks out certain types. The table itself
1864 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1865 is useful when for example there is only one permissible type
1866 (as in countryName) and using the mask might result in no valid
1870 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1871 SSL_get_peer_finished to allow applications to obtain the latest
1872 Finished messages sent to the peer or expected from the peer,
1873 respectively. (SSL_get_peer_finished is usually the Finished message
1874 actually received from the peer, otherwise the protocol will be aborted.)
1876 As the Finished message are message digests of the complete handshake
1877 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1878 be used for external authentication procedures when the authentication
1879 provided by SSL/TLS is not desired or is not enough.
1882 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1883 the host supports BWX extension and if Compaq C is present on the
1884 $PATH. Just exploiting of the BWX extension results in 20-30%
1885 performance kick for some algorithms, e.g. DES and RC4 to mention
1886 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1890 *) Add support for MS "fast SGC". This is arguably a violation of the
1891 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1892 weak crypto and after checking the certificate is SGC a second one
1893 with strong crypto. MS SGC stops the first handshake after receiving
1894 the server certificate message and sends a second client hello. Since
1895 a server will typically do all the time consuming operations before
1896 expecting any further messages from the client (server key exchange
1897 is the most expensive) there is little difference between the two.
1899 To get OpenSSL to support MS SGC we have to permit a second client
1900 hello message after we have sent server done. In addition we have to
1901 reset the MAC if we do get this second client hello.
1904 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1905 if a DER encoded private key is RSA or DSA traditional format. Changed
1906 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1907 format DER encoded private key. Newer code should use PKCS#8 format which
1908 has the key type encoded in the ASN1 structure. Added DER private key
1909 support to pkcs8 application.
1912 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1913 ciphersuites has been selected (as required by the SSL 3/TLS 1
1914 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1915 is set, we interpret this as a request to violate the specification
1916 (the worst that can happen is a handshake failure, and 'correct'
1917 behaviour would result in a handshake failure anyway).
1920 *) In SSL_CTX_add_session, take into account that there might be multiple
1921 SSL_SESSION structures with the same session ID (e.g. when two threads
1922 concurrently obtain them from an external cache).
1923 The internal cache can handle only one SSL_SESSION with a given ID,
1924 so if there's a conflict, we now throw out the old one to achieve
1928 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1929 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1930 some routines that use cipher OIDs: some ciphers do not have OIDs
1931 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1935 *) Simplify the trust setting structure and code. Now we just have
1936 two sequences of OIDs for trusted and rejected settings. These will
1937 typically have values the same as the extended key usage extension
1938 and any application specific purposes.
1940 The trust checking code now has a default behaviour: it will just
1941 check for an object with the same NID as the passed id. Functions can
1942 be provided to override either the default behaviour or the behaviour
1943 for a given id. SSL client, server and email already have functions
1944 in place for compatibility: they check the NID and also return "trusted"
1945 if the certificate is self signed.
1948 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1949 traditional format into an EVP_PKEY structure.
1952 *) Add a password callback function PEM_cb() which either prompts for
1953 a password if usr_data is NULL or otherwise assumes it is a null
1954 terminated password. Allow passwords to be passed on command line
1955 environment or config files in a few more utilities.
1958 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1959 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1960 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1961 Update documentation.
1964 *) Support for ASN1 "NULL" type. This could be handled before by using
1965 ASN1_TYPE but there wasn't any function that would try to read a NULL
1966 and produce an error if it couldn't. For compatibility we also have
1967 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1968 don't allocate anything because they don't need to.
1971 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1973 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1975 *) Rebuild of the memory allocation routines used by OpenSSL code and
1976 possibly others as well. The purpose is to make an interface that
1977 provide hooks so anyone can build a separate set of allocation and
1978 deallocation routines to be used by OpenSSL, for example memory
1979 pool implementations, or something else, which was previously hard
1980 since Malloc(), Realloc() and Free() were defined as macros having
1981 the values malloc, realloc and free, respectively (except for Win32
1982 compilations). The same is provided for memory debugging code.
1983 OpenSSL already comes with functionality to find memory leaks, but
1984 this gives people a chance to debug other memory problems.
1986 With these changes, a new set of functions and macros have appeared:
1988 CRYPTO_set_mem_debug_functions() [F]
1989 CRYPTO_get_mem_debug_functions() [F]
1990 CRYPTO_dbg_set_options() [F]
1991 CRYPTO_dbg_get_options() [F]
1992 CRYPTO_malloc_debug_init() [M]
1994 The memory debug functions are NULL by default, unless the library
1995 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1996 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1997 gives the standard debugging functions that come with OpenSSL) or
1998 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1999 provided by the library user) must be used. When the standard
2000 debugging functions are used, CRYPTO_dbg_set_options can be used to
2001 request additional information:
2002 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2003 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2005 Also, things like CRYPTO_set_mem_functions will always give the
2006 expected result (the new set of functions is used for allocation
2007 and deallocation) at all times, regardless of platform and compiler
2010 To finish it up, some functions that were never use in any other
2011 way than through macros have a new API and new semantic:
2014 CRYPTO_dbg_realloc()
2017 All macros of value have retained their old syntax.
2018 [Richard Levitte and Bodo Moeller]
2020 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2021 ordering of SMIMECapabilities wasn't in "strength order" and there
2022 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2026 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2027 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2028 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2030 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2031 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2032 functionality to handle multipart/signed properly) and a utility
2033 called 'smime' to call all this stuff. This is based on code I
2034 originally wrote for Celo who have kindly allowed it to be
2035 included in OpenSSL.
2038 *) Add variants des_set_key_checked and des_set_key_unchecked of
2039 des_set_key (aka des_key_sched). Global variable des_check_key
2040 decides which of these is called by des_set_key; this way
2041 des_check_key behaves as it always did, but applications and
2042 the library itself, which was buggy for des_check_key == 1,
2043 have a cleaner way to pick the version they need.
2046 *) New function PKCS12_newpass() which changes the password of a
2050 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2051 dynamic mix. In both cases the ids can be used as an index into the
2052 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2053 functions so they accept a list of the field values and the
2054 application doesn't need to directly manipulate the X509_TRUST
2058 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2062 *) Modify the way the V3 extension code looks up extensions. This now
2063 works in a similar way to the object code: we have some "standard"
2064 extensions in a static table which is searched with OBJ_bsearch()
2065 and the application can add dynamic ones if needed. The file
2066 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2067 updated whenever a new extension is added to the core code and kept
2068 in ext_nid order. There is a simple program 'tabtest.c' which checks
2069 this. New extensions are not added too often so this file can readily
2070 be maintained manually.
2072 There are two big advantages in doing things this way. The extensions
2073 can be looked up immediately and no longer need to be "added" using
2074 X509V3_add_standard_extensions(): this function now does nothing.
2075 [Side note: I get *lots* of email saying the extension code doesn't
2076 work because people forget to call this function]
2077 Also no dynamic allocation is done unless new extensions are added:
2078 so if we don't add custom extensions there is no need to call
2079 X509V3_EXT_cleanup().
2082 *) Modify enc utility's salting as follows: make salting the default. Add a
2083 magic header, so unsalted files fail gracefully instead of just decrypting
2084 to garbage. This is because not salting is a big security hole, so people
2085 should be discouraged from doing it.
2088 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2089 digest to be passed on the command line but it only used this
2090 parameter when signing a certificate. Modified so all relevant
2091 operations are affected by the digest parameter including the
2092 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2093 DSA key was used because it didn't fix the digest.
2096 *) Initial certificate chain verify code. Currently tests the untrusted
2097 certificates for consistency with the verify purpose (which is set
2098 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2100 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2101 this is because it will reject chains with invalid extensions whereas
2102 every previous version of OpenSSL and SSLeay made no checks at all.
2104 Trust code: checks the root CA for the relevant trust settings. Trust
2105 settings have an initial value consistent with the verify purpose: e.g.
2106 if the verify purpose is for SSL client use it expects the CA to be
2107 trusted for SSL client use. However the default value can be changed to
2108 permit custom trust settings: one example of this would be to only trust
2109 certificates from a specific "secure" set of CAs.
2111 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2112 which should be used for version portability: especially since the
2113 verify structure is likely to change more often now.
2115 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2116 to set them. If not set then assume SSL clients will verify SSL servers
2119 Two new options to the verify program: -untrusted allows a set of
2120 untrusted certificates to be passed in and -purpose which sets the
2121 intended purpose of the certificate. If a purpose is set then the
2122 new chain verify code is used to check extension consistency.
2125 *) Support for the authority information access extension.
2128 *) Modify RSA and DSA PEM read routines to transparently handle
2129 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2130 public keys in a format compatible with certificate
2131 SubjectPublicKeyInfo structures. Unfortunately there were already
2132 functions called *_PublicKey_* which used various odd formats so
2133 these are retained for compatibility: however the DSA variants were
2134 never in a public release so they have been deleted. Changed dsa/rsa
2135 utilities to handle the new format: note no releases ever handled public
2136 keys so we should be OK.
2138 The primary motivation for this change is to avoid the same fiasco
2139 that dogs private keys: there are several incompatible private key
2140 formats some of which are standard and some OpenSSL specific and
2141 require various evil hacks to allow partial transparent handling and
2142 even then it doesn't work with DER formats. Given the option anything
2143 other than PKCS#8 should be dumped: but the other formats have to
2144 stay in the name of compatibility.
2146 With public keys and the benefit of hindsight one standard format
2147 is used which works with EVP_PKEY, RSA or DSA structures: though
2148 it clearly returns an error if you try to read the wrong kind of key.
2150 Added a -pubkey option to the 'x509' utility to output the public key.
2151 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2152 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2153 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2154 that do the same as the EVP_PKEY_assign_*() except they up the
2155 reference count of the added key (they don't "swallow" the
2159 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2160 CRLs would fail if the file contained no certificates or no CRLs:
2161 added a new function to read in both types and return the number
2162 read: this means that if none are read it will be an error. The
2163 DER versions of the certificate and CRL reader would always fail
2164 because it isn't possible to mix certificates and CRLs in DER format
2165 without choking one or the other routine. Changed this to just read
2166 a certificate: this is the best we can do. Also modified the code
2167 in apps/verify.c to take notice of return codes: it was previously
2168 attempting to read in certificates from NULL pointers and ignoring
2169 any errors: this is one reason why the cert and CRL reader seemed
2170 to work. It doesn't check return codes from the default certificate
2171 routines: these may well fail if the certificates aren't installed.
2174 *) Code to support otherName option in GeneralName.
2177 *) First update to verify code. Change the verify utility
2178 so it warns if it is passed a self signed certificate:
2179 for consistency with the normal behaviour. X509_verify
2180 has been modified to it will now verify a self signed
2181 certificate if *exactly* the same certificate appears
2182 in the store: it was previously impossible to trust a
2183 single self signed certificate. This means that:
2184 openssl verify ss.pem
2185 now gives a warning about a self signed certificate but
2186 openssl verify -CAfile ss.pem ss.pem
2190 *) For servers, store verify_result in SSL_SESSION data structure
2191 (and add it to external session representation).
2192 This is needed when client certificate verifications fails,
2193 but an application-provided verification callback (set by
2194 SSL_CTX_set_cert_verify_callback) allows accepting the session
2195 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2196 but returns 1): When the session is reused, we have to set
2197 ssl->verify_result to the appropriate error code to avoid
2199 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2201 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2202 case in PKCS7_dataInit() where the signed PKCS7 structure
2203 didn't contain any existing data because it was being created.
2204 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2206 *) Add a salt to the key derivation routines in enc.c. This
2207 forms the first 8 bytes of the encrypted file. Also add a
2208 -S option to allow a salt to be input on the command line.
2211 *) New function X509_cmp(). Oddly enough there wasn't a function
2212 to compare two certificates. We do this by working out the SHA1
2213 hash and comparing that. X509_cmp() will be needed by the trust
2217 *) SSL_get1_session() is like SSL_get_session(), but increments
2218 the reference count in the SSL_SESSION returned.
2219 [Geoff Thorpe <geoff@eu.c2.net>]
2221 *) Fix for 'req': it was adding a null to request attributes.
2222 Also change the X509_LOOKUP and X509_INFO code to handle
2223 certificate auxiliary information.
2226 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2230 *) Add the possibility to add extra information to the memory leak
2231 detecting output, to form tracebacks, showing from where each
2232 allocation was originated: CRYPTO_push_info("constant string") adds
2233 the string plus current file name and line number to a per-thread
2234 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2235 is like calling CYRPTO_pop_info() until the stack is empty.
2236 Also updated memory leak detection code to be multi-thread-safe.
2239 *) Add options -text and -noout to pkcs7 utility and delete the
2240 encryption options which never did anything. Update docs.
2243 *) Add options to some of the utilities to allow the pass phrase
2244 to be included on either the command line (not recommended on
2245 OSes like Unix) or read from the environment. Update the
2246 manpages and fix a few bugs.
2249 *) Add a few manpages for some of the openssl commands.
2252 *) Fix the -revoke option in ca. It was freeing up memory twice,
2253 leaking and not finding already revoked certificates.
2256 *) Extensive changes to support certificate auxiliary information.
2257 This involves the use of X509_CERT_AUX structure and X509_AUX
2258 functions. An X509_AUX function such as PEM_read_X509_AUX()
2259 can still read in a certificate file in the usual way but it
2260 will also read in any additional "auxiliary information". By
2261 doing things this way a fair degree of compatibility can be
2262 retained: existing certificates can have this information added
2263 using the new 'x509' options.
2265 Current auxiliary information includes an "alias" and some trust
2266 settings. The trust settings will ultimately be used in enhanced
2267 certificate chain verification routines: currently a certificate
2268 can only be trusted if it is self signed and then it is trusted
2272 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2273 The problem was that one of the replacement routines had not been working
2274 since SSLeay releases. For now the offending routine has been replaced
2275 with non-optimised assembler. Even so, this now gives around 95%
2276 performance improvement for 1024 bit RSA signs.
2279 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2280 handling. Most clients have the effective key size in bits equal to
2281 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2282 A few however don't do this and instead use the size of the decrypted key
2283 to determine the RC2 key length and the AlgorithmIdentifier to determine
2284 the effective key length. In this case the effective key length can still
2285 be 40 bits but the key length can be 168 bits for example. This is fixed
2286 by manually forcing an RC2 key into the EVP_PKEY structure because the
2287 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2288 the key length and effective key length are equal.
2291 *) Add a bunch of functions that should simplify the creation of
2292 X509_NAME structures. Now you should be able to do:
2293 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2294 and have it automatically work out the correct field type and fill in
2295 the structures. The more adventurous can try:
2296 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2297 and it will (hopefully) work out the correct multibyte encoding.
2300 *) Change the 'req' utility to use the new field handling and multibyte
2301 copy routines. Before the DN field creation was handled in an ad hoc
2302 way in req, ca, and x509 which was rather broken and didn't support
2303 BMPStrings or UTF8Strings. Since some software doesn't implement
2304 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2305 using the dirstring_type option. See the new comment in the default
2306 openssl.cnf for more info.
2309 *) Make crypto/rand/md_rand.c more robust:
2310 - Assure unique random numbers after fork().
2311 - Make sure that concurrent threads access the global counter and
2312 md serializably so that we never lose entropy in them
2313 or use exactly the same state in multiple threads.
2314 Access to the large state is not always serializable because
2315 the additional locking could be a performance killer, and
2316 md should be large enough anyway.
2319 *) New file apps/app_rand.c with commonly needed functionality
2320 for handling the random seed file.
2322 Use the random seed file in some applications that previously did not:
2324 dsaparam -genkey (which also ignored its '-rand' option),
2327 x509 (when signing).
2328 Except on systems with /dev/urandom, it is crucial to have a random
2329 seed file at least for key creation, DSA signing, and for DH exchanges;
2330 for RSA signatures we could do without one.
2332 gendh and gendsa (unlike genrsa) used to read only the first byte
2333 of each file listed in the '-rand' option. The function as previously
2334 found in genrsa is now in app_rand.c and is used by all programs
2335 that support '-rand'.
2338 *) In RAND_write_file, use mode 0600 for creating files;
2339 don't just chmod when it may be too late.
2342 *) Report an error from X509_STORE_load_locations
2343 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2346 *) New function ASN1_mbstring_copy() this copies a string in either
2347 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2348 into an ASN1_STRING type. A mask of permissible types is passed
2349 and it chooses the "minimal" type to use or an error if not type
2353 *) Add function equivalents to the various macros in asn1.h. The old
2354 macros are retained with an M_ prefix. Code inside the library can
2355 use the M_ macros. External code (including the openssl utility)
2356 should *NOT* in order to be "shared library friendly".
2359 *) Add various functions that can check a certificate's extensions
2360 to see if it usable for various purposes such as SSL client,
2361 server or S/MIME and CAs of these types. This is currently
2362 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2363 verification. Also added a -purpose flag to x509 utility to
2364 print out all the purposes.
2367 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2371 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2372 for, obtain and decode and extension and obtain its critical flag.
2373 This allows all the necessary extension code to be handled in a
2374 single function call.
2377 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2378 platforms. See crypto/rc4/rc4_enc.c for further details.
2381 *) New -noout option to asn1parse. This causes no output to be produced
2382 its main use is when combined with -strparse and -out to extract data
2383 from a file (which may not be in ASN.1 format).
2386 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2387 when producing the local key id.
2388 [Richard Levitte <levitte@stacken.kth.se>]
2390 *) New option -dhparam in s_server. This allows a DH parameter file to be
2391 stated explicitly. If it is not stated then it tries the first server
2392 certificate file. The previous behaviour hard coded the filename
2396 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2397 a public key to be input or output. For example:
2398 openssl rsa -in key.pem -pubout -out pubkey.pem
2399 Also added necessary DSA public key functions to handle this.
2402 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2403 in the message. This was handled by allowing
2404 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2405 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2407 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2408 to the end of the strings whereas this didn't. This would cause problems
2409 if strings read with d2i_ASN1_bytes() were later modified.
2410 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2412 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2413 data and it contains EOF it will end up returning an error. This is
2414 caused by input 46 bytes long. The cause is due to the way base64
2415 BIOs find the start of base64 encoded data. They do this by trying a
2416 trial decode on each line until they find one that works. When they
2417 do a flag is set and it starts again knowing it can pass all the
2418 data directly through the decoder. Unfortunately it doesn't reset
2419 the context it uses. This means that if EOF is reached an attempt
2420 is made to pass two EOFs through the context and this causes the
2421 resulting error. This can also cause other problems as well. As is
2422 usual with these problems it takes *ages* to find and the fix is
2423 trivial: move one line.
2424 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2426 *) Ugly workaround to get s_client and s_server working under Windows. The
2427 old code wouldn't work because it needed to select() on sockets and the
2428 tty (for keypresses and to see if data could be written). Win32 only
2429 supports select() on sockets so we select() with a 1s timeout on the
2430 sockets and then see if any characters are waiting to be read, if none
2431 are present then we retry, we also assume we can always write data to
2432 the tty. This isn't nice because the code then blocks until we've
2433 received a complete line of data and it is effectively polling the
2434 keyboard at 1s intervals: however it's quite a bit better than not
2435 working at all :-) A dedicated Windows application might handle this
2436 with an event loop for example.
2439 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2440 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2441 will be called when RSA_sign() and RSA_verify() are used. This is useful
2442 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2443 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2444 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2445 This necessitated the support of an extra signature type NID_md5_sha1
2446 for SSL signatures and modifications to the SSL library to use it instead
2447 of calling RSA_public_decrypt() and RSA_private_encrypt().
2450 *) Add new -verify -CAfile and -CApath options to the crl program, these
2451 will lookup a CRL issuers certificate and verify the signature in a
2452 similar way to the verify program. Tidy up the crl program so it
2453 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2454 less strict. It will now permit CRL extensions even if it is not
2455 a V2 CRL: this will allow it to tolerate some broken CRLs.
2458 *) Initialize all non-automatic variables each time one of the openssl
2459 sub-programs is started (this is necessary as they may be started
2460 multiple times from the "OpenSSL>" prompt).
2461 [Lennart Bang, Bodo Moeller]
2463 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2464 removing all other RSA functionality (this is what NO_RSA does). This
2465 is so (for example) those in the US can disable those operations covered
2466 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2470 *) Non-copying interface to BIO pairs.
2471 (still largely untested)
2474 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2475 ASCII string. This was handled independently in various places before.
2478 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2479 UTF8 strings a character at a time.
2482 *) Use client_version from client hello to select the protocol
2483 (s23_srvr.c) and for RSA client key exchange verification
2484 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2487 *) Add various utility functions to handle SPKACs, these were previously
2488 handled by poking round in the structure internals. Added new function
2489 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2490 print, verify and generate SPKACs. Based on an original idea from
2491 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2494 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2497 *) Allow the config file extension section to be overwritten on the
2498 command line. Based on an original idea from Massimiliano Pala
2499 <madwolf@comune.modena.it>. The new option is called -extensions
2500 and can be applied to ca, req and x509. Also -reqexts to override
2501 the request extensions in req and -crlexts to override the crl extensions
2505 *) Add new feature to the SPKAC handling in ca. Now you can include
2506 the same field multiple times by preceding it by "XXXX." for example:
2509 this is the same syntax as used in the req config file.
2512 *) Allow certificate extensions to be added to certificate requests. These
2513 are specified in a 'req_extensions' option of the req section of the
2514 config file. They can be printed out with the -text option to req but
2515 are otherwise ignored at present.
2518 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2519 data read consists of only the final block it would not decrypted because
2520 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2521 A misplaced 'break' also meant the decrypted final block might not be
2522 copied until the next read.
2525 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2526 a few extra parameters to the DH structure: these will be useful if
2527 for example we want the value of 'q' or implement X9.42 DH.
2530 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2531 provides hooks that allow the default DSA functions or functions on a
2532 "per key" basis to be replaced. This allows hardware acceleration and
2533 hardware key storage to be handled without major modification to the
2534 library. Also added low level modexp hooks and CRYPTO_EX structure and
2535 associated functions.
2538 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2539 as "read only": it can't be written to and the buffer it points to will
2540 not be freed. Reading from a read only BIO is much more efficient than
2541 a normal memory BIO. This was added because there are several times when
2542 an area of memory needs to be read from a BIO. The previous method was
2543 to create a memory BIO and write the data to it, this results in two
2544 copies of the data and an O(n^2) reading algorithm. There is a new
2545 function BIO_new_mem_buf() which creates a read only memory BIO from
2546 an area of memory. Also modified the PKCS#7 routines to use read only
2550 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2551 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2552 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2553 but a retry condition occured while trying to read the rest.
2556 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2557 NID_pkcs7_encrypted by default: this was wrong since this should almost
2558 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2559 the encrypted data type: this is a more sensible place to put it and it
2560 allows the PKCS#12 code to be tidied up that duplicated this
2564 *) Changed obj_dat.pl script so it takes its input and output files on
2565 the command line. This should avoid shell escape redirection problems
2569 *) Initial support for certificate extension requests, these are included
2570 in things like Xenroll certificate requests. Included functions to allow
2571 extensions to be obtained and added.
2574 *) -crlf option to s_client and s_server for sending newlines as
2575 CRLF (as required by many protocols).
2578 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2580 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2581 [Ralf S. Engelschall]
2583 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2584 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2586 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2590 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2591 DH parameters/keys (q is lost during that conversion, but the resulting
2592 DH parameters contain its length).
2594 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2595 much faster than DH_generate_parameters (which creates parameters
2596 where p = 2*q + 1), and also the smaller q makes DH computations
2597 much more efficient (160-bit exponentiation instead of 1024-bit
2598 exponentiation); so this provides a convenient way to support DHE
2599 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2600 utter importance to use
2601 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2603 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2604 when such DH parameters are used, because otherwise small subgroup
2605 attacks may become possible!
2608 *) Avoid memory leak in i2d_DHparams.
2611 *) Allow the -k option to be used more than once in the enc program:
2612 this allows the same encrypted message to be read by multiple recipients.
2615 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2616 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2617 it will always use the numerical form of the OID, even if it has a short
2621 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2622 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2623 otherwise bn_mod_exp was called. In the case of hardware keys for example
2624 no private key components need be present and it might store extra data
2625 in the RSA structure, which cannot be accessed from bn_mod_exp.
2626 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2627 private key operations.
2630 *) Added support for SPARC Linux.
2633 *) pem_password_cb function type incompatibly changed from
2634 typedef int pem_password_cb(char *buf, int size, int rwflag);
2636 ....(char *buf, int size, int rwflag, void *userdata);
2637 so that applications can pass data to their callbacks:
2638 The PEM[_ASN1]_{read,write}... functions and macros now take an
2639 additional void * argument, which is just handed through whenever
2640 the password callback is called.
2641 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2643 New function SSL_CTX_set_default_passwd_cb_userdata.
2645 Compatibility note: As many C implementations push function arguments
2646 onto the stack in reverse order, the new library version is likely to
2647 interoperate with programs that have been compiled with the old
2648 pem_password_cb definition (PEM_whatever takes some data that
2649 happens to be on the stack as its last argument, and the callback
2650 just ignores this garbage); but there is no guarantee whatsoever that
2653 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2654 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2655 problems not only on Windows, but also on some Unix platforms.
2656 To avoid problematic command lines, these definitions are now in an
2657 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2658 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2661 *) MIPS III/IV assembler module is reimplemented.
2664 *) More DES library cleanups: remove references to srand/rand and
2665 delete an unused file.
2668 *) Add support for the the free Netwide assembler (NASM) under Win32,
2669 since not many people have MASM (ml) and it can be hard to obtain.
2670 This is currently experimental but it seems to work OK and pass all
2671 the tests. Check out INSTALL.W32 for info.
2674 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2675 without temporary keys kept an extra copy of the server key,
2676 and connections with temporary keys did not free everything in case
2680 *) New function RSA_check_key and new openssl rsa option -check
2681 for verifying the consistency of RSA keys.
2682 [Ulf Moeller, Bodo Moeller]
2684 *) Various changes to make Win32 compile work:
2685 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2686 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2687 comparison" warnings.
2688 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2691 *) Add a debugging option to PKCS#5 v2 key generation function: when
2692 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2693 derived keys are printed to stderr.
2696 *) Copy the flags in ASN1_STRING_dup().
2697 [Roman E. Pavlov <pre@mo.msk.ru>]
2699 *) The x509 application mishandled signing requests containing DSA
2700 keys when the signing key was also DSA and the parameters didn't match.
2702 It was supposed to omit the parameters when they matched the signing key:
2703 the verifying software was then supposed to automatically use the CA's
2704 parameters if they were absent from the end user certificate.
2706 Omitting parameters is no longer recommended. The test was also
2707 the wrong way round! This was probably due to unusual behaviour in
2708 EVP_cmp_parameters() which returns 1 if the parameters match.
2709 This meant that parameters were omitted when they *didn't* match and
2710 the certificate was useless. Certificates signed with 'ca' didn't have
2712 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2714 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2715 The interface is as follows:
2716 Applications can use
2717 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2718 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2719 "off" is now the default.
2720 The library internally uses
2721 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2722 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2723 to disable memory-checking temporarily.
2725 Some inconsistent states that previously were possible (and were
2726 even the default) are now avoided.
2728 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2729 with each memory chunk allocated; this is occasionally more helpful
2730 than just having a counter.
2732 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2734 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2738 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2739 which largely parallels "options", but is for changing API behaviour,
2740 whereas "options" are about protocol behaviour.
2741 Initial "mode" flags are:
2743 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2744 a single record has been written.
2745 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2746 retries use the same buffer location.
2747 (But all of the contents must be
2751 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2754 *) Fix problems with no-hmac etc.
2755 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2757 *) New functions RSA_get_default_method(), RSA_set_method() and
2758 RSA_get_method(). These allows replacement of RSA_METHODs without having
2759 to mess around with the internals of an RSA structure.
2762 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2763 Also really enable memory leak checks in openssl.c and in some
2765 [Chad C. Mulligan, Bodo Moeller]
2767 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2768 up the length of negative integers. This has now been simplified to just
2769 store the length when it is first determined and use it later, rather
2770 than trying to keep track of where data is copied and updating it to
2772 [Steve Henson, reported by Brien Wheeler
2773 <bwheeler@authentica-security.com>]
2775 *) Add a new function PKCS7_signatureVerify. This allows the verification
2776 of a PKCS#7 signature but with the signing certificate passed to the
2777 function itself. This contrasts with PKCS7_dataVerify which assumes the
2778 certificate is present in the PKCS#7 structure. This isn't always the
2779 case: certificates can be omitted from a PKCS#7 structure and be
2780 distributed by "out of band" means (such as a certificate database).
2783 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2784 function prototypes in pem.h, also change util/mkdef.pl to add the
2785 necessary function names.
2788 *) mk1mf.pl (used by Windows builds) did not properly read the
2789 options set by Configure in the top level Makefile, and Configure
2790 was not even able to write more than one option correctly.
2791 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2794 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2795 file to be loaded from a BIO or FILE pointer. The BIO version will
2796 for example allow memory BIOs to contain config info.
2799 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2800 Whoever hopes to achieve shared-library compatibility across versions
2801 must use this, not the compile-time macro.
2802 (Exercise 0.9.4: Which is the minimum library version required by
2804 Note: All this applies only to multi-threaded programs, others don't
2808 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2809 through a BIO pair triggered the default case, i.e.
2810 SSLerr(...,SSL_R_UNKNOWN_STATE).
2813 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2814 can use the SSL library even if none of the specific BIOs is
2818 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2819 for the encoded length.
2820 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2822 *) Add initial documentation of the X509V3 functions.
2825 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2826 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2827 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2828 secure PKCS#8 private key format with a high iteration count.
2831 *) Fix determination of Perl interpreter: A perl or perl5
2832 _directory_ in $PATH was also accepted as the interpreter.
2833 [Ralf S. Engelschall]
2835 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2836 wrong with it but it was very old and did things like calling
2837 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2841 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2842 to use the new extension code.
2845 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2846 with macros. This should make it easier to change their form, add extra
2847 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2851 *) Add to configuration table a new entry that can specify an alternative
2852 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2853 according to Mark Crispin <MRC@Panda.COM>.
2857 *) DES CBC did not update the IV. Weird.
2860 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2861 Changing the behaviour of the former might break existing programs --
2862 where IV updating is needed, des_ncbc_encrypt can be used.
2865 *) When bntest is run from "make test" it drives bc to check its
2866 calculations, as well as internally checking them. If an internal check
2867 fails, it needs to cause bc to give a non-zero result or make test carries
2868 on without noticing the failure. Fixed.
2871 *) DES library cleanups.
2874 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2875 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2876 ciphers. NOTE: although the key derivation function has been verified
2877 against some published test vectors it has not been extensively tested
2878 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2882 *) Instead of "mkdir -p", which is not fully portable, use new
2883 Perl script "util/mkdir-p.pl".
2886 *) Rewrite the way password based encryption (PBE) is handled. It used to
2887 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2888 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2889 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2890 the 'parameter' field of the AlgorithmIdentifier is passed to the
2891 underlying key generation function so it must do its own ASN1 parsing.
2892 This has also changed the EVP_PBE_CipherInit() function which now has a
2893 'parameter' argument instead of literal salt and iteration count values
2894 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2897 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2898 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2899 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2900 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2901 value was just used as a "magic string" and not used directly its
2902 value doesn't matter.
2905 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2909 *) "linux-sparc64" configuration (ultrapenguin).
2910 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2911 "linux-sparc" configuration.
2912 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2914 *) config now generates no-xxx options for missing ciphers.
2917 *) Support the EBCDIC character set (work in progress).
2918 File ebcdic.c not yet included because it has a different license.
2919 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2921 *) Support BS2000/OSD-POSIX.
2922 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2924 *) Make callbacks for key generation use void * instead of char *.
2927 *) Make S/MIME samples compile (not yet tested).
2930 *) Additional typesafe stacks.
2933 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2937 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2939 *) New configuration variant "sco5-gcc".
2941 *) Updated some demos.
2942 [Sean O Riordain, Wade Scholine]
2944 *) Add missing BIO_free at exit of pkcs12 application.
2947 *) Fix memory leak in conf.c.
2950 *) Updates for Win32 to assembler version of MD5.
2953 *) Set #! path to perl in apps/der_chop to where we found it
2954 instead of using a fixed path.
2957 *) SHA library changes for irix64-mips4-cc.
2960 *) Improvements for VMS support.
2964 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2966 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2967 This also avoids the problems with SC4.2 and unpatched SC5.
2968 [Andy Polyakov <appro@fy.chalmers.se>]
2970 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2971 These are required because of the typesafe stack would otherwise break
2972 existing code. If old code used a structure member which used to be STACK
2973 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2974 sk_num or sk_value it would produce an error because the num, data members
2975 are not present in STACK_OF. Now it just produces a warning. sk_set
2976 replaces the old method of assigning a value to sk_value
2977 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2978 that does this will no longer work (and should use sk_set instead) but
2979 this could be regarded as a "questionable" behaviour anyway.
2982 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2983 correctly handle encrypted S/MIME data.
2986 *) Change type of various DES function arguments from des_cblock
2987 (which means, in function argument declarations, pointer to char)
2988 to des_cblock * (meaning pointer to array with 8 char elements),
2989 which allows the compiler to do more typechecking; it was like
2990 that back in SSLeay, but with lots of ugly casts.
2992 Introduce new type const_des_cblock.
2995 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2996 problems: find RecipientInfo structure that matches recipient certificate
2997 and initialise the ASN1 structures properly based on passed cipher.
3000 *) Belatedly make the BN tests actually check the results.
3003 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3004 to and from BNs: it was completely broken. New compilation option
3005 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3006 key elements as negative integers.
3009 *) Reorganize and speed up MD5.
3010 [Andy Polyakov <appro@fy.chalmers.se>]
3013 [Richard Levitte <richard@levitte.org>]
3015 *) New option -out to asn1parse to allow the parsed structure to be
3016 output to a file. This is most useful when combined with the -strparse
3017 option to examine the output of things like OCTET STRINGS.
3020 *) Make SSL library a little more fool-proof by not requiring any longer
3021 that SSL_set_{accept,connect}_state be called before
3022 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3023 in many applications because usually everything *appeared* to work as
3024 intended anyway -- now it really works as intended).
3027 *) Move openssl.cnf out of lib/.
3030 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3031 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3032 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3033 [Ralf S. Engelschall]
3035 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3036 handle PKCS#7 enveloped data properly.
3037 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3039 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3040 copying pointers. The cert_st handling is changed by this in
3041 various ways (and thus what used to be known as ctx->default_cert
3042 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3043 any longer when s->cert does not give us what we need).
3044 ssl_cert_instantiate becomes obsolete by this change.
3045 As soon as we've got the new code right (possibly it already is?),
3046 we have solved a couple of bugs of the earlier code where s->cert
3047 was used as if it could not have been shared with other SSL structures.
3049 Note that using the SSL API in certain dirty ways now will result
3050 in different behaviour than observed with earlier library versions:
3051 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3052 does not influence s as it used to.
3054 In order to clean up things more thoroughly, inside SSL_SESSION
3055 we don't use CERT any longer, but a new structure SESS_CERT
3056 that holds per-session data (if available); currently, this is
3057 the peer's certificate chain and, for clients, the server's certificate
3058 and temporary key. CERT holds only those values that can have
3059 meaningful defaults in an SSL_CTX.
3062 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3063 from the internal representation. Various PKCS#7 fixes: remove some
3064 evil casts and set the enc_dig_alg field properly based on the signing
3068 *) Allow PKCS#12 password to be set from the command line or the
3069 environment. Let 'ca' get its config file name from the environment
3070 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3074 *) Allow certificate policies extension to use an IA5STRING for the
3075 organization field. This is contrary to the PKIX definition but
3076 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3080 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3081 without disallowing inline assembler and the like for non-pedantic builds.
3084 *) Support Borland C++ builder.
3085 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3090 *) SHA-1 cleanups and performance enhancements.
3091 [Andy Polyakov <appro@fy.chalmers.se>]
3093 *) Sparc v8plus assembler for the bignum library.
3094 [Andy Polyakov <appro@fy.chalmers.se>]
3096 *) Accept any -xxx and +xxx compiler options in Configure.
3099 *) Update HPUX configuration.
3102 *) Add missing sk_<type>_unshift() function to safestack.h
3103 [Ralf S. Engelschall]
3105 *) New function SSL_CTX_use_certificate_chain_file that sets the
3106 "extra_cert"s in addition to the certificate. (This makes sense
3107 only for "PEM" format files, as chains as a whole are not
3111 *) Support verify_depth from the SSL API.
3112 x509_vfy.c had what can be considered an off-by-one-error:
3113 Its depth (which was not part of the external interface)
3114 was actually counting the number of certificates in a chain;
3115 now it really counts the depth.
3118 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3119 instead of X509err, which often resulted in confusing error
3120 messages since the error codes are not globally unique
3121 (e.g. an alleged error in ssl3_accept when a certificate
3122 didn't match the private key).
3124 *) New function SSL_CTX_set_session_id_context that allows to set a default
3125 value (so that you don't need SSL_set_session_id_context for each
3126 connection using the SSL_CTX).
3129 *) OAEP decoding bug fix.
3132 *) Support INSTALL_PREFIX for package builders, as proposed by
3136 *) New Configure options "threads" and "no-threads". For systems
3137 where the proper compiler options are known (currently Solaris
3138 and Linux), "threads" is the default.
3141 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3144 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3145 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3146 such as /usr/local/bin.
3149 *) "make linux-shared" to build shared libraries.
3150 [Niels Poppe <niels@netbox.org>]
3152 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3155 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3156 extension adding in x509 utility.
3159 *) Remove NOPROTO sections and error code comments.
3162 *) Partial rewrite of the DEF file generator to now parse the ANSI
3166 *) New Configure options --prefix=DIR and --openssldir=DIR.
3169 *) Complete rewrite of the error code script(s). It is all now handled
3170 by one script at the top level which handles error code gathering,
3171 header rewriting and C source file generation. It should be much better
3172 than the old method: it now uses a modified version of Ulf's parser to
3173 read the ANSI prototypes in all header files (thus the old K&R definitions
3174 aren't needed for error creation any more) and do a better job of
3175 translating function codes into names. The old 'ASN1 error code imbedded
3176 in a comment' is no longer necessary and it doesn't use .err files which
3177 have now been deleted. Also the error code call doesn't have to appear all
3178 on one line (which resulted in some large lines...).
3181 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3184 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3185 0 (which usually indicates a closed connection), but continue reading.
3188 *) Fix some race conditions.
3191 *) Add support for CRL distribution points extension. Add Certificate
3192 Policies and CRL distribution points documentation.
3195 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3198 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3199 8 of keying material. Merlin has also confirmed interop with this fix
3200 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3201 [Merlin Hughes <merlin@baltimore.ie>]
3203 *) Fix lots of warnings.
3204 [Richard Levitte <levitte@stacken.kth.se>]
3206 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3207 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3208 [Richard Levitte <levitte@stacken.kth.se>]
3210 *) Fix problems with sizeof(long) == 8.
3211 [Andy Polyakov <appro@fy.chalmers.se>]
3213 *) Change functions to ANSI C.
3216 *) Fix typos in error codes.
3217 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3219 *) Remove defunct assembler files from Configure.
3222 *) SPARC v8 assembler BIGNUM implementation.
3223 [Andy Polyakov <appro@fy.chalmers.se>]
3225 *) Support for Certificate Policies extension: both print and set.
3226 Various additions to support the r2i method this uses.
3229 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3230 return a const string when you are expecting an allocated buffer.
3233 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3234 types DirectoryString and DisplayText.
3237 *) Add code to allow r2i extensions to access the configuration database,
3238 add an LHASH database driver and add several ctx helper functions.
3241 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3242 fail when they extended the size of a BIGNUM.
3245 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3246 support typesafe stack.
3249 *) Fix typo in SSL_[gs]et_options().
3250 [Nils Frostberg <nils@medcom.se>]
3252 *) Delete various functions and files that belonged to the (now obsolete)
3253 old X509V3 handling code.
3256 *) New Configure option "rsaref".
3259 *) Don't auto-generate pem.h.
3262 *) Introduce type-safe ASN.1 SETs.
3265 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3266 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3268 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3269 that links with OpenSSL (well at least cause lots of warnings), but fear
3270 not: the conversion is trivial, and it eliminates loads of evil casts. A
3271 few STACKed things have been converted already. Feel free to convert more.
3272 In the fullness of time, I'll do away with the STACK type altogether.
3275 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3276 specified in <certfile> by updating the entry in the index.txt file.
3277 This way one no longer has to edit the index.txt file manually for
3278 revoking a certificate. The -revoke option does the gory details now.
3279 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3281 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3282 `-text' option at all and this way the `-noout -text' combination was
3283 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3284 [Ralf S. Engelschall]
3286 *) Make sure a corresponding plain text error message exists for the
3287 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3288 verify callback function determined that a certificate was revoked.
3289 [Ralf S. Engelschall]
3291 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3292 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3293 all available cipers including rc5, which was forgotten until now.
3294 In order to let the testing shell script know which algorithms
3295 are available, a new (up to now undocumented) command
3296 "openssl list-cipher-commands" is used.
3299 *) Bugfix: s_client occasionally would sleep in select() when
3300 it should have checked SSL_pending() first.
3303 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3304 the raw DSA values prior to ASN.1 encoding.
3307 *) Tweaks to Configure
3308 [Niels Poppe <niels@netbox.org>]
3310 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3314 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3317 *) New config option to avoid instructions that are illegal on the 80386.
3318 The default code is faster, but requires at least a 486.
3321 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3322 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3323 same as SSL2_VERSION anyway.
3326 *) New "-showcerts" option for s_client.
3329 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3330 application. Various cleanups and fixes.
3333 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3334 modify error routines to work internally. Add error codes and PBE init
3335 to library startup routines.
3338 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3339 packing functions to asn1 and evp. Changed function names and error
3340 codes along the way.
3343 *) PKCS12 integration: and so it begins... First of several patches to
3344 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3345 objects to objects.h
3348 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3349 and display support for Thawte strong extranet extension.
3352 *) Add LinuxPPC support.
3353 [Jeff Dubrule <igor@pobox.org>]
3355 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3356 bn_div_words in alpha.s.
3357 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3359 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3360 OAEP isn't supported when OpenSSL is built with RSAref.
3361 [Ulf Moeller <ulf@fitug.de>]
3363 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3364 so they no longer are missing under -DNOPROTO.
3365 [Soren S. Jorvang <soren@t.dk>]
3368 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3370 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3371 doesn't work when the session is reused. Coming soon!
3374 *) Fix a security hole, that allows sessions to be reused in the wrong
3375 context thus bypassing client cert protection! All software that uses
3376 client certs and session caches in multiple contexts NEEDS PATCHING to
3377 allow session reuse! A fuller solution is in the works.
3378 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3380 *) Some more source tree cleanups (removed obsolete files
3381 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3382 permission on "config" script to be executable) and a fix for the INSTALL
3384 [Ulf Moeller <ulf@fitug.de>]
3386 *) Remove some legacy and erroneous uses of malloc, free instead of
3388 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3390 *) Make rsa_oaep_test return non-zero on error.
3391 [Ulf Moeller <ulf@fitug.de>]
3393 *) Add support for native Solaris shared libraries. Configure
3394 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3395 if someone would make that last step automatic.
3396 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3398 *) ctx_size was not built with the right compiler during "make links". Fixed.
3401 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3402 except NULL ciphers". This means the default cipher list will no longer
3403 enable NULL ciphers. They need to be specifically enabled e.g. with
3404 the string "DEFAULT:eNULL".
3407 *) Fix to RSA private encryption routines: if p < q then it would
3408 occasionally produce an invalid result. This will only happen with
3409 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3412 *) Be less restrictive and allow also `perl util/perlpath.pl
3413 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3414 because this way one can also use an interpreter named `perl5' (which is
3415 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3416 installed as `perl').
3417 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3419 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3420 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3422 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3423 advapi32.lib to Win32 build and change the pem test comparision
3424 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3425 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3426 and crypto/des/ede_cbcm_enc.c.
3429 *) DES quad checksum was broken on big-endian architectures. Fixed.
3432 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3433 Win32 test batch file so it (might) work again. The Win32 test batch file
3434 is horrible: I feel ill....
3437 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3438 in e_os.h. Audit of header files to check ANSI and non ANSI
3439 sections: 10 functions were absent from non ANSI section and not exported
3440 from Windows DLLs. Fixed up libeay.num for new functions.
3443 *) Make `openssl version' output lines consistent.
3444 [Ralf S. Engelschall]
3446 *) Fix Win32 symbol export lists for BIO functions: Added
3447 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3448 to ms/libeay{16,32}.def.
3449 [Ralf S. Engelschall]
3451 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3452 fine under Unix and passes some trivial tests I've now added. But the
3453 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3454 added to make sure no one expects that this stuff really works in the
3455 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3456 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3458 [Ralf S. Engelschall]
3460 *) Fix the generation of two part addresses in perl.
3461 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3463 *) Add config entry for Linux on MIPS.
3464 [John Tobey <jtobey@channel1.com>]
3466 *) Make links whenever Configure is run, unless we are on Windoze.
3469 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3470 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3474 *) Add a useful kludge to allow package maintainers to specify compiler and
3475 other platforms details on the command line without having to patch the
3476 Configure script everytime: One now can use ``perl Configure
3477 <id>:<details>'', i.e. platform ids are allowed to have details appended
3478 to them (seperated by colons). This is treated as there would be a static
3479 pre-configured entry in Configure's %table under key <id> with value
3480 <details> and ``perl Configure <id>'' is called. So, when you want to
3481 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3482 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3483 now, which overrides the FreeBSD-elf entry on-the-fly.
3484 [Ralf S. Engelschall]
3486 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3489 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3490 on the `perl Configure ...' command line. This way one can compile
3491 OpenSSL libraries with Position Independent Code (PIC) which is needed
3492 for linking it into DSOs.
3493 [Ralf S. Engelschall]
3495 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3499 *) Cleaned up the LICENSE document: The official contact for any license
3500 questions now is the OpenSSL core team under openssl-core@openssl.org.
3501 And add a paragraph about the dual-license situation to make sure people
3502 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3503 to the OpenSSL toolkit.
3504 [Ralf S. Engelschall]
3506 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3507 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3508 Additonally cleaned up the `make links' target: Remove unnecessary
3509 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3510 to speed processing and no longer clutter the display with confusing
3511 stuff. Instead only the actually done links are displayed.
3512 [Ralf S. Engelschall]
3514 *) Permit null encryption ciphersuites, used for authentication only. It used
3515 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3516 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3520 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3521 signed attributes when verifying signatures (this would break them),
3522 the detached data encoding was wrong and public keys obtained using
3523 X509_get_pubkey() weren't freed.
3526 *) Add text documentation for the BUFFER functions. Also added a work around
3527 to a Win95 console bug. This was triggered by the password read stuff: the
3528 last character typed gets carried over to the next fread(). If you were
3529 generating a new cert request using 'req' for example then the last
3530 character of the passphrase would be CR which would then enter the first
3534 *) Added the new `Includes OpenSSL Cryptography Software' button as
3535 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3536 button and can be used by applications based on OpenSSL to show the
3537 relationship to the OpenSSL project.
3538 [Ralf S. Engelschall]
3540 *) Remove confusing variables in function signatures in files
3541 ssl/ssl_lib.c and ssl/ssl.h.
3542 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3544 *) Don't install bss_file.c under PREFIX/include/
3545 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3547 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3548 functions that return function pointers and has support for NT specific
3549 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3550 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3551 unsigned to signed types: this was killing the Win32 compile.
3554 *) Add new certificate file to stack functions,
3555 SSL_add_dir_cert_subjects_to_stack() and
3556 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3557 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3558 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3559 This means that Apache-SSL and similar packages don't have to mess around
3560 to add as many CAs as they want to the preferred list.
3563 *) Experiment with doxygen documentation. Currently only partially applied to
3565 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3566 openssl.doxy as the configuration file.
3569 *) Get rid of remaining C++-style comments which strict C compilers hate.
3570 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3572 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3573 compiled in by default: it has problems with large keys.
3576 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3577 DH private keys and/or callback functions which directly correspond to
3578 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3579 is needed for applications which have to configure certificates on a
3580 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3582 For the RSA certificate situation is makes no difference, but
3583 for the DSA certificate situation this fixes the "no shared cipher"