4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7 keyUsage if basicConstraints absent for a CA.
10 *) Make SMIME_write_PKCS7() write mail header values with a format that
11 is more generally accepted (no spaces before the semicolon), since
12 some programs can't parse those values properly otherwise. Also make
13 sure BIO's that break lines after each write do not create invalid
17 *) Make sk_sort() tolerate a NULL argument.
18 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
20 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
21 passed by the function are trusted implicitly. If any of them signed the
22 reponse then it is assumed to be valid and is not verified.
25 *) Zero the premaster secret after deriving the master secret in
29 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
30 to data. This was previously part of the PKCS7 ASN1 code. This
31 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
32 [Steve Henson, reported by Kenneth R. Robinette
33 <support@securenetterm.com>]
35 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
36 routines: without these tracing memory leaks is very painful.
37 Fix leaks in PKCS12 and PKCS7 routines.
40 *) Fix for Irix with NO_ASM.
41 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
43 *) Add some EVP_add_digest_alias registrations (as found in
44 OpenSSL_add_all_digests()), to SSL_library_init()
45 aka OpenSSL_add_ssl_algorithms(). This provides improved
46 compatibility with peers using X.509 certificates
47 with unconventional AlgorithmIdentifier OIDs.
50 *) ./config script fixes.
51 [Ulf Moeller, Richard Levitte]
53 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
54 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
55 effectively meant GeneralizedTime would never be used. Now it
56 is initialised to -1 but X509_time_adj() now has to check the value
57 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
58 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
59 [Steve Henson, reported by Kenneth R. Robinette
60 <support@securenetterm.com>]
62 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
63 result in a zero length in the ASN1_INTEGER structure which was
64 not consistent with the structure when d2i_ASN1_INTEGER() was used
65 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
66 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
67 where it did not print out a minus for negative ASN1_INTEGER.
70 *) Fix 'openssl passwd -1'.
73 *) Add summary printout to ocsp utility. The various functions which
74 convert status values to strings have been renamed to:
75 OCSP_response_status_str(), OCSP_cert_status_str() and
76 OCSP_crl_reason_str() and are no longer static. New options
77 to verify nonce values and to disable verification. OCSP response
78 printout format cleaned up.
81 *) Add additional OCSP certificate checks. These are those specified
82 in RFC2560. This consists of two separate checks: the CA of the
83 certificate being checked must either be the OCSP signer certificate
84 or the issuer of the OCSP signer certificate. In the latter case the
85 OCSP signer certificate must contain the OCSP signing extended key
86 usage. This check is performed by attempting to match the OCSP
87 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
88 in the OCSP_CERTID structures of the response.
91 *) Initial OCSP certificate verification added to OCSP_basic_verify()
92 and related routines. This uses the standard OpenSSL certificate
93 verify routines to perform initial checks (just CA validity) and
94 to obtain the certificate chain. Then additional checks will be
95 performed on the chain. Currently the root CA is checked to see
96 if it is explicitly trusted for OCSP signing. This is used to set
97 a root CA as a global signing root: that is any certificate that
98 chains to that CA is an acceptable OCSP signing certificate.
101 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
102 extensions from a separate configuration file.
103 As when reading extensions from the main configuration file,
104 the '-extensions ...' option may be used for specifying the
106 [Massimiliano Pala <madwolf@comune.modena.it>]
108 *) Change PKCS12_key_gen_asc() so it can cope with non null
109 terminated strings whose length is passed in the passlen
110 parameter, for example from PEM callbacks. This was done
111 by adding an extra length parameter to asc2uni().
112 [Steve Henson, reported by <oddissey@samsung.co.kr>]
114 *) New OCSP utility. Allows OCSP requests to be generated or
115 read. The request can be sent to a responder and the output
116 parsed, outputed or printed in text form. Not complete yet:
117 still needs to check the OCSP response validity.
120 *) New subcommands for 'openssl ca':
121 'openssl ca -status <serial>' prints the status of the cert with
122 the given serial number (according to the index file).
123 'openssl ca -updatedb' updates the expiry status of certificates
125 [Massimiliano Pala <madwolf@comune.modena.it>]
127 *) New '-newreq-nodes' command option to CA.pl. This is like
128 '-newreq', but calls 'openssl req' with the '-nodes' option
129 so that the resulting key is not encrypted.
130 [Damien Miller <djm@mindrot.org>]
132 *) New configuration for the GNU Hurd.
133 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
135 *) Initial code to implement OCSP basic response verify. This
136 is currently incomplete. Currently just finds the signer's
137 certificate and verifies the signature on the response.
140 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
141 value of OPENSSLDIR. This is available via the new '-d' option
142 to 'openssl version', and is also included in 'openssl version -a'.
145 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
146 call failed, free the DSA structure.
149 *) Allowing defining memory allocation callbacks that will be given
150 file name and line number information in additional arguments
151 (a const char* and an int). The basic functionality remains, as
152 well as the original possibility to just replace malloc(),
153 realloc() and free() by functions that do not know about these
154 additional arguments. To register and find out the current
155 settings for extended allocation functions, the following
156 functions are provided:
158 CRYPTO_set_mem_ex_functions
159 CRYPTO_set_locked_mem_ex_functions
160 CRYPTO_get_mem_ex_functions
161 CRYPTO_get_locked_mem_ex_functions
163 These work the same way as CRYPTO_set_mem_functions and friends.
164 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
165 extended allocation function is enabled.
166 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
167 a conventional allocation function is enabled.
168 [Richard Levitte, Bodo Moeller]
170 *) Fix to uni2asc() to cope with zero length Unicode strings.
171 These are present in some PKCS#12 files.
174 *) Finish off removing the remaining LHASH function pointer casts.
175 There should no longer be any prototype-casting required when using
176 the LHASH abstraction, and any casts that remain are "bugs". See
177 the callback types and macros at the head of lhash.h for details
178 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
181 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
182 If an EGD or PRNGD is running and enough entropy is returned, automatic
183 seeding like with /dev/[u]random will be performed.
184 Positions tried are: /etc/entropy, /var/run/egd-pool.
187 *) Change the Unix RAND_poll() variant to be able to poll several
188 random devices, as specified by DEVRANDOM, until a sufficient amount
189 of data has been collected. We spend at most 10 ms on each file
190 (select timeout) and read in non-blocking mode. DEVRANDOM now
191 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
192 (previously it was just the string "/dev/urandom"), so on typical
193 platforms the 10 ms delay will never occur.
194 Also separate out the Unix variant to its own file, rand_unix.c.
195 For VMS, there's a currently-empty rand_vms.c.
198 *) Move OCSP client related routines to ocsp_cl.c. These
199 provide utility functions which an application needing
200 to issue a request to an OCSP responder and analyse the
201 response will typically need: as opposed to those which an
202 OCSP responder itself would need which will be added later.
204 OCSP_request_sign() signs an OCSP request with an API similar
205 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
206 response. OCSP_response_get1_basic() extracts basic response
207 from response. OCSP_resp_find_status(): finds and extracts status
208 information from an OCSP_CERTID structure (which will be created
209 when the request structure is built). These are built from lower
210 level functions which work on OCSP_SINGLERESP structures but
211 wont normally be used unless the application wishes to examine
212 extensions in the OCSP response for example.
214 Replace nonce routines with a pair of functions.
215 OCSP_request_add1_nonce() adds a nonce value and optionally
216 generates a random value. OCSP_check_nonce() checks the
217 validity of the nonce in an OCSP response.
220 *) Change function OCSP_request_add() to OCSP_request_add0_id().
221 This doesn't copy the supplied OCSP_CERTID and avoids the
222 need to free up the newly created id. Change return type
223 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
224 This can then be used to add extensions to the request.
225 Deleted OCSP_request_new(), since most of its functionality
226 is now in OCSP_REQUEST_new() (and the case insensitive name
227 clash) apart from the ability to set the request name which
228 will be added elsewhere.
231 *) Update OCSP API. Remove obsolete extensions argument from
232 various functions. Extensions are now handled using the new
233 OCSP extension code. New simple OCSP HTTP function which
234 can be used to send requests and parse the response.
237 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
238 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
239 uses the special reorder version of SET OF to sort the attributes
240 and reorder them to match the encoded order. This resolves a long
241 standing problem: a verify on a PKCS7 structure just after signing
242 it used to fail because the attribute order did not match the
243 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
244 it uses the received order. This is necessary to tolerate some broken
245 software that does not order SET OF. This is handled by encoding
246 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
247 to produce the required SET OF.
250 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
251 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
252 files to get correct declarations of the ASN.1 item variables.
255 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
256 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
257 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
258 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
259 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
260 ASN1_ITEM and no wrapper functions.
263 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
264 replace the old function pointer based I/O routines. Change most of
265 the *_d2i_bio() and *_d2i_fp() functions to use these.
268 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
269 lines, recognice more "algorithms" that can be deselected, and make
270 it complain about algorithm deselection that isn't recognised.
273 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
274 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
275 to use new functions. Add NO_ASN1_OLD which can be set to remove
276 some old style ASN1 functions: this can be used to determine if old
277 code will still work when these eventually go away.
280 *) New extension functions for OCSP structures, these follow the
281 same conventions as certificates and CRLs.
284 *) New function X509V3_add1_i2d(). This automatically encodes and
285 adds an extension. Its behaviour can be customised with various
286 flags to append, replace or delete. Various wrappers added for
287 certifcates and CRLs.
290 *) Fix to avoid calling the underlying ASN1 print routine when
291 an extension cannot be parsed. Correct a typo in the
292 OCSP_SERVICELOC extension. Tidy up print OCSP format.
295 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
296 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
297 when writing a 32767 byte record.
298 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
300 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
301 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
303 (RSA objects have a reference count access to which is protected
304 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
305 so they are meant to be shared between threads.)
306 [Bodo Moeller, Geoff Thorpe; original patch submitted by
307 "Reddie, Steven" <Steven.Reddie@ca.com>]
309 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
310 entries for variables.
313 *) Fix a deadlock in CRYPTO_mem_leaks().
316 *) Add functionality to apps/openssl.c for detecting locking
317 problems: As the program is single-threaded, all we have
318 to do is register a locking callback using an array for
319 storing which locks are currently held by the program.
322 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
323 SSL_get_ex_data_X509_STORE_idx(), which is used in
324 ssl_verify_cert_chain() and thus can be called at any time
325 during TLS/SSL handshakes so that thread-safety is essential.
326 Unfortunately, the ex_data design is not at all suited
327 for multi-threaded use, so it probably should be abolished.
330 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
331 [Broadcom, tweaked and integrated by Geoff Thorpe]
333 *) Move common extension printing code to new function
334 X509V3_print_extensions(). Reorganise OCSP print routines and
335 implement some needed OCSP ASN1 functions. Add OCSP extensions.
338 *) New function X509_signature_print() to remove duplication in some
342 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
343 set (this was treated exactly the same as SET OF previously). This
344 is used to reorder the STACK representing the structure to match the
345 encoding. This will be used to get round a problem where a PKCS7
346 structure which was signed could not be verified because the STACK
347 order did not reflect the encoded order.
350 *) Reimplement the OCSP ASN1 module using the new code.
353 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
354 for its ASN1 operations. The old style function pointers still exist
355 for now but they will eventually go away.
358 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
359 completely replaces the old ASN1 functionality with a table driven
360 encoder and decoder which interprets an ASN1_ITEM structure describing
361 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
362 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
363 has also been converted to the new form.
366 *) Change BN_mod_exp_recp so that negative moduli are tolerated
367 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
368 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
372 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
373 of not touching the result's sign bit.
376 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
380 *) Changed the LHASH code to use prototypes for callbacks, and created
381 macros to declare and implement thin (optionally static) functions
382 that provide type-safety and avoid function pointer casting for the
383 type-specific callbacks.
386 *) Use better test patterns in bntest.
389 *) Added Kerberos Cipher Suites to be used with TLS, as written in
391 [Veers Staats <staatsvr@asc.hpc.mil>,
392 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
394 *) rand_win.c fix for Borland C.
397 *) BN_rshift bugfix for n == 0.
400 *) Reformat the FAQ so the different questions and answers can be divided
401 in sections depending on the subject.
404 *) Have the zlib compression code load ZLIB.DLL dynamically under
408 *) New function BN_mod_sqrt for computing square roots modulo a prime
409 (using the probabilistic Tonelli-Shanks algorithm unless
410 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
411 be handled deterministically).
412 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
414 *) Store verify_result within SSL_SESSION also for client side to
415 avoid potential security hole. (Re-used sessions on the client side
416 always resulted in verify_result==X509_V_OK, not using the original
417 result of the server certificate verification.)
420 *) Make BN_mod_inverse faster by explicitly handling small quotients
421 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
422 512 bits], about 30% for larger ones [1024 or 2048 bits].)
425 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
426 SSL3_RT_APPLICATION_DATA, return 0.
427 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
431 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
432 releases, have been re-implemented by renaming the previous
433 implementations of ssl2_read and ssl3_read to ssl2_read_internal
434 and ssl3_read_internal, respectively, and adding 'peek' parameters
435 to them. The new ssl[23]_{read,peek} functions are calls to
436 ssl[23]_read_internal with the 'peek' flag set appropriately.
437 A 'peek' parameter has also been added to ssl3_read_bytes, which
438 does the actual work for ssl3_read_internal.
441 *) New function BN_kronecker.
444 *) Fix BN_gcd so that it works on negative inputs; the result is
445 positive unless both parameters are zero.
446 Previously something reasonably close to an infinite loop was
447 possible because numbers could be growing instead of shrinking
448 in the implementation of Euclid's algorithm.
451 *) Fix BN_is_word() and BN_is_one() macros to take into account the
452 sign of the number in question.
454 Fix BN_is_word(a,w) to work correctly for w == 0.
456 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
457 because its test if the absolute value of 'a' equals 'w'.
458 Note that BN_abs_is_word does *not* handle w == 0 reliably;
459 it exists mostly for use in the implementations of BN_is_zero(),
460 BN_is_one(), and BN_is_word().
463 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
464 the method-specific "init()" handler. Also clean up ex_data after
465 calling the method-specific "finish()" handler. Previously, this was
466 happening the other way round.
469 *) New function BN_swap.
472 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
473 the exponentiation functions are more likely to produce reasonable
474 results on negative inputs.
477 *) Change BN_mod_mul so that the result is always non-negative.
478 Previously, it could be negative if one of the factors was negative;
479 I don't think anyone really wanted that behaviour.
482 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
483 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
484 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
485 and add new functions:
498 These functions always generate non-negative results.
500 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
501 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
503 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
504 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
506 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
508 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
509 was actually never needed) and in BN_mul(). The removal in BN_mul()
510 required a small change in bn_mul_part_recursive() and the addition
511 of the functions bn_cmp_part_words(), bn_sub_part_words() and
512 bn_add_part_words(), which do the same thing as bn_cmp_words(),
513 bn_sub_words() and bn_add_words() except they take arrays with
517 *) In 'openssl passwd', verify passwords read from the terminal
518 unless the '-salt' option is used (which usually means that
519 verification would just waste user's time since the resulting
520 hash is going to be compared with some given password hash)
521 or the new '-noverify' option is used.
523 This is an incompatible change, but it does not affect
524 non-interactive use of 'openssl passwd' (passwords on the command
525 line, '-stdin' option, '-in ...' option) and thus should not
529 *) Remove all references to RSAref, since there's no more need for it.
532 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
533 The previous value, 12, was not always sufficient for BN_mod_exp().
536 *) Make DSO load along a path given through an environment variable
537 (SHLIB_PATH) with shl_load().
540 *) Constify the ENGINE code as a result of BIGNUM constification.
541 Also constify the RSA code and most things related to it. In a
542 few places, most notable in the depth of the ASN.1 code, ugly
543 casts back to non-const were required (to be solved at a later
547 *) Make it so the openssl application has all engines loaded by default.
550 *) Constify the BIGNUM routines a little more.
553 *) Make sure that shared libraries get the internal name engine with
554 the full version number and not just 0. This should mark the
555 shared libraries as not backward compatible. Of course, this should
556 be changed again when we can guarantee backward binary compatibility.
559 *) Add the following functions:
565 ENGINE_load_builtin_engines()
567 That way, an application can itself choose if external engines that
568 are built-in in OpenSSL shall ever be used or not. The benefit is
569 that applications won't have to be linked with libdl or other dso
570 libraries unless it's really needed.
572 Changed 'openssl engine' to load all engines on demand.
573 Changed the engine header files to avoid the duplication of some
574 declarations (they differed!).
577 *) 'openssl engine' can now list capabilities.
580 *) Better error reporting in 'openssl engine'.
583 *) Never call load_dh_param(NULL) in s_server.
586 *) Add engine application. It can currently list engines by name and
587 identity, and test if they are actually available.
590 *) Add support for shared libraries under Irix.
591 [Albert Chin-A-Young <china@thewrittenword.com>]
593 *) Improve RPM specification file by forcing symbolic linking and making
594 sure the installed documentation is also owned by root.root.
595 [Damien Miller <djm@mindrot.org>]
597 *) Add configuration option to build on Linux on both big-endian and
599 [Ralf Baechle <ralf@uni-koblenz.de>]
601 *) Give the OpenSSL applications more possibilities to make use of
602 keys (public as well as private) handled by engines.
605 *) Add OCSP code that comes from CertCo.
608 *) Add VMS support for the Rijndael code.
611 *) Added untested support for Nuron crypto accelerator.
614 *) Add support for external cryptographic devices. This code was
615 previously distributed separately as the "engine" branch.
616 [Geoff Thorpe, Richard Levitte]
618 *) Rework the filename-translation in the DSO code. It is now possible to
619 have far greater control over how a "name" is turned into a filename
620 depending on the operating environment and any oddities about the
621 different shared library filenames on each system.
624 *) Support threads on FreeBSD-elf in Configure.
627 *) Add the possibility to create shared libraries on HP-UX
630 *) Fix for SHA1 assembly problem with MASM: it produces
631 warnings about corrupt line number information when assembling
632 with debugging information. This is caused by the overlapping
634 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
637 NCONF_get_number() has no error checking at all. As a replacement,
638 NCONF_get_number_e() is defined (_e for "error checking") and is
639 promoted strongly. The old NCONF_get_number is kept around for
640 binary backward compatibility.
641 Make it possible for methods to load from something other than a BIO,
642 by providing a function pointer that is given a name instead of a BIO.
643 For example, this could be used to load configuration data from an
647 *) Fix typo in get_cert_by_subject() in by_dir.c
648 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
650 *) Rework the system to generate shared libraries:
652 - Make note of the expected extension for the shared libraries and
653 if there is a need for symbolic links from for example libcrypto.so.0
654 to libcrypto.so.0.9.7. There is extended info in Configure for
657 - Make as few rebuilds of the shared libraries as possible.
659 - Still avoid linking the OpenSSL programs with the shared libraries.
661 - When installing, install the shared libraries separately from the
665 *) Fix for non blocking accept BIOs. Added new I/O special reason
666 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
667 with non blocking I/O was not possible because no retry code was
668 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
672 *) Added the beginnings of Rijndael support.
675 *) Fix for bug in DirectoryString mask setting. Add support for
676 X509_NAME_print_ex() in 'req' and X509_print_ex() function
677 to allow certificate printing to more controllable, additional
678 'certopt' option to 'x509' to allow new printing options to be
682 *) Clean old EAY MD5 hack from e_os.h.
685 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
687 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
688 and not in SSL_clear because the latter is also used by the
689 accept/connect functions; previously, the settings made by
690 SSL_set_read_ahead would be lost during the handshake.
691 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
693 *) Correct util/mkdef.pl to be selective about disabled algorithms.
694 Previously, it would create entries for disableed algorithms no
698 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
700 *) In ssl23_get_client_hello, generate an error message when faced
701 with an initial SSL 3.0/TLS record that is too small to contain the
702 first two bytes of the ClientHello message, i.e. client_version.
703 (Note that this is a pathologic case that probably has never happened
704 in real life.) The previous approach was to use the version number
705 from the record header as a substitute; but our protocol choice
706 should not depend on that one because it is not authenticated
707 by the Finished messages.
710 *) More robust randomness gathering functions for Windows.
711 [Jeffrey Altman <jaltman@columbia.edu>]
713 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
714 not set then we don't setup the error code for issuer check errors
715 to avoid possibly overwriting other errors which the callback does
716 handle. If an application does set the flag then we assume it knows
717 what it is doing and can handle the new informational codes
721 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
722 a general "ANY" type, as such it should be able to decode anything
723 including tagged types. However it didn't check the class so it would
724 wrongly interpret tagged types in the same way as their universal
725 counterpart and unknown types were just rejected. Changed so that the
726 tagged and unknown types are handled in the same way as a SEQUENCE:
727 that is the encoding is stored intact. There is also a new type
728 "V_ASN1_OTHER" which is used when the class is not universal, in this
729 case we have no idea what the actual type is so we just lump them all
733 *) On VMS, stdout may very well lead to a file that is written to
734 in a record-oriented fashion. That means that every write() will
735 write a separate record, which will be read separately by the
736 programs trying to read from it. This can be very confusing.
738 The solution is to put a BIO filter in the way that will buffer
739 text until a linefeed is reached, and then write everything a
740 line at a time, so every record written will be an actual line,
741 not chunks of lines and not (usually doesn't happen, but I've
742 seen it once) several lines in one record. BIO_f_linebuffer() is
745 Currently, it's a VMS-only method, because that's where it has
746 been tested well enough.
749 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
750 it can return incorrect results.
751 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
752 but it was in 0.9.6-beta[12].)
755 *) Disable the check for content being present when verifying detached
756 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
757 include zero length content when signing messages.
760 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
761 BIO_ctrl (for BIO pairs).
764 *) Add DSO method for VMS.
767 *) Bug fix: Montgomery multiplication could produce results with the
771 *) Add RPM specification openssl.spec and modify it to build three
772 packages. The default package contains applications, application
773 documentation and run-time libraries. The devel package contains
774 include files, static libraries and function documentation. The
775 doc package contains the contents of the doc directory. The original
776 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
779 *) Add a large number of documentation files for many SSL routines.
780 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
782 *) Add a configuration entry for Sony News 4.
783 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
785 *) Don't set the two most significant bits to one when generating a
786 random number < q in the DSA library.
789 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
790 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
791 the underlying transport is blocking) if a handshake took place.
792 (The default behaviour is needed by applications such as s_client
793 and s_server that use select() to determine when to use SSL_read;
794 but for applications that know in advance when to expect data, it
795 just makes things more complicated.)
798 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
802 *) Add a few more EBCDIC conditionals that make `req' and `x509'
803 work better on such systems.
804 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
806 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
807 Update PKCS12_parse() so it copies the friendlyName and the
808 keyid to the certificates aux info.
811 *) Fix bug in PKCS7_verify() which caused an infinite loop
812 if there was more than one signature.
813 [Sven Uszpelkat <su@celocom.de>]
815 *) Major change in util/mkdef.pl to include extra information
816 about each symbol, as well as presentig variables as well
817 as functions. This change means that there's n more need
818 to rebuild the .num files when some algorithms are excluded.
821 *) Allow the verify time to be set by an application,
822 rather than always using the current time.
825 *) Phase 2 verify code reorganisation. The certificate
826 verify code now looks up an issuer certificate by a
827 number of criteria: subject name, authority key id
828 and key usage. It also verifies self signed certificates
829 by the same criteria. The main comparison function is
830 X509_check_issued() which performs these checks.
832 Lot of changes were necessary in order to support this
833 without completely rewriting the lookup code.
835 Authority and subject key identifier are now cached.
837 The LHASH 'certs' is X509_STORE has now been replaced
838 by a STACK_OF(X509_OBJECT). This is mainly because an
839 LHASH can't store or retrieve multiple objects with
842 As a result various functions (which were all internal
843 use only) have changed to handle the new X509_STORE
844 structure. This will break anything that messed round
845 with X509_STORE internally.
847 The functions X509_STORE_add_cert() now checks for an
848 exact match, rather than just subject name.
850 The X509_STORE API doesn't directly support the retrieval
851 of multiple certificates matching a given criteria, however
852 this can be worked round by performing a lookup first
853 (which will fill the cache with candidate certificates)
854 and then examining the cache for matches. This is probably
855 the best we can do without throwing out X509_LOOKUP
856 entirely (maybe later...).
858 The X509_VERIFY_CTX structure has been enhanced considerably.
860 All certificate lookup operations now go via a get_issuer()
861 callback. Although this currently uses an X509_STORE it
862 can be replaced by custom lookups. This is a simple way
863 to bypass the X509_STORE hackery necessary to make this
864 work and makes it possible to use more efficient techniques
865 in future. A very simple version which uses a simple
866 STACK for its trusted certificate store is also provided
867 using X509_STORE_CTX_trusted_stack().
869 The verify_cb() and verify() callbacks now have equivalents
870 in the X509_STORE_CTX structure.
872 X509_STORE_CTX also has a 'flags' field which can be used
873 to customise the verify behaviour.
876 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
877 excludes S/MIME capabilities.
880 *) When a certificate request is read in keep a copy of the
881 original encoding of the signed data and use it when outputing
882 again. Signatures then use the original encoding rather than
883 a decoded, encoded version which may cause problems if the
884 request is improperly encoded.
887 *) For consistency with other BIO_puts implementations, call
888 buffer_write(b, ...) directly in buffer_puts instead of calling
891 In BIO_puts, increment b->num_write as in BIO_write.
892 [Peter.Sylvester@EdelWeb.fr]
894 *) Fix BN_mul_word for the case where the word is 0. (We have to use
895 BN_zero, we may not return a BIGNUM with an array consisting of
899 *) Avoid calling abort() from within the library when problems are
900 detected, except if preprocessor symbols have been defined
901 (such as REF_CHECK, BN_DEBUG etc.).
904 *) New openssl application 'rsautl'. This utility can be
905 used for low level RSA operations. DER public key
906 BIO/fp routines also added.
909 *) New Configure entry and patches for compiling on QNX 4.
910 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
912 *) A demo state-machine implementation was sponsored by
913 Nuron (http://www.nuron.com/) and is now available in
917 *) New options added to the 'dgst' utility for signature
918 generation and verification.
921 *) Unrecognized PKCS#7 content types are now handled via a
922 catch all ASN1_TYPE structure. This allows unsupported
923 types to be stored as a "blob" and an application can
924 encode and decode it manually.
927 *) Fix various signed/unsigned issues to make a_strex.c
929 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
931 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
932 length if passed a buffer. ASN1_INTEGER_to_BN failed
933 if passed a NULL BN and its argument was negative.
934 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
936 *) Modification to PKCS#7 encoding routines to output definite
937 length encoding. Since currently the whole structures are in
938 memory there's not real point in using indefinite length
939 constructed encoding. However if OpenSSL is compiled with
940 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
943 *) Added BIO_vprintf() and BIO_vsnprintf().
946 *) Added more prefixes to parse for in the the strings written
947 through a logging bio, to cover all the levels that are available
948 through syslog. The prefixes are now:
950 PANIC, EMERG, EMR => LOG_EMERG
951 ALERT, ALR => LOG_ALERT
952 CRIT, CRI => LOG_CRIT
953 ERROR, ERR => LOG_ERR
954 WARNING, WARN, WAR => LOG_WARNING
955 NOTICE, NOTE, NOT => LOG_NOTICE
956 INFO, INF => LOG_INFO
957 DEBUG, DBG => LOG_DEBUG
959 and as before, if none of those prefixes are present at the
960 beginning of the string, LOG_ERR is chosen.
962 On Win32, the LOG_* levels are mapped according to this:
964 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
965 LOG_WARNING => EVENTLOG_WARNING_TYPE
966 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
970 *) Made it possible to reconfigure with just the configuration
971 argument "reconf" or "reconfigure". The command line arguments
972 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
973 and are retrieved from there when reconfiguring.
977 [Assar Westerlund <assar@sics.se>, Richard Levitte]
979 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
982 *) The obj_dat.pl script was messing up the sorting of object
983 names. The reason was that it compared the quoted version
984 of strings as a result "OCSP" > "OCSP Signing" because
985 " > SPACE. Changed script to store unquoted versions of
986 names and add quotes on output. It was also omitting some
987 names from the lookup table if they were given a default
988 value (that is if SN is missing it is given the same
989 value as LN and vice versa), these are now added on the
990 grounds that if an object has a name we should be able to
991 look it up. Finally added warning output when duplicate
992 short or long names are found.
995 *) Changes needed for Tandem NSK.
996 [Scott Uroff <scott@xypro.com>]
998 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
999 RSA_padding_check_SSLv23(), special padding was never detected
1000 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1001 version rollback attacks was not effective.
1003 In s23_clnt.c, don't use special rollback-attack detection padding
1004 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1005 client; similarly, in s23_srvr.c, don't do the rollback check if
1006 SSL 2.0 is the only protocol enabled in the server.
1009 *) Make it possible to get hexdumps of unprintable data with 'openssl
1010 asn1parse'. By implication, the functions ASN1_parse_dump() and
1011 BIO_dump_indent() are added.
1014 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1015 these print out strings and name structures based on various
1016 flags including RFC2253 support and proper handling of
1017 multibyte characters. Added options to the 'x509' utility
1018 to allow the various flags to be set.
1021 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1022 Also change the functions X509_cmp_current_time() and
1023 X509_gmtime_adj() work with an ASN1_TIME structure,
1024 this will enable certificates using GeneralizedTime in validity
1025 dates to be checked.
1028 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1029 negative public key encodings) on by default,
1030 NO_NEG_PUBKEY_BUG can be set to disable it.
1033 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1034 content octets. An i2c_ASN1_OBJECT is unnecessary because
1035 the encoding can be trivially obtained from the structure.
1038 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1039 not read locks (CRYPTO_r_[un]lock).
1042 *) A first attempt at creating official support for shared
1043 libraries through configuration. I've kept it so the
1044 default is static libraries only, and the OpenSSL programs
1045 are always statically linked for now, but there are
1046 preparations for dynamic linking in place.
1047 This has been tested on Linux and True64.
1050 *) Randomness polling function for Win9x, as described in:
1051 Peter Gutmann, Software Generation of Practically Strong
1055 *) Fix so PRNG is seeded in req if using an already existing
1059 *) New options to smime application. -inform and -outform
1060 allow alternative formats for the S/MIME message including
1061 PEM and DER. The -content option allows the content to be
1062 specified separately. This should allow things like Netscape
1063 form signing output easier to verify.
1066 *) Fix the ASN1 encoding of tags using the 'long form'.
1069 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1070 STRING types. These convert content octets to and from the
1071 underlying type. The actual tag and length octets are
1072 already assumed to have been read in and checked. These
1073 are needed because all other string types have virtually
1074 identical handling apart from the tag. By having versions
1075 of the ASN1 functions that just operate on content octets
1076 IMPLICIT tagging can be handled properly. It also allows
1077 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1078 and ASN1_INTEGER are identical apart from the tag.
1081 *) Change the handling of OID objects as follows:
1083 - New object identifiers are inserted in objects.txt, following
1084 the syntax given in objects.README.
1085 - objects.pl is used to process obj_mac.num and create a new
1087 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1090 This is currently kind of a hack, and the perl code in objects.pl
1091 isn't very elegant, but it works as I intended. The simplest way
1092 to check that it worked correctly is to look in obj_dat.h and
1093 check the array nid_objs and make sure the objects haven't moved
1094 around (this is important!). Additions are OK, as well as
1095 consistent name changes.
1098 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1101 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1102 The given file adds to whatever has already been seeded into the
1103 random pool through the RANDFILE configuration file option or
1104 environment variable, or the default random state file.
1107 *) mkstack.pl now sorts each macro group into lexical order.
1108 Previously the output order depended on the order the files
1109 appeared in the directory, resulting in needless rewriting
1113 *) Patches to make OpenSSL compile under Win32 again. Mostly
1114 work arounds for the VC++ problem that it treats func() as
1115 func(void). Also stripped out the parts of mkdef.pl that
1116 added extra typesafe functions: these no longer exist.
1119 *) Reorganisation of the stack code. The macros are now all
1120 collected in safestack.h . Each macro is defined in terms of
1121 a "stack macro" of the form SKM_<name>(type, a, b). The
1122 DEBUG_SAFESTACK is now handled in terms of function casts,
1123 this has the advantage of retaining type safety without the
1124 use of additional functions. If DEBUG_SAFESTACK is not defined
1125 then the non typesafe macros are used instead. Also modified the
1126 mkstack.pl script to handle the new form. Needs testing to see
1127 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1128 the default if no major problems. Similar behaviour for ASN1_SET_OF
1129 and PKCS12_STACK_OF.
1132 *) When some versions of IIS use the 'NET' form of private key the
1133 key derivation algorithm is different. Normally MD5(password) is
1134 used as a 128 bit RC4 key. In the modified case
1135 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1136 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1137 as the old Netscape_RSA functions except they have an additional
1138 'sgckey' parameter which uses the modified algorithm. Also added
1139 an -sgckey command line option to the rsa utility. Thanks to
1140 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1141 algorithm to openssl-dev.
1144 *) The evp_local.h macros were using 'c.##kname' which resulted in
1145 invalid expansion on some systems (SCO 5.0.5 for example).
1146 Corrected to 'c.kname'.
1147 [Phillip Porch <root@theporch.com>]
1149 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1150 a STACK of email addresses from a certificate or request, these look
1151 in the subject name and the subject alternative name extensions and
1152 omit any duplicate addresses.
1155 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1156 This makes DSA verification about 2 % faster.
1159 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1160 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1161 plus overhead for 1024 bit moduli).
1162 This makes exponentiations about 0.5 % faster for 1024 bit
1163 exponents (as measured by "openssl speed rsa2048").
1166 *) Rename memory handling macros to avoid conflicts with other
1168 Malloc => OPENSSL_malloc
1169 Malloc_locked => OPENSSL_malloc_locked
1170 Realloc => OPENSSL_realloc
1171 Free => OPENSSL_free
1174 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1175 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1178 *) CygWin32 support.
1179 [John Jarvie <jjarvie@newsguy.com>]
1181 *) The type-safe stack code has been rejigged. It is now only compiled
1182 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1183 by default all type-specific stack functions are "#define"d back to
1184 standard stack functions. This results in more streamlined output
1185 but retains the type-safety checking possibilities of the original
1189 *) The STACK code has been cleaned up, and certain type declarations
1190 that didn't make a lot of sense have been brought in line. This has
1191 also involved a cleanup of sorts in safestack.h to more correctly
1192 map type-safe stack functions onto their plain stack counterparts.
1193 This work has also resulted in a variety of "const"ifications of
1194 lots of the code, especially "_cmp" operations which should normally
1195 be prototyped with "const" parameters anyway.
1198 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1199 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1200 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1201 where all of 'md' is used each time the PRNG is used, but 'state'
1202 is used only indexed by a cyclic counter. As entropy may not be
1203 well distributed from the beginning, 'md' is important as a
1204 chaining variable. However, the output function chains only half
1205 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1206 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1207 in all of 'state' being rewritten, with the new values depending
1208 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1211 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1212 the handshake is continued after ssl_verify_cert_chain();
1213 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1214 can lead to 'unexplainable' connection aborts later.
1215 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1217 *) Major EVP API cipher revision.
1218 Add hooks for extra EVP features. This allows various cipher
1219 parameters to be set in the EVP interface. Support added for variable
1220 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1221 setting of RC2 and RC5 parameters.
1223 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1226 Remove lots of duplicated code from the EVP library. For example *every*
1227 cipher init() function handles the 'iv' in the same way according to the
1228 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1229 for CFB and OFB modes they zero ctx->num.
1231 New functionality allows removal of S/MIME code RC2 hack.
1233 Most of the routines have the same form and so can be declared in terms
1236 By shifting this to the top level EVP_CipherInit() it can be removed from
1237 all individual ciphers. If the cipher wants to handle IVs or keys
1238 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1241 Change lots of functions like EVP_EncryptUpdate() to now return a
1242 value: although software versions of the algorithms cannot fail
1243 any installed hardware versions can.
1246 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1247 this option is set, tolerate broken clients that send the negotiated
1248 protocol version number instead of the requested protocol version
1252 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1253 i.e. non-zero for export ciphersuites, zero otherwise.
1254 Previous versions had this flag inverted, inconsistent with
1255 rsa_tmp_cb (..._TMP_RSA_CB).
1256 [Bodo Moeller; problem reported by Amit Chopra]
1258 *) Add missing DSA library text string. Work around for some IIS
1259 key files with invalid SEQUENCE encoding.
1262 *) Add a document (doc/standards.txt) that list all kinds of standards
1263 and so on that are implemented in OpenSSL.
1266 *) Enhance c_rehash script. Old version would mishandle certificates
1267 with the same subject name hash and wouldn't handle CRLs at all.
1268 Added -fingerprint option to crl utility, to support new c_rehash
1272 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1275 *) Fix for SSL server purpose checking. Server checking was
1276 rejecting certificates which had extended key usage present
1277 but no ssl client purpose.
1278 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1280 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1281 is a little unclear about how a blank password is handled.
1282 Since the password in encoded as a BMPString with terminating
1283 double NULL a zero length password would end up as just the
1284 double NULL. However no password at all is different and is
1285 handled differently in the PKCS#12 key generation code. NS
1286 treats a blank password as zero length. MSIE treats it as no
1287 password on export: but it will try both on import. We now do
1288 the same: PKCS12_parse() tries zero length and no password if
1289 the password is set to "" or NULL (NULL is now a valid password:
1290 it wasn't before) as does the pkcs12 application.
1293 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1294 perror when PEM_read_bio_X509_REQ fails, the error message must
1295 be obtained from the error queue.
1298 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1299 it in ERR_remove_state if appropriate, and change ERR_get_state
1300 accordingly to avoid race conditions (this is necessary because
1301 thread_hash is no longer constant once set).
1304 *) Bugfix for linux-elf makefile.one.
1307 *) RSA_get_default_method() will now cause a default
1308 RSA_METHOD to be chosen if one doesn't exist already.
1309 Previously this was only set during a call to RSA_new()
1310 or RSA_new_method(NULL) meaning it was possible for
1311 RSA_get_default_method() to return NULL.
1314 *) Added native name translation to the existing DSO code
1315 that will convert (if the flag to do so is set) filenames
1316 that are sufficiently small and have no path information
1317 into a canonical native form. Eg. "blah" converted to
1318 "libblah.so" or "blah.dll" etc.
1321 *) New function ERR_error_string_n(e, buf, len) which is like
1322 ERR_error_string(e, buf), but writes at most 'len' bytes
1323 including the 0 terminator. For ERR_error_string_n, 'buf'
1325 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1327 *) CONF library reworked to become more general. A new CONF
1328 configuration file reader "class" is implemented as well as a
1329 new functions (NCONF_*, for "New CONF") to handle it. The now
1330 old CONF_* functions are still there, but are reimplemented to
1331 work in terms of the new functions. Also, a set of functions
1332 to handle the internal storage of the configuration data is
1333 provided to make it easier to write new configuration file
1334 reader "classes" (I can definitely see something reading a
1335 configuration file in XML format, for example), called _CONF_*,
1336 or "the configuration storage API"...
1338 The new configuration file reading functions are:
1340 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1341 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1343 NCONF_default, NCONF_WIN32
1345 NCONF_dump_fp, NCONF_dump_bio
1347 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1348 NCONF_new creates a new CONF object. This works in the same way
1349 as other interfaces in OpenSSL, like the BIO interface.
1350 NCONF_dump_* dump the internal storage of the configuration file,
1351 which is useful for debugging. All other functions take the same
1352 arguments as the old CONF_* functions wth the exception of the
1353 first that must be a `CONF *' instead of a `LHASH *'.
1355 To make it easer to use the new classes with the old CONF_* functions,
1356 the function CONF_set_default_method is provided.
1359 *) Add '-tls1' option to 'openssl ciphers', which was already
1360 mentioned in the documentation but had not been implemented.
1361 (This option is not yet really useful because even the additional
1362 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1365 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1366 OpenSSL-based applications) load shared libraries and bind to
1367 them in a portable way.
1368 [Geoff Thorpe, with contributions from Richard Levitte]
1370 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1372 *) Make sure _lrotl and _lrotr are only used with MSVC.
1374 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1375 (the default implementation of RAND_status).
1377 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1378 to '-clrext' (= clear extensions), as intended and documented.
1379 [Bodo Moeller; inconsistency pointed out by Michael Attili
1382 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1383 was larger than the MD block size.
1384 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1386 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1387 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1388 using the passed key: if the passed key was a private key the result
1389 of X509_print(), for example, would be to print out all the private key
1393 *) des_quad_cksum() byte order bug fix.
1394 [Ulf Möller, using the problem description in krb4-0.9.7, where
1395 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1397 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1399 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1401 *) For easily testing in shell scripts whether some command
1402 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1403 returns with exit code 0 iff no command of the given name is available.
1404 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1405 the output goes to stdout and nothing is printed to stderr.
1406 Additional arguments are always ignored.
1408 Since for each cipher there is a command of the same name,
1409 the 'no-cipher' compilation switches can be tested this way.
1411 ('openssl no-XXX' is not able to detect pseudo-commands such
1412 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1415 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1418 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1419 is set; it will be thrown away anyway because each handshake creates
1421 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1422 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1423 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1424 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1427 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1428 'Q' and 'R' lose their special meanings (quit/renegotiate).
1429 This is part of what -quiet does; unlike -quiet, -ign_eof
1430 does not suppress any output.
1433 *) Add compatibility options to the purpose and trust code. The
1434 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1435 accepts a certificate or CA, this was the previous behaviour,
1436 with all the associated security issues.
1438 X509_TRUST_COMPAT is the old trust behaviour: only and
1439 automatically trust self signed roots in certificate store. A
1440 new trust setting X509_TRUST_DEFAULT is used to specify that
1441 a purpose has no associated trust setting and it should instead
1442 use the value in the default purpose.
1445 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1446 and fix a memory leak.
1449 *) In util/mkerr.pl (which implements 'make errors'), preserve
1450 reason strings from the previous version of the .c file, as
1451 the default to have only downcase letters (and digits) in
1452 automatically generated reasons codes is not always appropriate.
1455 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1456 using strerror. Previously, ERR_reason_error_string() returned
1457 library names as reason strings for SYSerr; but SYSerr is a special
1458 case where small numbers are errno values, not library numbers.
1461 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1462 converts DSA parameters into DH parameters. (When creating parameters,
1463 DSA_generate_parameters is used.)
1466 *) Include 'length' (recommended exponent length) in C code generated
1467 by 'openssl dhparam -C'.
1470 *) The second argument to set_label in perlasm was already being used
1471 so couldn't be used as a "file scope" flag. Moved to third argument
1475 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1476 instead of RAND_bytes for encryption IVs and salts.
1479 *) Include RAND_status() into RAND_METHOD instead of implementing
1480 it only for md_rand.c Otherwise replacing the PRNG by calling
1481 RAND_set_rand_method would be impossible.
1484 *) Don't let DSA_generate_key() enter an infinite loop if the random
1485 number generation fails.
1488 *) New 'rand' application for creating pseudo-random output.
1491 *) Added configuration support for Linux/IA64
1492 [Rolf Haberrecker <rolf@suse.de>]
1494 *) Assembler module support for Mingw32.
1497 *) Shared library support for HPUX (in shlib/).
1498 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1500 *) Shared library support for Solaris gcc.
1501 [Lutz Behnke <behnke@trustcenter.de>]
1503 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1505 *) PKCS7_encrypt() was adding text MIME headers twice because they
1506 were added manually and by SMIME_crlf_copy().
1509 *) In bntest.c don't call BN_rand with zero bits argument.
1510 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1512 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1513 case was implemented. This caused BN_div_recp() to fail occasionally.
1516 *) Add an optional second argument to the set_label() in the perl
1517 assembly language builder. If this argument exists and is set
1518 to 1 it signals that the assembler should use a symbol whose
1519 scope is the entire file, not just the current function. This
1520 is needed with MASM which uses the format label:: for this scope.
1521 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1523 *) Change the ASN1 types so they are typedefs by default. Before
1524 almost all types were #define'd to ASN1_STRING which was causing
1525 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1529 *) Change names of new functions to the new get1/get0 naming
1530 convention: After 'get1', the caller owns a reference count
1531 and has to call ..._free; 'get0' returns a pointer to some
1532 data structure without incrementing reference counters.
1533 (Some of the existing 'get' functions increment a reference
1534 counter, some don't.)
1535 Similarly, 'set1' and 'add1' functions increase reference
1536 counters or duplicate objects.
1539 *) Allow for the possibility of temp RSA key generation failure:
1540 the code used to assume it always worked and crashed on failure.
1543 *) Fix potential buffer overrun problem in BIO_printf().
1544 [Ulf Möller, using public domain code by Patrick Powell; problem
1545 pointed out by David Sacerdote <das33@cornell.edu>]
1547 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1548 RAND_egd() and RAND_status(). In the command line application,
1549 the EGD socket can be specified like a seed file using RANDFILE
1553 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1554 Some CAs (e.g. Verisign) distribute certificates in this form.
1557 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1558 list to exclude them. This means that no special compilation option
1559 is needed to use anonymous DH: it just needs to be included in the
1563 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1564 EVP_MD_type. The old functionality is available in a new macro called
1565 EVP_MD_md(). Change code that uses it and update docs.
1568 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1569 where the 'void *' argument is replaced by a function pointer argument.
1570 Previously 'void *' was abused to point to functions, which works on
1571 many platforms, but is not correct. As these functions are usually
1572 called by macros defined in OpenSSL header files, most source code
1573 should work without changes.
1576 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1577 sections with information on -D... compiler switches used for
1578 compiling the library so that applications can see them. To enable
1579 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1580 must be defined. E.g.,
1581 #define OPENSSL_ALGORITHM_DEFINES
1582 #include <openssl/opensslconf.h>
1583 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1584 [Richard Levitte, Ulf and Bodo Möller]
1586 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1590 *) Change the 'other' type in certificate aux info to a STACK_OF
1591 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1592 the required ASN1 format: arbitrary types determined by an OID.
1595 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1596 argument to 'req'. This is not because the function is newer or
1597 better than others it just uses the work 'NEW' in the certificate
1598 request header lines. Some software needs this.
1601 *) Reorganise password command line arguments: now passwords can be
1602 obtained from various sources. Delete the PEM_cb function and make
1603 it the default behaviour: i.e. if the callback is NULL and the
1604 usrdata argument is not NULL interpret it as a null terminated pass
1605 phrase. If usrdata and the callback are NULL then the pass phrase
1606 is prompted for as usual.
1609 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1610 the support is automatically enabled. The resulting binaries will
1611 autodetect the card and use it if present.
1612 [Ben Laurie and Compaq Inc.]
1614 *) Work around for Netscape hang bug. This sends certificate request
1615 and server done in one record. Since this is perfectly legal in the
1616 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1617 the bugs/SSLv3 entry for more info.
1620 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1623 *) Add -rand argument to smime and pkcs12 applications and read/write
1627 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1630 *) Add command line password options to the remaining applications.
1633 *) Bug fix for BN_div_recp() for numerators with an even number of
1637 *) More tests in bntest.c, and changed test_bn output.
1640 *) ./config recognizes MacOS X now.
1643 *) Bug fix for BN_div() when the first words of num and divsor are
1644 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1647 *) Add support for various broken PKCS#8 formats, and command line
1648 options to produce them.
1651 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1652 get temporary BIGNUMs from a BN_CTX.
1655 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1659 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1660 include a #define from the old name to the new. The original intent
1661 was that statically linked binaries could for example just call
1662 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1663 link with digests. This never worked becayse SSLeay_add_all_digests()
1664 and SSLeay_add_all_ciphers() were in the same source file so calling
1665 one would link with the other. They are now in separate source files.
1668 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1671 *) Use a less unusual form of the Miller-Rabin primality test (it used
1672 a binary algorithm for exponentiation integrated into the Miller-Rabin
1673 loop, our standard modexp algorithms are faster).
1676 *) Support for the EBCDIC character set completed.
1677 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1679 *) Source code cleanups: use const where appropriate, eliminate casts,
1680 use void * instead of char * in lhash.
1683 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1684 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1685 this the server could overwrite ephemeral keys that the client
1689 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1690 using 50 iterations of the Rabin-Miller test.
1692 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1693 iterations of the Rabin-Miller test as required by the appendix
1694 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1695 As BN_is_prime_fasttest includes trial division, DSA parameter
1696 generation becomes much faster.
1698 This implies a change for the callback functions in DSA_is_prime
1699 and DSA_generate_parameters: The callback function is called once
1700 for each positive witness in the Rabin-Miller test, not just
1701 occasionally in the inner loop; and the parameters to the
1702 callback function now provide an iteration count for the outer
1703 loop rather than for the current invocation of the inner loop.
1704 DSA_generate_parameters additionally can call the callback
1705 function with an 'iteration count' of -1, meaning that a
1706 candidate has passed the trial division test (when q is generated
1707 from an application-provided seed, trial division is skipped).
1710 *) New function BN_is_prime_fasttest that optionally does trial
1711 division before starting the Rabin-Miller test and has
1712 an additional BN_CTX * argument (whereas BN_is_prime always
1713 has to allocate at least one BN_CTX).
1714 'callback(1, -1, cb_arg)' is called when a number has passed the
1715 trial division stage.
1718 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1722 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1725 *) New function BN_pseudo_rand().
1728 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1729 bignum version of BN_from_montgomery() with the working code from
1730 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1734 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1735 made it impossible to use the same SSL_SESSION data structure in
1736 SSL2 clients in multiple threads.
1739 *) The return value of RAND_load_file() no longer counts bytes obtained
1740 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1741 to seed the PRNG (previously an explicit byte count was required).
1742 [Ulf Möller, Bodo Möller]
1744 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1745 used (char *) instead of (void *) and had casts all over the place.
1748 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1751 *) Retain source code compatibility for BN_prime_checks macro:
1752 BN_is_prime(..., BN_prime_checks, ...) now uses
1753 BN_prime_checks_for_size to determine the appropriate number of
1754 Rabin-Miller iterations.
1757 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1758 DH_CHECK_P_NOT_SAFE_PRIME.
1759 (Check if this is true? OpenPGP calls them "strong".)
1762 *) Merge the functionality of "dh" and "gendh" programs into a new program
1763 "dhparam". The old programs are retained for now but will handle DH keys
1764 (instead of parameters) in future.
1767 *) Make the ciphers, s_server and s_client programs check the return values
1768 when a new cipher list is set.
1771 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1772 ciphers. Before when the 56bit ciphers were enabled the sorting was
1775 The syntax for the cipher sorting has been extended to support sorting by
1776 cipher-strength (using the strength_bits hard coded in the tables).
1777 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1779 Fix a bug in the cipher-command parser: when supplying a cipher command
1780 string with an "undefined" symbol (neither command nor alphanumeric
1781 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1782 an error is flagged.
1784 Due to the strength-sorting extension, the code of the
1785 ssl_create_cipher_list() function was completely rearranged. I hope that
1786 the readability was also increased :-)
1787 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1789 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1790 for the first serial number and places 2 in the serial number file. This
1791 avoids problems when the root CA is created with serial number zero and
1792 the first user certificate has the same issuer name and serial number
1796 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1797 the new code. Add documentation for this stuff.
1800 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1801 X509_*() to X509at_*() on the grounds that they don't handle X509
1802 structures and behave in an analagous way to the X509v3 functions:
1803 they shouldn't be called directly but wrapper functions should be used
1806 So we also now have some wrapper functions that call the X509at functions
1807 when passed certificate requests. (TO DO: similar things can be done with
1808 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1809 things. Some of these need some d2i or i2d and print functionality
1810 because they handle more complex structures.)
1813 *) Add missing #ifndefs that caused missing symbols when building libssl
1814 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1815 NO_RSA in ssl/s2*.c.
1816 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1818 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1819 has a return value which indicates the quality of the random data
1820 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1821 error queue. New function RAND_pseudo_bytes() generates output that is
1822 guaranteed to be unique but not unpredictable. RAND_add is like
1823 RAND_seed, but takes an extra argument for an entropy estimate
1824 (RAND_seed always assumes full entropy).
1827 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1828 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1829 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1830 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1831 false-positive rate of at most 2^-80 for random input.
1834 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1837 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1838 in the 0.9.5 release), this returns the chain
1839 from an X509_CTX structure with a dup of the stack and all
1840 the X509 reference counts upped: so the stack will exist
1841 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1844 Also make SSL_SESSION_print() print out the verify return
1848 *) Add manpage for the pkcs12 command. Also change the default
1849 behaviour so MAC iteration counts are used unless the new
1850 -nomaciter option is used. This improves file security and
1851 only older versions of MSIE (4.0 for example) need it.
1854 *) Honor the no-xxx Configure options when creating .DEF files.
1857 *) Add PKCS#10 attributes to field table: challengePassword,
1858 unstructuredName and unstructuredAddress. These are taken from
1859 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1860 international characters are used.
1862 More changes to X509_ATTRIBUTE code: allow the setting of types
1863 based on strings. Remove the 'loc' parameter when adding
1864 attributes because these will be a SET OF encoding which is sorted
1868 *) Initial changes to the 'req' utility to allow request generation
1869 automation. This will allow an application to just generate a template
1870 file containing all the field values and have req construct the
1873 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1874 used all over the place including certificate requests and PKCS#7
1875 structures. They are currently handled manually where necessary with
1876 some primitive wrappers for PKCS#7. The new functions behave in a
1877 manner analogous to the X509 extension functions: they allow
1878 attributes to be looked up by NID and added.
1880 Later something similar to the X509V3 code would be desirable to
1881 automatically handle the encoding, decoding and printing of the
1882 more complex types. The string types like challengePassword can
1883 be handled by the string table functions.
1885 Also modified the multi byte string table handling. Now there is
1886 a 'global mask' which masks out certain types. The table itself
1887 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1888 is useful when for example there is only one permissible type
1889 (as in countryName) and using the mask might result in no valid
1893 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1894 SSL_get_peer_finished to allow applications to obtain the latest
1895 Finished messages sent to the peer or expected from the peer,
1896 respectively. (SSL_get_peer_finished is usually the Finished message
1897 actually received from the peer, otherwise the protocol will be aborted.)
1899 As the Finished message are message digests of the complete handshake
1900 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1901 be used for external authentication procedures when the authentication
1902 provided by SSL/TLS is not desired or is not enough.
1905 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1906 the host supports BWX extension and if Compaq C is present on the
1907 $PATH. Just exploiting of the BWX extension results in 20-30%
1908 performance kick for some algorithms, e.g. DES and RC4 to mention
1909 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1913 *) Add support for MS "fast SGC". This is arguably a violation of the
1914 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1915 weak crypto and after checking the certificate is SGC a second one
1916 with strong crypto. MS SGC stops the first handshake after receiving
1917 the server certificate message and sends a second client hello. Since
1918 a server will typically do all the time consuming operations before
1919 expecting any further messages from the client (server key exchange
1920 is the most expensive) there is little difference between the two.
1922 To get OpenSSL to support MS SGC we have to permit a second client
1923 hello message after we have sent server done. In addition we have to
1924 reset the MAC if we do get this second client hello.
1927 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1928 if a DER encoded private key is RSA or DSA traditional format. Changed
1929 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1930 format DER encoded private key. Newer code should use PKCS#8 format which
1931 has the key type encoded in the ASN1 structure. Added DER private key
1932 support to pkcs8 application.
1935 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1936 ciphersuites has been selected (as required by the SSL 3/TLS 1
1937 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1938 is set, we interpret this as a request to violate the specification
1939 (the worst that can happen is a handshake failure, and 'correct'
1940 behaviour would result in a handshake failure anyway).
1943 *) In SSL_CTX_add_session, take into account that there might be multiple
1944 SSL_SESSION structures with the same session ID (e.g. when two threads
1945 concurrently obtain them from an external cache).
1946 The internal cache can handle only one SSL_SESSION with a given ID,
1947 so if there's a conflict, we now throw out the old one to achieve
1951 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1952 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1953 some routines that use cipher OIDs: some ciphers do not have OIDs
1954 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1958 *) Simplify the trust setting structure and code. Now we just have
1959 two sequences of OIDs for trusted and rejected settings. These will
1960 typically have values the same as the extended key usage extension
1961 and any application specific purposes.
1963 The trust checking code now has a default behaviour: it will just
1964 check for an object with the same NID as the passed id. Functions can
1965 be provided to override either the default behaviour or the behaviour
1966 for a given id. SSL client, server and email already have functions
1967 in place for compatibility: they check the NID and also return "trusted"
1968 if the certificate is self signed.
1971 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1972 traditional format into an EVP_PKEY structure.
1975 *) Add a password callback function PEM_cb() which either prompts for
1976 a password if usr_data is NULL or otherwise assumes it is a null
1977 terminated password. Allow passwords to be passed on command line
1978 environment or config files in a few more utilities.
1981 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1982 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1983 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1984 Update documentation.
1987 *) Support for ASN1 "NULL" type. This could be handled before by using
1988 ASN1_TYPE but there wasn't any function that would try to read a NULL
1989 and produce an error if it couldn't. For compatibility we also have
1990 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1991 don't allocate anything because they don't need to.
1994 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1996 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1998 *) Rebuild of the memory allocation routines used by OpenSSL code and
1999 possibly others as well. The purpose is to make an interface that
2000 provide hooks so anyone can build a separate set of allocation and
2001 deallocation routines to be used by OpenSSL, for example memory
2002 pool implementations, or something else, which was previously hard
2003 since Malloc(), Realloc() and Free() were defined as macros having
2004 the values malloc, realloc and free, respectively (except for Win32
2005 compilations). The same is provided for memory debugging code.
2006 OpenSSL already comes with functionality to find memory leaks, but
2007 this gives people a chance to debug other memory problems.
2009 With these changes, a new set of functions and macros have appeared:
2011 CRYPTO_set_mem_debug_functions() [F]
2012 CRYPTO_get_mem_debug_functions() [F]
2013 CRYPTO_dbg_set_options() [F]
2014 CRYPTO_dbg_get_options() [F]
2015 CRYPTO_malloc_debug_init() [M]
2017 The memory debug functions are NULL by default, unless the library
2018 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2019 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2020 gives the standard debugging functions that come with OpenSSL) or
2021 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2022 provided by the library user) must be used. When the standard
2023 debugging functions are used, CRYPTO_dbg_set_options can be used to
2024 request additional information:
2025 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2026 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2028 Also, things like CRYPTO_set_mem_functions will always give the
2029 expected result (the new set of functions is used for allocation
2030 and deallocation) at all times, regardless of platform and compiler
2033 To finish it up, some functions that were never use in any other
2034 way than through macros have a new API and new semantic:
2037 CRYPTO_dbg_realloc()
2040 All macros of value have retained their old syntax.
2041 [Richard Levitte and Bodo Moeller]
2043 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2044 ordering of SMIMECapabilities wasn't in "strength order" and there
2045 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2049 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2050 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2051 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2053 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2054 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2055 functionality to handle multipart/signed properly) and a utility
2056 called 'smime' to call all this stuff. This is based on code I
2057 originally wrote for Celo who have kindly allowed it to be
2058 included in OpenSSL.
2061 *) Add variants des_set_key_checked and des_set_key_unchecked of
2062 des_set_key (aka des_key_sched). Global variable des_check_key
2063 decides which of these is called by des_set_key; this way
2064 des_check_key behaves as it always did, but applications and
2065 the library itself, which was buggy for des_check_key == 1,
2066 have a cleaner way to pick the version they need.
2069 *) New function PKCS12_newpass() which changes the password of a
2073 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2074 dynamic mix. In both cases the ids can be used as an index into the
2075 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2076 functions so they accept a list of the field values and the
2077 application doesn't need to directly manipulate the X509_TRUST
2081 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2085 *) Modify the way the V3 extension code looks up extensions. This now
2086 works in a similar way to the object code: we have some "standard"
2087 extensions in a static table which is searched with OBJ_bsearch()
2088 and the application can add dynamic ones if needed. The file
2089 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2090 updated whenever a new extension is added to the core code and kept
2091 in ext_nid order. There is a simple program 'tabtest.c' which checks
2092 this. New extensions are not added too often so this file can readily
2093 be maintained manually.
2095 There are two big advantages in doing things this way. The extensions
2096 can be looked up immediately and no longer need to be "added" using
2097 X509V3_add_standard_extensions(): this function now does nothing.
2098 [Side note: I get *lots* of email saying the extension code doesn't
2099 work because people forget to call this function]
2100 Also no dynamic allocation is done unless new extensions are added:
2101 so if we don't add custom extensions there is no need to call
2102 X509V3_EXT_cleanup().
2105 *) Modify enc utility's salting as follows: make salting the default. Add a
2106 magic header, so unsalted files fail gracefully instead of just decrypting
2107 to garbage. This is because not salting is a big security hole, so people
2108 should be discouraged from doing it.
2111 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2112 digest to be passed on the command line but it only used this
2113 parameter when signing a certificate. Modified so all relevant
2114 operations are affected by the digest parameter including the
2115 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2116 DSA key was used because it didn't fix the digest.
2119 *) Initial certificate chain verify code. Currently tests the untrusted
2120 certificates for consistency with the verify purpose (which is set
2121 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2123 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2124 this is because it will reject chains with invalid extensions whereas
2125 every previous version of OpenSSL and SSLeay made no checks at all.
2127 Trust code: checks the root CA for the relevant trust settings. Trust
2128 settings have an initial value consistent with the verify purpose: e.g.
2129 if the verify purpose is for SSL client use it expects the CA to be
2130 trusted for SSL client use. However the default value can be changed to
2131 permit custom trust settings: one example of this would be to only trust
2132 certificates from a specific "secure" set of CAs.
2134 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2135 which should be used for version portability: especially since the
2136 verify structure is likely to change more often now.
2138 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2139 to set them. If not set then assume SSL clients will verify SSL servers
2142 Two new options to the verify program: -untrusted allows a set of
2143 untrusted certificates to be passed in and -purpose which sets the
2144 intended purpose of the certificate. If a purpose is set then the
2145 new chain verify code is used to check extension consistency.
2148 *) Support for the authority information access extension.
2151 *) Modify RSA and DSA PEM read routines to transparently handle
2152 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2153 public keys in a format compatible with certificate
2154 SubjectPublicKeyInfo structures. Unfortunately there were already
2155 functions called *_PublicKey_* which used various odd formats so
2156 these are retained for compatibility: however the DSA variants were
2157 never in a public release so they have been deleted. Changed dsa/rsa
2158 utilities to handle the new format: note no releases ever handled public
2159 keys so we should be OK.
2161 The primary motivation for this change is to avoid the same fiasco
2162 that dogs private keys: there are several incompatible private key
2163 formats some of which are standard and some OpenSSL specific and
2164 require various evil hacks to allow partial transparent handling and
2165 even then it doesn't work with DER formats. Given the option anything
2166 other than PKCS#8 should be dumped: but the other formats have to
2167 stay in the name of compatibility.
2169 With public keys and the benefit of hindsight one standard format
2170 is used which works with EVP_PKEY, RSA or DSA structures: though
2171 it clearly returns an error if you try to read the wrong kind of key.
2173 Added a -pubkey option to the 'x509' utility to output the public key.
2174 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2175 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2176 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2177 that do the same as the EVP_PKEY_assign_*() except they up the
2178 reference count of the added key (they don't "swallow" the
2182 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2183 CRLs would fail if the file contained no certificates or no CRLs:
2184 added a new function to read in both types and return the number
2185 read: this means that if none are read it will be an error. The
2186 DER versions of the certificate and CRL reader would always fail
2187 because it isn't possible to mix certificates and CRLs in DER format
2188 without choking one or the other routine. Changed this to just read
2189 a certificate: this is the best we can do. Also modified the code
2190 in apps/verify.c to take notice of return codes: it was previously
2191 attempting to read in certificates from NULL pointers and ignoring
2192 any errors: this is one reason why the cert and CRL reader seemed
2193 to work. It doesn't check return codes from the default certificate
2194 routines: these may well fail if the certificates aren't installed.
2197 *) Code to support otherName option in GeneralName.
2200 *) First update to verify code. Change the verify utility
2201 so it warns if it is passed a self signed certificate:
2202 for consistency with the normal behaviour. X509_verify
2203 has been modified to it will now verify a self signed
2204 certificate if *exactly* the same certificate appears
2205 in the store: it was previously impossible to trust a
2206 single self signed certificate. This means that:
2207 openssl verify ss.pem
2208 now gives a warning about a self signed certificate but
2209 openssl verify -CAfile ss.pem ss.pem
2213 *) For servers, store verify_result in SSL_SESSION data structure
2214 (and add it to external session representation).
2215 This is needed when client certificate verifications fails,
2216 but an application-provided verification callback (set by
2217 SSL_CTX_set_cert_verify_callback) allows accepting the session
2218 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2219 but returns 1): When the session is reused, we have to set
2220 ssl->verify_result to the appropriate error code to avoid
2222 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2224 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2225 case in PKCS7_dataInit() where the signed PKCS7 structure
2226 didn't contain any existing data because it was being created.
2227 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2229 *) Add a salt to the key derivation routines in enc.c. This
2230 forms the first 8 bytes of the encrypted file. Also add a
2231 -S option to allow a salt to be input on the command line.
2234 *) New function X509_cmp(). Oddly enough there wasn't a function
2235 to compare two certificates. We do this by working out the SHA1
2236 hash and comparing that. X509_cmp() will be needed by the trust
2240 *) SSL_get1_session() is like SSL_get_session(), but increments
2241 the reference count in the SSL_SESSION returned.
2242 [Geoff Thorpe <geoff@eu.c2.net>]
2244 *) Fix for 'req': it was adding a null to request attributes.
2245 Also change the X509_LOOKUP and X509_INFO code to handle
2246 certificate auxiliary information.
2249 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2253 *) Add the possibility to add extra information to the memory leak
2254 detecting output, to form tracebacks, showing from where each
2255 allocation was originated: CRYPTO_push_info("constant string") adds
2256 the string plus current file name and line number to a per-thread
2257 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2258 is like calling CYRPTO_pop_info() until the stack is empty.
2259 Also updated memory leak detection code to be multi-thread-safe.
2262 *) Add options -text and -noout to pkcs7 utility and delete the
2263 encryption options which never did anything. Update docs.
2266 *) Add options to some of the utilities to allow the pass phrase
2267 to be included on either the command line (not recommended on
2268 OSes like Unix) or read from the environment. Update the
2269 manpages and fix a few bugs.
2272 *) Add a few manpages for some of the openssl commands.
2275 *) Fix the -revoke option in ca. It was freeing up memory twice,
2276 leaking and not finding already revoked certificates.
2279 *) Extensive changes to support certificate auxiliary information.
2280 This involves the use of X509_CERT_AUX structure and X509_AUX
2281 functions. An X509_AUX function such as PEM_read_X509_AUX()
2282 can still read in a certificate file in the usual way but it
2283 will also read in any additional "auxiliary information". By
2284 doing things this way a fair degree of compatibility can be
2285 retained: existing certificates can have this information added
2286 using the new 'x509' options.
2288 Current auxiliary information includes an "alias" and some trust
2289 settings. The trust settings will ultimately be used in enhanced
2290 certificate chain verification routines: currently a certificate
2291 can only be trusted if it is self signed and then it is trusted
2295 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2296 The problem was that one of the replacement routines had not been working
2297 since SSLeay releases. For now the offending routine has been replaced
2298 with non-optimised assembler. Even so, this now gives around 95%
2299 performance improvement for 1024 bit RSA signs.
2302 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2303 handling. Most clients have the effective key size in bits equal to
2304 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2305 A few however don't do this and instead use the size of the decrypted key
2306 to determine the RC2 key length and the AlgorithmIdentifier to determine
2307 the effective key length. In this case the effective key length can still
2308 be 40 bits but the key length can be 168 bits for example. This is fixed
2309 by manually forcing an RC2 key into the EVP_PKEY structure because the
2310 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2311 the key length and effective key length are equal.
2314 *) Add a bunch of functions that should simplify the creation of
2315 X509_NAME structures. Now you should be able to do:
2316 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2317 and have it automatically work out the correct field type and fill in
2318 the structures. The more adventurous can try:
2319 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2320 and it will (hopefully) work out the correct multibyte encoding.
2323 *) Change the 'req' utility to use the new field handling and multibyte
2324 copy routines. Before the DN field creation was handled in an ad hoc
2325 way in req, ca, and x509 which was rather broken and didn't support
2326 BMPStrings or UTF8Strings. Since some software doesn't implement
2327 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2328 using the dirstring_type option. See the new comment in the default
2329 openssl.cnf for more info.
2332 *) Make crypto/rand/md_rand.c more robust:
2333 - Assure unique random numbers after fork().
2334 - Make sure that concurrent threads access the global counter and
2335 md serializably so that we never lose entropy in them
2336 or use exactly the same state in multiple threads.
2337 Access to the large state is not always serializable because
2338 the additional locking could be a performance killer, and
2339 md should be large enough anyway.
2342 *) New file apps/app_rand.c with commonly needed functionality
2343 for handling the random seed file.
2345 Use the random seed file in some applications that previously did not:
2347 dsaparam -genkey (which also ignored its '-rand' option),
2350 x509 (when signing).
2351 Except on systems with /dev/urandom, it is crucial to have a random
2352 seed file at least for key creation, DSA signing, and for DH exchanges;
2353 for RSA signatures we could do without one.
2355 gendh and gendsa (unlike genrsa) used to read only the first byte
2356 of each file listed in the '-rand' option. The function as previously
2357 found in genrsa is now in app_rand.c and is used by all programs
2358 that support '-rand'.
2361 *) In RAND_write_file, use mode 0600 for creating files;
2362 don't just chmod when it may be too late.
2365 *) Report an error from X509_STORE_load_locations
2366 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2369 *) New function ASN1_mbstring_copy() this copies a string in either
2370 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2371 into an ASN1_STRING type. A mask of permissible types is passed
2372 and it chooses the "minimal" type to use or an error if not type
2376 *) Add function equivalents to the various macros in asn1.h. The old
2377 macros are retained with an M_ prefix. Code inside the library can
2378 use the M_ macros. External code (including the openssl utility)
2379 should *NOT* in order to be "shared library friendly".
2382 *) Add various functions that can check a certificate's extensions
2383 to see if it usable for various purposes such as SSL client,
2384 server or S/MIME and CAs of these types. This is currently
2385 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2386 verification. Also added a -purpose flag to x509 utility to
2387 print out all the purposes.
2390 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2394 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2395 for, obtain and decode and extension and obtain its critical flag.
2396 This allows all the necessary extension code to be handled in a
2397 single function call.
2400 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2401 platforms. See crypto/rc4/rc4_enc.c for further details.
2404 *) New -noout option to asn1parse. This causes no output to be produced
2405 its main use is when combined with -strparse and -out to extract data
2406 from a file (which may not be in ASN.1 format).
2409 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2410 when producing the local key id.
2411 [Richard Levitte <levitte@stacken.kth.se>]
2413 *) New option -dhparam in s_server. This allows a DH parameter file to be
2414 stated explicitly. If it is not stated then it tries the first server
2415 certificate file. The previous behaviour hard coded the filename
2419 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2420 a public key to be input or output. For example:
2421 openssl rsa -in key.pem -pubout -out pubkey.pem
2422 Also added necessary DSA public key functions to handle this.
2425 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2426 in the message. This was handled by allowing
2427 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2428 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2430 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2431 to the end of the strings whereas this didn't. This would cause problems
2432 if strings read with d2i_ASN1_bytes() were later modified.
2433 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2435 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2436 data and it contains EOF it will end up returning an error. This is
2437 caused by input 46 bytes long. The cause is due to the way base64
2438 BIOs find the start of base64 encoded data. They do this by trying a
2439 trial decode on each line until they find one that works. When they
2440 do a flag is set and it starts again knowing it can pass all the
2441 data directly through the decoder. Unfortunately it doesn't reset
2442 the context it uses. This means that if EOF is reached an attempt
2443 is made to pass two EOFs through the context and this causes the
2444 resulting error. This can also cause other problems as well. As is
2445 usual with these problems it takes *ages* to find and the fix is
2446 trivial: move one line.
2447 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2449 *) Ugly workaround to get s_client and s_server working under Windows. The
2450 old code wouldn't work because it needed to select() on sockets and the
2451 tty (for keypresses and to see if data could be written). Win32 only
2452 supports select() on sockets so we select() with a 1s timeout on the
2453 sockets and then see if any characters are waiting to be read, if none
2454 are present then we retry, we also assume we can always write data to
2455 the tty. This isn't nice because the code then blocks until we've
2456 received a complete line of data and it is effectively polling the
2457 keyboard at 1s intervals: however it's quite a bit better than not
2458 working at all :-) A dedicated Windows application might handle this
2459 with an event loop for example.
2462 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2463 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2464 will be called when RSA_sign() and RSA_verify() are used. This is useful
2465 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2466 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2467 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2468 This necessitated the support of an extra signature type NID_md5_sha1
2469 for SSL signatures and modifications to the SSL library to use it instead
2470 of calling RSA_public_decrypt() and RSA_private_encrypt().
2473 *) Add new -verify -CAfile and -CApath options to the crl program, these
2474 will lookup a CRL issuers certificate and verify the signature in a
2475 similar way to the verify program. Tidy up the crl program so it
2476 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2477 less strict. It will now permit CRL extensions even if it is not
2478 a V2 CRL: this will allow it to tolerate some broken CRLs.
2481 *) Initialize all non-automatic variables each time one of the openssl
2482 sub-programs is started (this is necessary as they may be started
2483 multiple times from the "OpenSSL>" prompt).
2484 [Lennart Bang, Bodo Moeller]
2486 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2487 removing all other RSA functionality (this is what NO_RSA does). This
2488 is so (for example) those in the US can disable those operations covered
2489 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2493 *) Non-copying interface to BIO pairs.
2494 (still largely untested)
2497 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2498 ASCII string. This was handled independently in various places before.
2501 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2502 UTF8 strings a character at a time.
2505 *) Use client_version from client hello to select the protocol
2506 (s23_srvr.c) and for RSA client key exchange verification
2507 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2510 *) Add various utility functions to handle SPKACs, these were previously
2511 handled by poking round in the structure internals. Added new function
2512 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2513 print, verify and generate SPKACs. Based on an original idea from
2514 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2517 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2520 *) Allow the config file extension section to be overwritten on the
2521 command line. Based on an original idea from Massimiliano Pala
2522 <madwolf@comune.modena.it>. The new option is called -extensions
2523 and can be applied to ca, req and x509. Also -reqexts to override
2524 the request extensions in req and -crlexts to override the crl extensions
2528 *) Add new feature to the SPKAC handling in ca. Now you can include
2529 the same field multiple times by preceding it by "XXXX." for example:
2532 this is the same syntax as used in the req config file.
2535 *) Allow certificate extensions to be added to certificate requests. These
2536 are specified in a 'req_extensions' option of the req section of the
2537 config file. They can be printed out with the -text option to req but
2538 are otherwise ignored at present.
2541 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2542 data read consists of only the final block it would not decrypted because
2543 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2544 A misplaced 'break' also meant the decrypted final block might not be
2545 copied until the next read.
2548 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2549 a few extra parameters to the DH structure: these will be useful if
2550 for example we want the value of 'q' or implement X9.42 DH.
2553 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2554 provides hooks that allow the default DSA functions or functions on a
2555 "per key" basis to be replaced. This allows hardware acceleration and
2556 hardware key storage to be handled without major modification to the
2557 library. Also added low level modexp hooks and CRYPTO_EX structure and
2558 associated functions.
2561 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2562 as "read only": it can't be written to and the buffer it points to will
2563 not be freed. Reading from a read only BIO is much more efficient than
2564 a normal memory BIO. This was added because there are several times when
2565 an area of memory needs to be read from a BIO. The previous method was
2566 to create a memory BIO and write the data to it, this results in two
2567 copies of the data and an O(n^2) reading algorithm. There is a new
2568 function BIO_new_mem_buf() which creates a read only memory BIO from
2569 an area of memory. Also modified the PKCS#7 routines to use read only
2573 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2574 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2575 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2576 but a retry condition occured while trying to read the rest.
2579 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2580 NID_pkcs7_encrypted by default: this was wrong since this should almost
2581 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2582 the encrypted data type: this is a more sensible place to put it and it
2583 allows the PKCS#12 code to be tidied up that duplicated this
2587 *) Changed obj_dat.pl script so it takes its input and output files on
2588 the command line. This should avoid shell escape redirection problems
2592 *) Initial support for certificate extension requests, these are included
2593 in things like Xenroll certificate requests. Included functions to allow
2594 extensions to be obtained and added.
2597 *) -crlf option to s_client and s_server for sending newlines as
2598 CRLF (as required by many protocols).
2601 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2603 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2604 [Ralf S. Engelschall]
2606 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2607 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2609 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2613 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2614 DH parameters/keys (q is lost during that conversion, but the resulting
2615 DH parameters contain its length).
2617 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2618 much faster than DH_generate_parameters (which creates parameters
2619 where p = 2*q + 1), and also the smaller q makes DH computations
2620 much more efficient (160-bit exponentiation instead of 1024-bit
2621 exponentiation); so this provides a convenient way to support DHE
2622 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2623 utter importance to use
2624 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2626 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2627 when such DH parameters are used, because otherwise small subgroup
2628 attacks may become possible!
2631 *) Avoid memory leak in i2d_DHparams.
2634 *) Allow the -k option to be used more than once in the enc program:
2635 this allows the same encrypted message to be read by multiple recipients.
2638 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2639 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2640 it will always use the numerical form of the OID, even if it has a short
2644 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2645 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2646 otherwise bn_mod_exp was called. In the case of hardware keys for example
2647 no private key components need be present and it might store extra data
2648 in the RSA structure, which cannot be accessed from bn_mod_exp.
2649 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2650 private key operations.
2653 *) Added support for SPARC Linux.
2656 *) pem_password_cb function type incompatibly changed from
2657 typedef int pem_password_cb(char *buf, int size, int rwflag);
2659 ....(char *buf, int size, int rwflag, void *userdata);
2660 so that applications can pass data to their callbacks:
2661 The PEM[_ASN1]_{read,write}... functions and macros now take an
2662 additional void * argument, which is just handed through whenever
2663 the password callback is called.
2664 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2666 New function SSL_CTX_set_default_passwd_cb_userdata.
2668 Compatibility note: As many C implementations push function arguments
2669 onto the stack in reverse order, the new library version is likely to
2670 interoperate with programs that have been compiled with the old
2671 pem_password_cb definition (PEM_whatever takes some data that
2672 happens to be on the stack as its last argument, and the callback
2673 just ignores this garbage); but there is no guarantee whatsoever that
2676 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2677 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2678 problems not only on Windows, but also on some Unix platforms.
2679 To avoid problematic command lines, these definitions are now in an
2680 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2681 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2684 *) MIPS III/IV assembler module is reimplemented.
2687 *) More DES library cleanups: remove references to srand/rand and
2688 delete an unused file.
2691 *) Add support for the the free Netwide assembler (NASM) under Win32,
2692 since not many people have MASM (ml) and it can be hard to obtain.
2693 This is currently experimental but it seems to work OK and pass all
2694 the tests. Check out INSTALL.W32 for info.
2697 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2698 without temporary keys kept an extra copy of the server key,
2699 and connections with temporary keys did not free everything in case
2703 *) New function RSA_check_key and new openssl rsa option -check
2704 for verifying the consistency of RSA keys.
2705 [Ulf Moeller, Bodo Moeller]
2707 *) Various changes to make Win32 compile work:
2708 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2709 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2710 comparison" warnings.
2711 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2714 *) Add a debugging option to PKCS#5 v2 key generation function: when
2715 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2716 derived keys are printed to stderr.
2719 *) Copy the flags in ASN1_STRING_dup().
2720 [Roman E. Pavlov <pre@mo.msk.ru>]
2722 *) The x509 application mishandled signing requests containing DSA
2723 keys when the signing key was also DSA and the parameters didn't match.
2725 It was supposed to omit the parameters when they matched the signing key:
2726 the verifying software was then supposed to automatically use the CA's
2727 parameters if they were absent from the end user certificate.
2729 Omitting parameters is no longer recommended. The test was also
2730 the wrong way round! This was probably due to unusual behaviour in
2731 EVP_cmp_parameters() which returns 1 if the parameters match.
2732 This meant that parameters were omitted when they *didn't* match and
2733 the certificate was useless. Certificates signed with 'ca' didn't have
2735 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2737 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2738 The interface is as follows:
2739 Applications can use
2740 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2741 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2742 "off" is now the default.
2743 The library internally uses
2744 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2745 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2746 to disable memory-checking temporarily.
2748 Some inconsistent states that previously were possible (and were
2749 even the default) are now avoided.
2751 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2752 with each memory chunk allocated; this is occasionally more helpful
2753 than just having a counter.
2755 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2757 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2761 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2762 which largely parallels "options", but is for changing API behaviour,
2763 whereas "options" are about protocol behaviour.
2764 Initial "mode" flags are:
2766 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2767 a single record has been written.
2768 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2769 retries use the same buffer location.
2770 (But all of the contents must be
2774 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2777 *) Fix problems with no-hmac etc.
2778 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2780 *) New functions RSA_get_default_method(), RSA_set_method() and
2781 RSA_get_method(). These allows replacement of RSA_METHODs without having
2782 to mess around with the internals of an RSA structure.
2785 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2786 Also really enable memory leak checks in openssl.c and in some
2788 [Chad C. Mulligan, Bodo Moeller]
2790 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2791 up the length of negative integers. This has now been simplified to just
2792 store the length when it is first determined and use it later, rather
2793 than trying to keep track of where data is copied and updating it to
2795 [Steve Henson, reported by Brien Wheeler
2796 <bwheeler@authentica-security.com>]
2798 *) Add a new function PKCS7_signatureVerify. This allows the verification
2799 of a PKCS#7 signature but with the signing certificate passed to the
2800 function itself. This contrasts with PKCS7_dataVerify which assumes the
2801 certificate is present in the PKCS#7 structure. This isn't always the
2802 case: certificates can be omitted from a PKCS#7 structure and be
2803 distributed by "out of band" means (such as a certificate database).
2806 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2807 function prototypes in pem.h, also change util/mkdef.pl to add the
2808 necessary function names.
2811 *) mk1mf.pl (used by Windows builds) did not properly read the
2812 options set by Configure in the top level Makefile, and Configure
2813 was not even able to write more than one option correctly.
2814 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2817 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2818 file to be loaded from a BIO or FILE pointer. The BIO version will
2819 for example allow memory BIOs to contain config info.
2822 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2823 Whoever hopes to achieve shared-library compatibility across versions
2824 must use this, not the compile-time macro.
2825 (Exercise 0.9.4: Which is the minimum library version required by
2827 Note: All this applies only to multi-threaded programs, others don't
2831 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2832 through a BIO pair triggered the default case, i.e.
2833 SSLerr(...,SSL_R_UNKNOWN_STATE).
2836 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2837 can use the SSL library even if none of the specific BIOs is
2841 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2842 for the encoded length.
2843 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2845 *) Add initial documentation of the X509V3 functions.
2848 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2849 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2850 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2851 secure PKCS#8 private key format with a high iteration count.
2854 *) Fix determination of Perl interpreter: A perl or perl5
2855 _directory_ in $PATH was also accepted as the interpreter.
2856 [Ralf S. Engelschall]
2858 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2859 wrong with it but it was very old and did things like calling
2860 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2864 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2865 to use the new extension code.
2868 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2869 with macros. This should make it easier to change their form, add extra
2870 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2874 *) Add to configuration table a new entry that can specify an alternative
2875 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2876 according to Mark Crispin <MRC@Panda.COM>.
2880 *) DES CBC did not update the IV. Weird.
2883 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2884 Changing the behaviour of the former might break existing programs --
2885 where IV updating is needed, des_ncbc_encrypt can be used.
2888 *) When bntest is run from "make test" it drives bc to check its
2889 calculations, as well as internally checking them. If an internal check
2890 fails, it needs to cause bc to give a non-zero result or make test carries
2891 on without noticing the failure. Fixed.
2894 *) DES library cleanups.
2897 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2898 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2899 ciphers. NOTE: although the key derivation function has been verified
2900 against some published test vectors it has not been extensively tested
2901 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2905 *) Instead of "mkdir -p", which is not fully portable, use new
2906 Perl script "util/mkdir-p.pl".
2909 *) Rewrite the way password based encryption (PBE) is handled. It used to
2910 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2911 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2912 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2913 the 'parameter' field of the AlgorithmIdentifier is passed to the
2914 underlying key generation function so it must do its own ASN1 parsing.
2915 This has also changed the EVP_PBE_CipherInit() function which now has a
2916 'parameter' argument instead of literal salt and iteration count values
2917 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2920 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2921 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2922 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2923 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2924 value was just used as a "magic string" and not used directly its
2925 value doesn't matter.
2928 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2932 *) "linux-sparc64" configuration (ultrapenguin).
2933 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2934 "linux-sparc" configuration.
2935 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2937 *) config now generates no-xxx options for missing ciphers.
2940 *) Support the EBCDIC character set (work in progress).
2941 File ebcdic.c not yet included because it has a different license.
2942 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2944 *) Support BS2000/OSD-POSIX.
2945 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2947 *) Make callbacks for key generation use void * instead of char *.
2950 *) Make S/MIME samples compile (not yet tested).
2953 *) Additional typesafe stacks.
2956 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2960 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2962 *) New configuration variant "sco5-gcc".
2964 *) Updated some demos.
2965 [Sean O Riordain, Wade Scholine]
2967 *) Add missing BIO_free at exit of pkcs12 application.
2970 *) Fix memory leak in conf.c.
2973 *) Updates for Win32 to assembler version of MD5.
2976 *) Set #! path to perl in apps/der_chop to where we found it
2977 instead of using a fixed path.
2980 *) SHA library changes for irix64-mips4-cc.
2983 *) Improvements for VMS support.
2987 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2989 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2990 This also avoids the problems with SC4.2 and unpatched SC5.
2991 [Andy Polyakov <appro@fy.chalmers.se>]
2993 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2994 These are required because of the typesafe stack would otherwise break
2995 existing code. If old code used a structure member which used to be STACK
2996 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2997 sk_num or sk_value it would produce an error because the num, data members
2998 are not present in STACK_OF. Now it just produces a warning. sk_set
2999 replaces the old method of assigning a value to sk_value
3000 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3001 that does this will no longer work (and should use sk_set instead) but
3002 this could be regarded as a "questionable" behaviour anyway.
3005 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3006 correctly handle encrypted S/MIME data.
3009 *) Change type of various DES function arguments from des_cblock
3010 (which means, in function argument declarations, pointer to char)
3011 to des_cblock * (meaning pointer to array with 8 char elements),
3012 which allows the compiler to do more typechecking; it was like
3013 that back in SSLeay, but with lots of ugly casts.
3015 Introduce new type const_des_cblock.
3018 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3019 problems: find RecipientInfo structure that matches recipient certificate
3020 and initialise the ASN1 structures properly based on passed cipher.
3023 *) Belatedly make the BN tests actually check the results.
3026 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3027 to and from BNs: it was completely broken. New compilation option
3028 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3029 key elements as negative integers.
3032 *) Reorganize and speed up MD5.
3033 [Andy Polyakov <appro@fy.chalmers.se>]
3036 [Richard Levitte <richard@levitte.org>]
3038 *) New option -out to asn1parse to allow the parsed structure to be
3039 output to a file. This is most useful when combined with the -strparse
3040 option to examine the output of things like OCTET STRINGS.
3043 *) Make SSL library a little more fool-proof by not requiring any longer
3044 that SSL_set_{accept,connect}_state be called before
3045 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3046 in many applications because usually everything *appeared* to work as
3047 intended anyway -- now it really works as intended).
3050 *) Move openssl.cnf out of lib/.
3053 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3054 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3055 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3056 [Ralf S. Engelschall]
3058 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3059 handle PKCS#7 enveloped data properly.
3060 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3062 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3063 copying pointers. The cert_st handling is changed by this in
3064 various ways (and thus what used to be known as ctx->default_cert
3065 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3066 any longer when s->cert does not give us what we need).
3067 ssl_cert_instantiate becomes obsolete by this change.
3068 As soon as we've got the new code right (possibly it already is?),
3069 we have solved a couple of bugs of the earlier code where s->cert
3070 was used as if it could not have been shared with other SSL structures.
3072 Note that using the SSL API in certain dirty ways now will result
3073 in different behaviour than observed with earlier library versions:
3074 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3075 does not influence s as it used to.
3077 In order to clean up things more thoroughly, inside SSL_SESSION
3078 we don't use CERT any longer, but a new structure SESS_CERT
3079 that holds per-session data (if available); currently, this is
3080 the peer's certificate chain and, for clients, the server's certificate
3081 and temporary key. CERT holds only those values that can have
3082 meaningful defaults in an SSL_CTX.
3085 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3086 from the internal representation. Various PKCS#7 fixes: remove some
3087 evil casts and set the enc_dig_alg field properly based on the signing
3091 *) Allow PKCS#12 password to be set from the command line or the
3092 environment. Let 'ca' get its config file name from the environment
3093 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3097 *) Allow certificate policies extension to use an IA5STRING for the
3098 organization field. This is contrary to the PKIX definition but
3099 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3103 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3104 without disallowing inline assembler and the like for non-pedantic builds.
3107 *) Support Borland C++ builder.
3108 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3113 *) SHA-1 cleanups and performance enhancements.
3114 [Andy Polyakov <appro@fy.chalmers.se>]
3116 *) Sparc v8plus assembler for the bignum library.
3117 [Andy Polyakov <appro@fy.chalmers.se>]
3119 *) Accept any -xxx and +xxx compiler options in Configure.
3122 *) Update HPUX configuration.
3125 *) Add missing sk_<type>_unshift() function to safestack.h
3126 [Ralf S. Engelschall]
3128 *) New function SSL_CTX_use_certificate_chain_file that sets the
3129 "extra_cert"s in addition to the certificate. (This makes sense
3130 only for "PEM" format files, as chains as a whole are not
3134 *) Support verify_depth from the SSL API.
3135 x509_vfy.c had what can be considered an off-by-one-error:
3136 Its depth (which was not part of the external interface)
3137 was actually counting the number of certificates in a chain;
3138 now it really counts the depth.
3141 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3142 instead of X509err, which often resulted in confusing error
3143 messages since the error codes are not globally unique
3144 (e.g. an alleged error in ssl3_accept when a certificate
3145 didn't match the private key).
3147 *) New function SSL_CTX_set_session_id_context that allows to set a default
3148 value (so that you don't need SSL_set_session_id_context for each
3149 connection using the SSL_CTX).
3152 *) OAEP decoding bug fix.
3155 *) Support INSTALL_PREFIX for package builders, as proposed by
3159 *) New Configure options "threads" and "no-threads". For systems
3160 where the proper compiler options are known (currently Solaris
3161 and Linux), "threads" is the default.
3164 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3167 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3168 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3169 such as /usr/local/bin.
3172 *) "make linux-shared" to build shared libraries.
3173 [Niels Poppe <niels@netbox.org>]
3175 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3178 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3179 extension adding in x509 utility.
3182 *) Remove NOPROTO sections and error code comments.
3185 *) Partial rewrite of the DEF file generator to now parse the ANSI
3189 *) New Configure options --prefix=DIR and --openssldir=DIR.
3192 *) Complete rewrite of the error code script(s). It is all now handled
3193 by one script at the top level which handles error code gathering,
3194 header rewriting and C source file generation. It should be much better
3195 than the old method: it now uses a modified version of Ulf's parser to
3196 read the ANSI prototypes in all header files (thus the old K&R definitions
3197 aren't needed for error creation any more) and do a better job of
3198 translating function codes into names. The old 'ASN1 error code imbedded
3199 in a comment' is no longer necessary and it doesn't use .err files which
3200 have now been deleted. Also the error code call doesn't have to appear all
3201 on one line (which resulted in some large lines...).
3204 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3207 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3208 0 (which usually indicates a closed connection), but continue reading.
3211 *) Fix some race conditions.
3214 *) Add support for CRL distribution points extension. Add Certificate
3215 Policies and CRL distribution points documentation.
3218 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3221 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3222 8 of keying material. Merlin has also confirmed interop with this fix
3223 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3224 [Merlin Hughes <merlin@baltimore.ie>]
3226 *) Fix lots of warnings.
3227 [Richard Levitte <levitte@stacken.kth.se>]
3229 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3230 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3231 [Richard Levitte <levitte@stacken.kth.se>]
3233 *) Fix problems with sizeof(long) == 8.
3234 [Andy Polyakov <appro@fy.chalmers.se>]
3236 *) Change functions to ANSI C.
3239 *) Fix typos in error codes.
3240 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3242 *) Remove defunct assembler files from Configure.
3245 *) SPARC v8 assembler BIGNUM implementation.
3246 [Andy Polyakov <appro@fy.chalmers.se>]
3248 *) Support for Certificate Policies extension: both print and set.
3249 Various additions to support the r2i method this uses.
3252 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3253 return a const string when you are expecting an allocated buffer.
3256 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3257 types DirectoryString and DisplayText.
3260 *) Add code to allow r2i extensions to access the configuration database,
3261 add an LHASH database driver and add several ctx helper functions.
3264 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3265 fail when they extended the size of a BIGNUM.
3268 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3269 support typesafe stack.
3272 *) Fix typo in SSL_[gs]et_options().
3273 [Nils Frostberg <nils@medcom.se>]
3275 *) Delete various functions and files that belonged to the (now obsolete)
3276 old X509V3 handling code.
3279 *) New Configure option "rsaref".
3282 *) Don't auto-generate pem.h.
3285 *) Introduce type-safe ASN.1 SETs.
3288 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3289 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3291 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3292 that links with OpenSSL (well at least cause lots of warnings), but fear
3293 not: the conversion is trivial, and it eliminates loads of evil casts. A
3294 few STACKed things have been converted already. Feel free to convert more.
3295 In the fullness of time, I'll do away with the STACK type altogether.
3298 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3299 specified in <certfile> by updating the entry in the index.txt file.
3300 This way one no longer has to edit the index.txt file manually for
3301 revoking a certificate. The -revoke option does the gory details now.
3302 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3304 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3305 `-text' option at all and this way the `-noout -text' combination was
3306 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3307 [Ralf S. Engelschall]
3309 *) Make sure a corresponding plain text error message exists for the
3310 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3311 verify callback function determined that a certificate was revoked.
3312 [Ralf S. Engelschall]
3314 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3315 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3316 all available cipers including rc5, which was forgotten until now.
3317 In order to let the testing shell script know which algorithms
3318 are available, a new (up to now undocumented) command
3319 "openssl list-cipher-commands" is used.
3322 *) Bugfix: s_client occasionally would sleep in select() when
3323 it should have checked SSL_pending() first.
3326 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3327 the raw DSA values prior to ASN.1 encoding.
3330 *) Tweaks to Configure
3331 [Niels Poppe <niels@netbox.org>]
3333 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3337 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3340 *) New config option to avoid instructions that are illegal on the 80386.
3341 The default code is faster, but requires at least a 486.
3344 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3345 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3346 same as SSL2_VERSION anyway.
3349 *) New "-showcerts" option for s_client.
3352 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3353 application. Various cleanups and fixes.
3356 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3357 modify error routines to work internally. Add error codes and PBE init
3358 to library startup routines.
3361 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3362 packing functions to asn1 and evp. Changed function names and error
3363 codes along the way.
3366 *) PKCS12 integration: and so it begins... First of several patches to
3367 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3368 objects to objects.h
3371 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3372 and display support for Thawte strong extranet extension.
3375 *) Add LinuxPPC support.
3376 [Jeff Dubrule <igor@pobox.org>]
3378 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3379 bn_div_words in alpha.s.
3380 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3382 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3383 OAEP isn't supported when OpenSSL is built with RSAref.
3384 [Ulf Moeller <ulf@fitug.de>]
3386 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3387 so they no longer are missing under -DNOPROTO.
3388 [Soren S. Jorvang <soren@t.dk>]
3391 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3393 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3394 doesn't work when the session is reused. Coming soon!
3397 *) Fix a security hole, that allows sessions to be reused in the wrong
3398 context thus bypassing client cert protection! All software that uses
3399 client certs and session caches in multiple contexts NEEDS PATCHING to
3400 allow session reuse! A fuller solution is in the works.
3401 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3403 *) Some more source tree cleanups (removed obsolete files
3404 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3405 permission on "config" script to be executable) and a fix for the INSTALL
3407 [Ulf Moeller <ulf@fitug.de>]
3409 *) Remove some legacy and erroneous uses of malloc, free instead of
3411 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3413 *) Make rsa_oaep_test return non-zero on error.
3414 [Ulf Moeller <ulf@fitug.de>]
3416 *) Add support for native Solaris shared libraries. Configure
3417 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3418 if someone would make that last step automatic.
3419 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3421 *) ctx_size was not built with the right compiler during "make links". Fixed.
3424 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3425 except NULL ciphers". This means the default cipher list will no longer
3426 enable NULL ciphers. They need to be specifically enabled e.g. with
3427 the string "DEFAULT:eNULL".
3430 *) Fix to RSA private encryption routines: if p < q then it would
3431 occasionally produce an invalid result. This will only happen with
3432 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3435 *) Be less restrictive and allow also `perl util/perlpath.pl
3436 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3437 because this way one can also use an interpreter named `perl5' (which is
3438 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3439 installed as `perl').
3440 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3442 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3443 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3445 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3446 advapi32.lib to Win32 build and change the pem test comparision
3447 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3448 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3449 and crypto/des/ede_cbcm_enc.c.
3452 *) DES quad checksum was broken on big-endian architectures. Fixed.
3455 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3456 Win32 test batch file so it (might) work again. The Win32 test batch file
3457 is horrible: I feel ill....
3460 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3461 in e_os.h. Audit of header files to check ANSI and non ANSI
3462 sections: 10 functions were absent from non ANSI section and not exported
3463 from Windows DLLs. Fixed up libeay.num for new functions.
3466 *) Make `openssl version' output lines consistent.
3467 [Ralf S. Engelschall]
3469 *) Fix Win32 symbol export lists for BIO functions: Added
3470 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3471 to ms/libeay{16,32}.def.
3472 [Ralf S. Engelschall]
3474 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3475 fine under Unix and passes some trivial tests I've now added. But the
3476 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3477 added to make sure no one expects that this stuff really works in the
3478 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3479 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3481 [Ralf S. Engelschall]
3483 *) Fix the generation of two part addresses in perl.
3484 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3486 *) Add config entry for Linux on MIPS.
3487 [John Tobey <jtobey@channel1.com>]
3489 *) Make links whenever Configure is run, unless we are on Windoze.
3492 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3493 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3497 *) Add a useful kludge to allow package maintainers to specify compiler and
3498 other platforms details on the command line without having to patch the
3499 Configure script everytime: One now can use ``perl Configure
3500 <id>:<details>'', i.e. platform ids are allowed to have details appended
3501 to them (seperated by colons). This is treated as there would be a static
3502 pre-configured entry in Configure's %table under key <id> with value
3503 <details> and ``perl Configure <id>'' is called. So, when you want to
3504 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3505 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3506 now, which overrides the FreeBSD-elf entry on-the-fly.
3507 [Ralf S. Engelschall]
3509 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3512 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3513 on the `perl Configure ...' command line. This way one can compile
3514 OpenSSL libraries with Position Independent Code (PIC) which is needed
3515 for linking it into DSOs.
3516 [Ralf S. Engelschall]
3518 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3522 *) Cleaned up the LICENSE document: The official contact for any license
3523 questions now is the OpenSSL core team under openssl-core@openssl.org.
3524 And add a paragraph about the dual-license situation to make sure people
3525 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3526 to the OpenSSL toolkit.
3527 [Ralf S. Engelschall]
3529 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3530 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3531 Additonally cleaned up the `make links' target: Remove unnecessary
3532 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3533 to speed processing and no longer clutter the display with confusing
3534 stuff. Instead only the actually done links are displayed.
3535 [Ralf S. Engelschall]
3537 *) Permit null encryption ciphersuites, used for authentication only. It used
3538 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3539 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3543 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3544 signed attributes when verifying signatures (this would break them),
3545 the detached data encoding was wrong and public keys obtained using
3546 X509_get_pubkey() weren't freed.
3549 *) Add text documentation for the BUFFER functions. Also added a work around
3550 to a Win95 console bug. This was triggered by the password read stuff: the
3551 last character typed gets carried over to the next fread(). If you were
3552 generating a new cert request using 'req' for example then the last
3553 character of the passphrase would be CR which would then enter the first
3557 *) Added the new `Includes OpenSSL Cryptography Software' button as
3558 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3559 button and can be used by applications based on OpenSSL to show the
3560 relationship to the OpenSSL project.
3561 [Ralf S. Engelschall]
3563 *) Remove confusing variables in function signatures in files
3564 ssl/ssl_lib.c and ssl/ssl.h.
3565 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3567 *) Don't install bss_file.c under PREFIX/include/
3568 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3570 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3571 functions that return function pointers and has support for NT specific
3572 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3573 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3574 unsigned to signed types: this was killing the Win32 compile.
3577 *) Add new certificate file to stack functions,
3578 SSL_add_dir_cert_subjects_to_stack() and
3579 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3580 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3581 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3582 This means that Apache-SSL and similar packages don't have to mess around
3583 to add as many CAs as they want to the preferred list.
3586 *) Experiment with doxygen documentation. Currently only partially applied to
3588 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3589 openssl.doxy as the configuration file.
3592 *) Get rid of remaining C++-style comments which strict C compilers hate.
3593 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3595 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3596 compiled in by default: it has problems with large keys.
3599 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3600 DH private keys and/or callback functions which directly correspond to
3601 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3602 is needed for applications which have to configure certificates on a
3603 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3605 For the RSA certificate situation is makes no difference, but
3606 for the DSA certificate situation this fixes the "no shared cipher"
3607 problem where the OpenSSL cipher selection procedure failed because the
3608 temporary keys were not overtaken from the context and the API provided
3609 no way to reconfigure them.
3610 The new functions now let applications reconfigure the stuff and they
3611 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3612 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3613 non-public-API function ssl_cert_instantiate() is used as a helper
3614 function and also to reduce code redundancy inside ssl_rsa.c.
3615 [Ralf S. Engelschall]
3617 *) Move s_server -dcert and -dkey options out of the undocumented feature
3618 area because they are useful for the DSA situation and should be
3619 recognized by the users.
3620 [Ralf S. Engelschall]
3622 *) Fix the cipher decision scheme for export ciphers: the export bits are
3623 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3624 SSL_EXP_MASK. So, the original variable has to be used instead of the
3625 already masked variable.
3626 [Richard Levitte <levitte@stacken.kth.se>]
3628 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3629 [Richard Levitte <levitte@stacken.kth.se>]
3631 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3632 from `int' to `unsigned int' because it's a length and initialized by
3633 EVP_DigestFinal() which expects an `unsigned int *'.
3634 [Richard Levitte <levitte@stacken.kth.se>]
3636 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3637 script. Instead use the usual Shell->Perl transition trick.
3638 [Ralf S. Engelschall]
3640 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3641 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3642 -noout -modulus' as it's already the case for `openssl rsa -noout
3643 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3644 currently the public key is printed (a decision which was already done by
3645 `openssl dsa -modulus' in the past) which serves a similar purpose.
3646 Additionally the NO_RSA no longer completely removes the whole -modulus
3647 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3649 [Ralf S. Engelschall]
3651 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3652 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3653 [Arne Ansper <arne@ats.cyber.ee>]
3655 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3656 to be added. Now both 'req' and 'ca' can use new objects defined in the
3660 *) Add cool BIO that does syslog (or event log on NT).
3661 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3663 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3664 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3665 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3666 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3669 *) Add preliminary config info for new extension code.
3672 *) Make RSA_NO_PADDING really use no padding.
3673 [Ulf Moeller <ulf@fitug.de>]
3675 *) Generate errors when private/public key check is done.
3678 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3679 for some CRL extensions and new objects added.
3682 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3683 key usage extension and fuller support for authority key id.
3686 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3687 padding method for RSA, which is recommended for new applications in PKCS
3688 #1 v2.0 (RFC 2437, October 1998).
3689 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3690 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3691 against Bleichbacher's attack on RSA.
3692 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3695 *) Updates to the new SSL compression code
3696 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3698 *) Fix so that the version number in the master secret, when passed
3699 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3700 (because the server will not accept higher), that the version number
3701 is 0x03,0x01, not 0x03,0x00
3702 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3704 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3705 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3706 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3709 *) Support for RAW extensions where an arbitrary extension can be
3710 created by including its DER encoding. See apps/openssl.cnf for
3714 *) Make sure latest Perl versions don't interpret some generated C array
3715 code as Perl array code in the crypto/err/err_genc.pl script.
3716 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3718 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3719 not many people have the assembler. Various Win32 compilation fixes and
3720 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3724 *) Modify configure script 'Configure' to automatically create crypto/date.h
3725 file under Win32 and also build pem.h from pem.org. New script
3726 util/mkfiles.pl to create the MINFO file on environments that can't do a
3727 'make files': perl util/mkfiles.pl >MINFO should work.
3730 *) Major rework of DES function declarations, in the pursuit of correctness
3731 and purity. As a result, many evil casts evaporated, and some weirdness,
3732 too. You may find this causes warnings in your code. Zapping your evil
3733 casts will probably fix them. Mostly.
3736 *) Fix for a typo in asn1.h. Bug fix to object creation script
3737 obj_dat.pl. It considered a zero in an object definition to mean
3738 "end of object": none of the objects in objects.h have any zeros
3739 so it wasn't spotted.
3740 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3742 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3743 Masking (CBCM). In the absence of test vectors, the best I have been able
3744 to do is check that the decrypt undoes the encrypt, so far. Send me test
3745 vectors if you have them.
3748 *) Correct calculation of key length for export ciphers (too much space was
3749 allocated for null ciphers). This has not been tested!
3752 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3753 message is now correct (it understands "crypto" and "ssl" on its
3754 command line). There is also now an "update" option. This will update
3755 the util/ssleay.num and util/libeay.num files with any new functions.
3757 perl util/mkdef.pl crypto ssl update
3758 it will update them.
3761 *) Overhauled the Perl interface (perl/*):
3762 - ported BN stuff to OpenSSL's different BN library
3763 - made the perl/ source tree CVS-aware
3764 - renamed the package from SSLeay to OpenSSL (the files still contain
3765 their history because I've copied them in the repository)
3766 - removed obsolete files (the test scripts will be replaced
3767 by better Test::Harness variants in the future)
3768 [Ralf S. Engelschall]
3770 *) First cut for a very conservative source tree cleanup:
3771 1. merge various obsolete readme texts into doc/ssleay.txt
3772 where we collect the old documents and readme texts.
3773 2. remove the first part of files where I'm already sure that we no
3774 longer need them because of three reasons: either they are just temporary
3775 files which were left by Eric or they are preserved original files where
3776 I've verified that the diff is also available in the CVS via "cvs diff
3777 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3778 the crypto/md/ stuff).
3779 [Ralf S. Engelschall]
3781 *) More extension code. Incomplete support for subject and issuer alt
3782 name, issuer and authority key id. Change the i2v function parameters
3783 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3784 what that's for :-) Fix to ASN1 macro which messed up
3785 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3788 *) Preliminary support for ENUMERATED type. This is largely copied from the
3792 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3793 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3795 *) Make sure `make rehash' target really finds the `openssl' program.
3796 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3798 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3799 like to hear about it if this slows down other processors.
3802 *) Add CygWin32 platform information to Configure script.
3803 [Alan Batie <batie@aahz.jf.intel.com>]
3805 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3806 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3808 *) New program nseq to manipulate netscape certificate sequences
3811 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3815 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3816 but the BN code had some problems that would cause failures when
3817 doing certificate verification and some other functions.
3818 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3820 *) Add ASN1 and PEM code to support netscape certificate sequences.
3823 *) Add ASN1 and PEM code to support netscape certificate sequences.
3826 *) Add several PKIX and private extended key usage OIDs.
3829 *) Modify the 'ca' program to handle the new extension code. Modify
3830 openssl.cnf for new extension format, add comments.
3833 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3834 and add a sample to openssl.cnf so req -x509 now adds appropriate
3838 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3839 error code, add initial support to X509_print() and x509 application.
3842 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3843 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3844 stuff is currently isolated and isn't even compiled yet.
3847 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3848 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3849 Removed the versions check from X509 routines when loading extensions:
3850 this allows certain broken certificates that don't set the version
3851 properly to be processed.
3854 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3855 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3856 can still be regenerated with "make depend".
3859 *) Spelling mistake in C version of CAST-128.
3860 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3862 *) Changes to the error generation code. The perl script err-code.pl
3863 now reads in the old error codes and retains the old numbers, only
3864 adding new ones if necessary. It also only changes the .err files if new
3865 codes are added. The makefiles have been modified to only insert errors
3866 when needed (to avoid needlessly modifying header files). This is done
3867 by only inserting errors if the .err file is newer than the auto generated
3868 C file. To rebuild all the error codes from scratch (the old behaviour)
3869 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3870 or delete all the .err files.
3873 *) CAST-128 was incorrectly implemented for short keys. The C version has
3874 been fixed, but is untested. The assembler versions are also fixed, but
3875 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3876 to regenerate it if needed.
3877 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3878 Hagino <itojun@kame.net>]
3880 *) File was opened incorrectly in randfile.c.
3881 [Ulf Möller <ulf@fitug.de>]
3883 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3884 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3885 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3886 al: it's just almost always a UTCTime. Note this patch adds new error
3887 codes so do a "make errors" if there are problems.
3890 *) Correct Linux 1 recognition in config.
3891 [Ulf Möller <ulf@fitug.de>]
3893 *) Remove pointless MD5 hash when using DSA keys in ca.
3894 [Anonymous <nobody@replay.com>]
3896 *) Generate an error if given an empty string as a cert directory. Also
3897 generate an error if handed NULL (previously returned 0 to indicate an
3898 error, but didn't set one).
3899 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3901 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3904 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3905 parameters. This was causing a warning which killed off the Win32 compile.
3908 *) Remove C++ style comments from crypto/bn/bn_local.h.
3909 [Neil Costigan <neil.costigan@celocom.com>]
3911 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3912 based on a text string, looking up short and long names and finally
3913 "dot" format. The "dot" format stuff didn't work. Added new function
3914 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3915 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3916 OID is not part of the table.
3919 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3920 X509_LOOKUP_by_alias().
3923 *) Sort openssl functions by name.
3926 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3927 encryption from sample DSA keys (in case anyone is interested the password
3931 *) Make _all_ *_free functions accept a NULL pointer.
3932 [Frans Heymans <fheymans@isaserver.be>]
3934 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3936 [Anonymous <nobody@replay.com>]
3938 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3939 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3941 *) Don't blow it for numeric -newkey arguments to apps/req.
3942 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3944 *) Temp key "for export" tests were wrong in s3_srvr.c.
3945 [Anonymous <nobody@replay.com>]
3947 *) Add prototype for temp key callback functions
3948 SSL_CTX_set_tmp_{rsa,dh}_callback().
3951 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3952 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3955 *) X509_name_add_entry() freed the wrong thing after an error.
3956 [Arne Ansper <arne@ats.cyber.ee>]
3958 *) rsa_eay.c would attempt to free a NULL context.
3959 [Arne Ansper <arne@ats.cyber.ee>]
3961 *) BIO_s_socket() had a broken should_retry() on Windoze.
3962 [Arne Ansper <arne@ats.cyber.ee>]
3964 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3965 [Arne Ansper <arne@ats.cyber.ee>]
3967 *) Make sure the already existing X509_STORE->depth variable is initialized
3968 in X509_STORE_new(), but document the fact that this variable is still
3969 unused in the certificate verification process.
3970 [Ralf S. Engelschall]
3972 *) Fix the various library and apps files to free up pkeys obtained from
3973 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3976 *) Fix reference counting in X509_PUBKEY_get(). This makes
3977 demos/maurice/example2.c work, amongst others, probably.
3978 [Steve Henson and Ben Laurie]
3980 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3981 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3982 are no longer created. This way we have a single and consistent command
3983 line interface `openssl <command>', similar to `cvs <command>'.
3984 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3986 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3987 BIT STRING wrapper always have zero unused bits.
3990 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3993 *) Make the top-level INSTALL documentation easier to understand.
3996 *) Makefiles updated to exit if an error occurs in a sub-directory
3997 make (including if user presses ^C) [Paul Sutton]
3999 *) Make Montgomery context stuff explicit in RSA data structure.
4002 *) Fix build order of pem and err to allow for generated pem.h.
4005 *) Fix renumbering bug in X509_NAME_delete_entry().
4008 *) Enhanced the err-ins.pl script so it makes the error library number
4009 global and can add a library name. This is needed for external ASN1 and
4010 other error libraries.
4013 *) Fixed sk_insert which never worked properly.
4016 *) Fix ASN1 macros so they can handle indefinite length construted
4017 EXPLICIT tags. Some non standard certificates use these: they can now
4021 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4022 into a single doc/ssleay.txt bundle. This way the information is still
4023 preserved but no longer messes up this directory. Now it's new room for
4024 the new set of documenation files.
4025 [Ralf S. Engelschall]
4027 *) SETs were incorrectly DER encoded. This was a major pain, because they
4028 shared code with SEQUENCEs, which aren't coded the same. This means that
4029 almost everything to do with SETs or SEQUENCEs has either changed name or
4030 number of arguments.
4031 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4033 *) Fix test data to work with the above.
4036 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4037 was already fixed by Eric for 0.9.1 it seems.
4038 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4040 *) Autodetect FreeBSD3.
4043 *) Fix various bugs in Configure. This affects the following platforms:
4047 unixware-2.0-pentium
4051 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4052 before they are needed.
4055 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4059 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4061 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4062 changed SSLeay to OpenSSL in version strings.
4063 [Ralf S. Engelschall]
4065 *) Some fixups to the top-level documents.
4068 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4069 because the symlink to include/ was missing.
4070 [Ralf S. Engelschall]
4072 *) Incorporated the popular no-RSA/DSA-only patches
4073 which allow to compile a RSA-free SSLeay.
4074 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4076 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4077 when "ssleay" is still not found.
4078 [Ralf S. Engelschall]
4080 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4081 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4083 *) Updated the README file.
4084 [Ralf S. Engelschall]
4086 *) Added various .cvsignore files in the CVS repository subdirs
4087 to make a "cvs update" really silent.
4088 [Ralf S. Engelschall]
4090 *) Recompiled the error-definition header files and added
4091 missing symbols to the Win32 linker tables.
4092 [Ralf S. Engelschall]
4094 *) Cleaned up the top-level documents;
4095 o new files: CHANGES and LICENSE
4096 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4097 o merged COPYRIGHT into LICENSE
4098 o removed obsolete TODO file
4099 o renamed MICROSOFT to INSTALL.W32
4100 [Ralf S. Engelschall]
4102 *) Removed dummy files from the 0.9.1b source tree:
4103 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4104 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4105 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4106 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4107 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4108 [Ralf S. Engelschall]
4110 *) Added various platform portability fixes.
4113 *) The Genesis of the OpenSSL rpject:
4114 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4115 Young and Tim J. Hudson created while they were working for C2Net until
4117 [The OpenSSL Project]
4120 Changes between 0.9.0b and 0.9.1b [not released]
4122 *) Updated a few CA certificates under certs/
4125 *) Changed some BIGNUM api stuff.
4128 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4129 DGUX x86, Linux Alpha, etc.
4132 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4133 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4137 *) Add -strparse option to asn1pars program which parses nested
4139 [Dr Stephen Henson <shenson@bigfoot.com>]
4141 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4144 *) DSA fix for "ca" program.
4147 *) Added "-genkey" option to "dsaparam" program.
4150 *) Added RIPE MD160 (rmd160) message digest.
4153 *) Added -a (all) option to "ssleay version" command.
4156 *) Added PLATFORM define which is the id given to Configure.
4159 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4162 *) Extended the ASN.1 parser routines.
4165 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4168 *) Added a BN_CTX to the BN library.
4171 *) Fixed the weak key values in DES library
4174 *) Changed API in EVP library for cipher aliases.
4177 *) Added support for RC2/64bit cipher.
4180 *) Converted the lhash library to the crypto/mem.c functions.
4183 *) Added more recognized ASN.1 object ids.
4186 *) Added more RSA padding checks for SSL/TLS.
4189 *) Added BIO proxy/filter functionality.
4192 *) Added extra_certs to SSL_CTX which can be used
4193 send extra CA certificates to the client in the CA cert chain sending
4194 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4197 *) Now Fortezza is denied in the authentication phase because
4198 this is key exchange mechanism is not supported by SSLeay at all.
4201 *) Additional PKCS1 checks.
4204 *) Support the string "TLSv1" for all TLS v1 ciphers.
4207 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4208 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4211 *) Fixed a few memory leaks.
4214 *) Fixed various code and comment typos.
4217 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4218 bytes sent in the client random.
4219 [Edward Bishop <ebishop@spyglass.com>]