4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8 extract information from a certificate request. OCSP_response_create()
9 creates a response and optionally adds a basic response structure.
10 OCSP_basic_add1_status() adds a complete single response to a basic
11 reponse and returns the OCSP_SINGLERESP structure just added (to allow
12 extensions to be included for example). OCSP_basic_add1_cert() adds a
13 certificate to a basic response and OCSP_basic_sign() signs a basic
14 response with various flags. New helper functions ASN1_TIME_check()
15 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
16 (converts ASN1_TIME to GeneralizedTime).
19 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
20 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
21 structure from a certificate. X509_pubkey_digest() digests tha public_key
22 contents: this is used in various key identifiers.
25 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
26 keyUsage if basicConstraints absent for a CA.
29 *) Make SMIME_write_PKCS7() write mail header values with a format that
30 is more generally accepted (no spaces before the semicolon), since
31 some programs can't parse those values properly otherwise. Also make
32 sure BIO's that break lines after each write do not create invalid
36 *) Make sk_sort() tolerate a NULL argument.
37 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
39 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
40 passed by the function are trusted implicitly. If any of them signed the
41 reponse then it is assumed to be valid and is not verified.
44 *) Zero the premaster secret after deriving the master secret in
48 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
49 to data. This was previously part of the PKCS7 ASN1 code. This
50 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
51 [Steve Henson, reported by Kenneth R. Robinette
52 <support@securenetterm.com>]
54 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
55 routines: without these tracing memory leaks is very painful.
56 Fix leaks in PKCS12 and PKCS7 routines.
59 *) Fix for Irix with NO_ASM.
60 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
62 *) Add some EVP_add_digest_alias registrations (as found in
63 OpenSSL_add_all_digests()), to SSL_library_init()
64 aka OpenSSL_add_ssl_algorithms(). This provides improved
65 compatibility with peers using X.509 certificates
66 with unconventional AlgorithmIdentifier OIDs.
69 *) ./config script fixes.
70 [Ulf Moeller, Richard Levitte]
72 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
73 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
74 effectively meant GeneralizedTime would never be used. Now it
75 is initialised to -1 but X509_time_adj() now has to check the value
76 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
77 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
78 [Steve Henson, reported by Kenneth R. Robinette
79 <support@securenetterm.com>]
81 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
82 result in a zero length in the ASN1_INTEGER structure which was
83 not consistent with the structure when d2i_ASN1_INTEGER() was used
84 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
85 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
86 where it did not print out a minus for negative ASN1_INTEGER.
89 *) Fix 'openssl passwd -1'.
92 *) Add summary printout to ocsp utility. The various functions which
93 convert status values to strings have been renamed to:
94 OCSP_response_status_str(), OCSP_cert_status_str() and
95 OCSP_crl_reason_str() and are no longer static. New options
96 to verify nonce values and to disable verification. OCSP response
97 printout format cleaned up.
100 *) Add additional OCSP certificate checks. These are those specified
101 in RFC2560. This consists of two separate checks: the CA of the
102 certificate being checked must either be the OCSP signer certificate
103 or the issuer of the OCSP signer certificate. In the latter case the
104 OCSP signer certificate must contain the OCSP signing extended key
105 usage. This check is performed by attempting to match the OCSP
106 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
107 in the OCSP_CERTID structures of the response.
110 *) Initial OCSP certificate verification added to OCSP_basic_verify()
111 and related routines. This uses the standard OpenSSL certificate
112 verify routines to perform initial checks (just CA validity) and
113 to obtain the certificate chain. Then additional checks will be
114 performed on the chain. Currently the root CA is checked to see
115 if it is explicitly trusted for OCSP signing. This is used to set
116 a root CA as a global signing root: that is any certificate that
117 chains to that CA is an acceptable OCSP signing certificate.
120 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
121 extensions from a separate configuration file.
122 As when reading extensions from the main configuration file,
123 the '-extensions ...' option may be used for specifying the
125 [Massimiliano Pala <madwolf@comune.modena.it>]
127 *) Change PKCS12_key_gen_asc() so it can cope with non null
128 terminated strings whose length is passed in the passlen
129 parameter, for example from PEM callbacks. This was done
130 by adding an extra length parameter to asc2uni().
131 [Steve Henson, reported by <oddissey@samsung.co.kr>]
133 *) New OCSP utility. Allows OCSP requests to be generated or
134 read. The request can be sent to a responder and the output
135 parsed, outputed or printed in text form. Not complete yet:
136 still needs to check the OCSP response validity.
139 *) New subcommands for 'openssl ca':
140 'openssl ca -status <serial>' prints the status of the cert with
141 the given serial number (according to the index file).
142 'openssl ca -updatedb' updates the expiry status of certificates
144 [Massimiliano Pala <madwolf@comune.modena.it>]
146 *) New '-newreq-nodes' command option to CA.pl. This is like
147 '-newreq', but calls 'openssl req' with the '-nodes' option
148 so that the resulting key is not encrypted.
149 [Damien Miller <djm@mindrot.org>]
151 *) New configuration for the GNU Hurd.
152 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
154 *) Initial code to implement OCSP basic response verify. This
155 is currently incomplete. Currently just finds the signer's
156 certificate and verifies the signature on the response.
159 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
160 value of OPENSSLDIR. This is available via the new '-d' option
161 to 'openssl version', and is also included in 'openssl version -a'.
164 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
165 call failed, free the DSA structure.
168 *) Allowing defining memory allocation callbacks that will be given
169 file name and line number information in additional arguments
170 (a const char* and an int). The basic functionality remains, as
171 well as the original possibility to just replace malloc(),
172 realloc() and free() by functions that do not know about these
173 additional arguments. To register and find out the current
174 settings for extended allocation functions, the following
175 functions are provided:
177 CRYPTO_set_mem_ex_functions
178 CRYPTO_set_locked_mem_ex_functions
179 CRYPTO_get_mem_ex_functions
180 CRYPTO_get_locked_mem_ex_functions
182 These work the same way as CRYPTO_set_mem_functions and friends.
183 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
184 extended allocation function is enabled.
185 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
186 a conventional allocation function is enabled.
187 [Richard Levitte, Bodo Moeller]
189 *) Fix to uni2asc() to cope with zero length Unicode strings.
190 These are present in some PKCS#12 files.
193 *) Finish off removing the remaining LHASH function pointer casts.
194 There should no longer be any prototype-casting required when using
195 the LHASH abstraction, and any casts that remain are "bugs". See
196 the callback types and macros at the head of lhash.h for details
197 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
200 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
201 If an EGD or PRNGD is running and enough entropy is returned, automatic
202 seeding like with /dev/[u]random will be performed.
203 Positions tried are: /etc/entropy, /var/run/egd-pool.
206 *) Change the Unix RAND_poll() variant to be able to poll several
207 random devices, as specified by DEVRANDOM, until a sufficient amount
208 of data has been collected. We spend at most 10 ms on each file
209 (select timeout) and read in non-blocking mode. DEVRANDOM now
210 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
211 (previously it was just the string "/dev/urandom"), so on typical
212 platforms the 10 ms delay will never occur.
213 Also separate out the Unix variant to its own file, rand_unix.c.
214 For VMS, there's a currently-empty rand_vms.c.
217 *) Move OCSP client related routines to ocsp_cl.c. These
218 provide utility functions which an application needing
219 to issue a request to an OCSP responder and analyse the
220 response will typically need: as opposed to those which an
221 OCSP responder itself would need which will be added later.
223 OCSP_request_sign() signs an OCSP request with an API similar
224 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
225 response. OCSP_response_get1_basic() extracts basic response
226 from response. OCSP_resp_find_status(): finds and extracts status
227 information from an OCSP_CERTID structure (which will be created
228 when the request structure is built). These are built from lower
229 level functions which work on OCSP_SINGLERESP structures but
230 wont normally be used unless the application wishes to examine
231 extensions in the OCSP response for example.
233 Replace nonce routines with a pair of functions.
234 OCSP_request_add1_nonce() adds a nonce value and optionally
235 generates a random value. OCSP_check_nonce() checks the
236 validity of the nonce in an OCSP response.
239 *) Change function OCSP_request_add() to OCSP_request_add0_id().
240 This doesn't copy the supplied OCSP_CERTID and avoids the
241 need to free up the newly created id. Change return type
242 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
243 This can then be used to add extensions to the request.
244 Deleted OCSP_request_new(), since most of its functionality
245 is now in OCSP_REQUEST_new() (and the case insensitive name
246 clash) apart from the ability to set the request name which
247 will be added elsewhere.
250 *) Update OCSP API. Remove obsolete extensions argument from
251 various functions. Extensions are now handled using the new
252 OCSP extension code. New simple OCSP HTTP function which
253 can be used to send requests and parse the response.
256 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
257 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
258 uses the special reorder version of SET OF to sort the attributes
259 and reorder them to match the encoded order. This resolves a long
260 standing problem: a verify on a PKCS7 structure just after signing
261 it used to fail because the attribute order did not match the
262 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
263 it uses the received order. This is necessary to tolerate some broken
264 software that does not order SET OF. This is handled by encoding
265 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
266 to produce the required SET OF.
269 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
270 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
271 files to get correct declarations of the ASN.1 item variables.
274 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
275 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
276 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
277 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
278 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
279 ASN1_ITEM and no wrapper functions.
282 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
283 replace the old function pointer based I/O routines. Change most of
284 the *_d2i_bio() and *_d2i_fp() functions to use these.
287 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
288 lines, recognice more "algorithms" that can be deselected, and make
289 it complain about algorithm deselection that isn't recognised.
292 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
293 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
294 to use new functions. Add NO_ASN1_OLD which can be set to remove
295 some old style ASN1 functions: this can be used to determine if old
296 code will still work when these eventually go away.
299 *) New extension functions for OCSP structures, these follow the
300 same conventions as certificates and CRLs.
303 *) New function X509V3_add1_i2d(). This automatically encodes and
304 adds an extension. Its behaviour can be customised with various
305 flags to append, replace or delete. Various wrappers added for
306 certifcates and CRLs.
309 *) Fix to avoid calling the underlying ASN1 print routine when
310 an extension cannot be parsed. Correct a typo in the
311 OCSP_SERVICELOC extension. Tidy up print OCSP format.
314 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
315 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
316 when writing a 32767 byte record.
317 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
319 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
320 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
322 (RSA objects have a reference count access to which is protected
323 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
324 so they are meant to be shared between threads.)
325 [Bodo Moeller, Geoff Thorpe; original patch submitted by
326 "Reddie, Steven" <Steven.Reddie@ca.com>]
328 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
329 entries for variables.
332 *) Fix a deadlock in CRYPTO_mem_leaks().
335 *) Add functionality to apps/openssl.c for detecting locking
336 problems: As the program is single-threaded, all we have
337 to do is register a locking callback using an array for
338 storing which locks are currently held by the program.
341 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
342 SSL_get_ex_data_X509_STORE_idx(), which is used in
343 ssl_verify_cert_chain() and thus can be called at any time
344 during TLS/SSL handshakes so that thread-safety is essential.
345 Unfortunately, the ex_data design is not at all suited
346 for multi-threaded use, so it probably should be abolished.
349 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
350 [Broadcom, tweaked and integrated by Geoff Thorpe]
352 *) Move common extension printing code to new function
353 X509V3_print_extensions(). Reorganise OCSP print routines and
354 implement some needed OCSP ASN1 functions. Add OCSP extensions.
357 *) New function X509_signature_print() to remove duplication in some
361 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
362 set (this was treated exactly the same as SET OF previously). This
363 is used to reorder the STACK representing the structure to match the
364 encoding. This will be used to get round a problem where a PKCS7
365 structure which was signed could not be verified because the STACK
366 order did not reflect the encoded order.
369 *) Reimplement the OCSP ASN1 module using the new code.
372 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
373 for its ASN1 operations. The old style function pointers still exist
374 for now but they will eventually go away.
377 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
378 completely replaces the old ASN1 functionality with a table driven
379 encoder and decoder which interprets an ASN1_ITEM structure describing
380 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
381 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
382 has also been converted to the new form.
385 *) Change BN_mod_exp_recp so that negative moduli are tolerated
386 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
387 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
391 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
392 of not touching the result's sign bit.
395 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
399 *) Changed the LHASH code to use prototypes for callbacks, and created
400 macros to declare and implement thin (optionally static) functions
401 that provide type-safety and avoid function pointer casting for the
402 type-specific callbacks.
405 *) Use better test patterns in bntest.
408 *) Added Kerberos Cipher Suites to be used with TLS, as written in
410 [Veers Staats <staatsvr@asc.hpc.mil>,
411 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
413 *) rand_win.c fix for Borland C.
416 *) BN_rshift bugfix for n == 0.
419 *) Reformat the FAQ so the different questions and answers can be divided
420 in sections depending on the subject.
423 *) Have the zlib compression code load ZLIB.DLL dynamically under
427 *) New function BN_mod_sqrt for computing square roots modulo a prime
428 (using the probabilistic Tonelli-Shanks algorithm unless
429 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
430 be handled deterministically).
431 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
433 *) Store verify_result within SSL_SESSION also for client side to
434 avoid potential security hole. (Re-used sessions on the client side
435 always resulted in verify_result==X509_V_OK, not using the original
436 result of the server certificate verification.)
439 *) Make BN_mod_inverse faster by explicitly handling small quotients
440 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
441 512 bits], about 30% for larger ones [1024 or 2048 bits].)
444 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
445 SSL3_RT_APPLICATION_DATA, return 0.
446 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
450 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
451 releases, have been re-implemented by renaming the previous
452 implementations of ssl2_read and ssl3_read to ssl2_read_internal
453 and ssl3_read_internal, respectively, and adding 'peek' parameters
454 to them. The new ssl[23]_{read,peek} functions are calls to
455 ssl[23]_read_internal with the 'peek' flag set appropriately.
456 A 'peek' parameter has also been added to ssl3_read_bytes, which
457 does the actual work for ssl3_read_internal.
460 *) New function BN_kronecker.
463 *) Fix BN_gcd so that it works on negative inputs; the result is
464 positive unless both parameters are zero.
465 Previously something reasonably close to an infinite loop was
466 possible because numbers could be growing instead of shrinking
467 in the implementation of Euclid's algorithm.
470 *) Fix BN_is_word() and BN_is_one() macros to take into account the
471 sign of the number in question.
473 Fix BN_is_word(a,w) to work correctly for w == 0.
475 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
476 because its test if the absolute value of 'a' equals 'w'.
477 Note that BN_abs_is_word does *not* handle w == 0 reliably;
478 it exists mostly for use in the implementations of BN_is_zero(),
479 BN_is_one(), and BN_is_word().
482 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
483 the method-specific "init()" handler. Also clean up ex_data after
484 calling the method-specific "finish()" handler. Previously, this was
485 happening the other way round.
488 *) New function BN_swap.
491 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
492 the exponentiation functions are more likely to produce reasonable
493 results on negative inputs.
496 *) Change BN_mod_mul so that the result is always non-negative.
497 Previously, it could be negative if one of the factors was negative;
498 I don't think anyone really wanted that behaviour.
501 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
502 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
503 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
504 and add new functions:
517 These functions always generate non-negative results.
519 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
520 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
522 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
523 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
525 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
527 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
528 was actually never needed) and in BN_mul(). The removal in BN_mul()
529 required a small change in bn_mul_part_recursive() and the addition
530 of the functions bn_cmp_part_words(), bn_sub_part_words() and
531 bn_add_part_words(), which do the same thing as bn_cmp_words(),
532 bn_sub_words() and bn_add_words() except they take arrays with
536 *) In 'openssl passwd', verify passwords read from the terminal
537 unless the '-salt' option is used (which usually means that
538 verification would just waste user's time since the resulting
539 hash is going to be compared with some given password hash)
540 or the new '-noverify' option is used.
542 This is an incompatible change, but it does not affect
543 non-interactive use of 'openssl passwd' (passwords on the command
544 line, '-stdin' option, '-in ...' option) and thus should not
548 *) Remove all references to RSAref, since there's no more need for it.
551 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
552 The previous value, 12, was not always sufficient for BN_mod_exp().
555 *) Make DSO load along a path given through an environment variable
556 (SHLIB_PATH) with shl_load().
559 *) Constify the ENGINE code as a result of BIGNUM constification.
560 Also constify the RSA code and most things related to it. In a
561 few places, most notable in the depth of the ASN.1 code, ugly
562 casts back to non-const were required (to be solved at a later
566 *) Make it so the openssl application has all engines loaded by default.
569 *) Constify the BIGNUM routines a little more.
572 *) Make sure that shared libraries get the internal name engine with
573 the full version number and not just 0. This should mark the
574 shared libraries as not backward compatible. Of course, this should
575 be changed again when we can guarantee backward binary compatibility.
578 *) Add the following functions:
584 ENGINE_load_builtin_engines()
586 That way, an application can itself choose if external engines that
587 are built-in in OpenSSL shall ever be used or not. The benefit is
588 that applications won't have to be linked with libdl or other dso
589 libraries unless it's really needed.
591 Changed 'openssl engine' to load all engines on demand.
592 Changed the engine header files to avoid the duplication of some
593 declarations (they differed!).
596 *) 'openssl engine' can now list capabilities.
599 *) Better error reporting in 'openssl engine'.
602 *) Never call load_dh_param(NULL) in s_server.
605 *) Add engine application. It can currently list engines by name and
606 identity, and test if they are actually available.
609 *) Add support for shared libraries under Irix.
610 [Albert Chin-A-Young <china@thewrittenword.com>]
612 *) Improve RPM specification file by forcing symbolic linking and making
613 sure the installed documentation is also owned by root.root.
614 [Damien Miller <djm@mindrot.org>]
616 *) Add configuration option to build on Linux on both big-endian and
618 [Ralf Baechle <ralf@uni-koblenz.de>]
620 *) Give the OpenSSL applications more possibilities to make use of
621 keys (public as well as private) handled by engines.
624 *) Add OCSP code that comes from CertCo.
627 *) Add VMS support for the Rijndael code.
630 *) Added untested support for Nuron crypto accelerator.
633 *) Add support for external cryptographic devices. This code was
634 previously distributed separately as the "engine" branch.
635 [Geoff Thorpe, Richard Levitte]
637 *) Rework the filename-translation in the DSO code. It is now possible to
638 have far greater control over how a "name" is turned into a filename
639 depending on the operating environment and any oddities about the
640 different shared library filenames on each system.
643 *) Support threads on FreeBSD-elf in Configure.
646 *) Add the possibility to create shared libraries on HP-UX
649 *) Fix for SHA1 assembly problem with MASM: it produces
650 warnings about corrupt line number information when assembling
651 with debugging information. This is caused by the overlapping
653 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
656 NCONF_get_number() has no error checking at all. As a replacement,
657 NCONF_get_number_e() is defined (_e for "error checking") and is
658 promoted strongly. The old NCONF_get_number is kept around for
659 binary backward compatibility.
660 Make it possible for methods to load from something other than a BIO,
661 by providing a function pointer that is given a name instead of a BIO.
662 For example, this could be used to load configuration data from an
666 *) Fix typo in get_cert_by_subject() in by_dir.c
667 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
669 *) Rework the system to generate shared libraries:
671 - Make note of the expected extension for the shared libraries and
672 if there is a need for symbolic links from for example libcrypto.so.0
673 to libcrypto.so.0.9.7. There is extended info in Configure for
676 - Make as few rebuilds of the shared libraries as possible.
678 - Still avoid linking the OpenSSL programs with the shared libraries.
680 - When installing, install the shared libraries separately from the
684 *) Fix for non blocking accept BIOs. Added new I/O special reason
685 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
686 with non blocking I/O was not possible because no retry code was
687 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
691 *) Added the beginnings of Rijndael support.
694 *) Fix for bug in DirectoryString mask setting. Add support for
695 X509_NAME_print_ex() in 'req' and X509_print_ex() function
696 to allow certificate printing to more controllable, additional
697 'certopt' option to 'x509' to allow new printing options to be
701 *) Clean old EAY MD5 hack from e_os.h.
704 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
706 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
707 and not in SSL_clear because the latter is also used by the
708 accept/connect functions; previously, the settings made by
709 SSL_set_read_ahead would be lost during the handshake.
710 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
712 *) Correct util/mkdef.pl to be selective about disabled algorithms.
713 Previously, it would create entries for disableed algorithms no
717 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
719 *) In ssl23_get_client_hello, generate an error message when faced
720 with an initial SSL 3.0/TLS record that is too small to contain the
721 first two bytes of the ClientHello message, i.e. client_version.
722 (Note that this is a pathologic case that probably has never happened
723 in real life.) The previous approach was to use the version number
724 from the record header as a substitute; but our protocol choice
725 should not depend on that one because it is not authenticated
726 by the Finished messages.
729 *) More robust randomness gathering functions for Windows.
730 [Jeffrey Altman <jaltman@columbia.edu>]
732 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
733 not set then we don't setup the error code for issuer check errors
734 to avoid possibly overwriting other errors which the callback does
735 handle. If an application does set the flag then we assume it knows
736 what it is doing and can handle the new informational codes
740 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
741 a general "ANY" type, as such it should be able to decode anything
742 including tagged types. However it didn't check the class so it would
743 wrongly interpret tagged types in the same way as their universal
744 counterpart and unknown types were just rejected. Changed so that the
745 tagged and unknown types are handled in the same way as a SEQUENCE:
746 that is the encoding is stored intact. There is also a new type
747 "V_ASN1_OTHER" which is used when the class is not universal, in this
748 case we have no idea what the actual type is so we just lump them all
752 *) On VMS, stdout may very well lead to a file that is written to
753 in a record-oriented fashion. That means that every write() will
754 write a separate record, which will be read separately by the
755 programs trying to read from it. This can be very confusing.
757 The solution is to put a BIO filter in the way that will buffer
758 text until a linefeed is reached, and then write everything a
759 line at a time, so every record written will be an actual line,
760 not chunks of lines and not (usually doesn't happen, but I've
761 seen it once) several lines in one record. BIO_f_linebuffer() is
764 Currently, it's a VMS-only method, because that's where it has
765 been tested well enough.
768 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
769 it can return incorrect results.
770 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
771 but it was in 0.9.6-beta[12].)
774 *) Disable the check for content being present when verifying detached
775 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
776 include zero length content when signing messages.
779 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
780 BIO_ctrl (for BIO pairs).
783 *) Add DSO method for VMS.
786 *) Bug fix: Montgomery multiplication could produce results with the
790 *) Add RPM specification openssl.spec and modify it to build three
791 packages. The default package contains applications, application
792 documentation and run-time libraries. The devel package contains
793 include files, static libraries and function documentation. The
794 doc package contains the contents of the doc directory. The original
795 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
798 *) Add a large number of documentation files for many SSL routines.
799 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
801 *) Add a configuration entry for Sony News 4.
802 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
804 *) Don't set the two most significant bits to one when generating a
805 random number < q in the DSA library.
808 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
809 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
810 the underlying transport is blocking) if a handshake took place.
811 (The default behaviour is needed by applications such as s_client
812 and s_server that use select() to determine when to use SSL_read;
813 but for applications that know in advance when to expect data, it
814 just makes things more complicated.)
817 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
821 *) Add a few more EBCDIC conditionals that make `req' and `x509'
822 work better on such systems.
823 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
825 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
826 Update PKCS12_parse() so it copies the friendlyName and the
827 keyid to the certificates aux info.
830 *) Fix bug in PKCS7_verify() which caused an infinite loop
831 if there was more than one signature.
832 [Sven Uszpelkat <su@celocom.de>]
834 *) Major change in util/mkdef.pl to include extra information
835 about each symbol, as well as presentig variables as well
836 as functions. This change means that there's n more need
837 to rebuild the .num files when some algorithms are excluded.
840 *) Allow the verify time to be set by an application,
841 rather than always using the current time.
844 *) Phase 2 verify code reorganisation. The certificate
845 verify code now looks up an issuer certificate by a
846 number of criteria: subject name, authority key id
847 and key usage. It also verifies self signed certificates
848 by the same criteria. The main comparison function is
849 X509_check_issued() which performs these checks.
851 Lot of changes were necessary in order to support this
852 without completely rewriting the lookup code.
854 Authority and subject key identifier are now cached.
856 The LHASH 'certs' is X509_STORE has now been replaced
857 by a STACK_OF(X509_OBJECT). This is mainly because an
858 LHASH can't store or retrieve multiple objects with
861 As a result various functions (which were all internal
862 use only) have changed to handle the new X509_STORE
863 structure. This will break anything that messed round
864 with X509_STORE internally.
866 The functions X509_STORE_add_cert() now checks for an
867 exact match, rather than just subject name.
869 The X509_STORE API doesn't directly support the retrieval
870 of multiple certificates matching a given criteria, however
871 this can be worked round by performing a lookup first
872 (which will fill the cache with candidate certificates)
873 and then examining the cache for matches. This is probably
874 the best we can do without throwing out X509_LOOKUP
875 entirely (maybe later...).
877 The X509_VERIFY_CTX structure has been enhanced considerably.
879 All certificate lookup operations now go via a get_issuer()
880 callback. Although this currently uses an X509_STORE it
881 can be replaced by custom lookups. This is a simple way
882 to bypass the X509_STORE hackery necessary to make this
883 work and makes it possible to use more efficient techniques
884 in future. A very simple version which uses a simple
885 STACK for its trusted certificate store is also provided
886 using X509_STORE_CTX_trusted_stack().
888 The verify_cb() and verify() callbacks now have equivalents
889 in the X509_STORE_CTX structure.
891 X509_STORE_CTX also has a 'flags' field which can be used
892 to customise the verify behaviour.
895 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
896 excludes S/MIME capabilities.
899 *) When a certificate request is read in keep a copy of the
900 original encoding of the signed data and use it when outputing
901 again. Signatures then use the original encoding rather than
902 a decoded, encoded version which may cause problems if the
903 request is improperly encoded.
906 *) For consistency with other BIO_puts implementations, call
907 buffer_write(b, ...) directly in buffer_puts instead of calling
910 In BIO_puts, increment b->num_write as in BIO_write.
911 [Peter.Sylvester@EdelWeb.fr]
913 *) Fix BN_mul_word for the case where the word is 0. (We have to use
914 BN_zero, we may not return a BIGNUM with an array consisting of
918 *) Avoid calling abort() from within the library when problems are
919 detected, except if preprocessor symbols have been defined
920 (such as REF_CHECK, BN_DEBUG etc.).
923 *) New openssl application 'rsautl'. This utility can be
924 used for low level RSA operations. DER public key
925 BIO/fp routines also added.
928 *) New Configure entry and patches for compiling on QNX 4.
929 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
931 *) A demo state-machine implementation was sponsored by
932 Nuron (http://www.nuron.com/) and is now available in
936 *) New options added to the 'dgst' utility for signature
937 generation and verification.
940 *) Unrecognized PKCS#7 content types are now handled via a
941 catch all ASN1_TYPE structure. This allows unsupported
942 types to be stored as a "blob" and an application can
943 encode and decode it manually.
946 *) Fix various signed/unsigned issues to make a_strex.c
948 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
950 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
951 length if passed a buffer. ASN1_INTEGER_to_BN failed
952 if passed a NULL BN and its argument was negative.
953 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
955 *) Modification to PKCS#7 encoding routines to output definite
956 length encoding. Since currently the whole structures are in
957 memory there's not real point in using indefinite length
958 constructed encoding. However if OpenSSL is compiled with
959 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
962 *) Added BIO_vprintf() and BIO_vsnprintf().
965 *) Added more prefixes to parse for in the the strings written
966 through a logging bio, to cover all the levels that are available
967 through syslog. The prefixes are now:
969 PANIC, EMERG, EMR => LOG_EMERG
970 ALERT, ALR => LOG_ALERT
971 CRIT, CRI => LOG_CRIT
972 ERROR, ERR => LOG_ERR
973 WARNING, WARN, WAR => LOG_WARNING
974 NOTICE, NOTE, NOT => LOG_NOTICE
975 INFO, INF => LOG_INFO
976 DEBUG, DBG => LOG_DEBUG
978 and as before, if none of those prefixes are present at the
979 beginning of the string, LOG_ERR is chosen.
981 On Win32, the LOG_* levels are mapped according to this:
983 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
984 LOG_WARNING => EVENTLOG_WARNING_TYPE
985 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
989 *) Made it possible to reconfigure with just the configuration
990 argument "reconf" or "reconfigure". The command line arguments
991 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
992 and are retrieved from there when reconfiguring.
996 [Assar Westerlund <assar@sics.se>, Richard Levitte]
998 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1001 *) The obj_dat.pl script was messing up the sorting of object
1002 names. The reason was that it compared the quoted version
1003 of strings as a result "OCSP" > "OCSP Signing" because
1004 " > SPACE. Changed script to store unquoted versions of
1005 names and add quotes on output. It was also omitting some
1006 names from the lookup table if they were given a default
1007 value (that is if SN is missing it is given the same
1008 value as LN and vice versa), these are now added on the
1009 grounds that if an object has a name we should be able to
1010 look it up. Finally added warning output when duplicate
1011 short or long names are found.
1014 *) Changes needed for Tandem NSK.
1015 [Scott Uroff <scott@xypro.com>]
1017 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1018 RSA_padding_check_SSLv23(), special padding was never detected
1019 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1020 version rollback attacks was not effective.
1022 In s23_clnt.c, don't use special rollback-attack detection padding
1023 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1024 client; similarly, in s23_srvr.c, don't do the rollback check if
1025 SSL 2.0 is the only protocol enabled in the server.
1028 *) Make it possible to get hexdumps of unprintable data with 'openssl
1029 asn1parse'. By implication, the functions ASN1_parse_dump() and
1030 BIO_dump_indent() are added.
1033 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1034 these print out strings and name structures based on various
1035 flags including RFC2253 support and proper handling of
1036 multibyte characters. Added options to the 'x509' utility
1037 to allow the various flags to be set.
1040 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1041 Also change the functions X509_cmp_current_time() and
1042 X509_gmtime_adj() work with an ASN1_TIME structure,
1043 this will enable certificates using GeneralizedTime in validity
1044 dates to be checked.
1047 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1048 negative public key encodings) on by default,
1049 NO_NEG_PUBKEY_BUG can be set to disable it.
1052 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1053 content octets. An i2c_ASN1_OBJECT is unnecessary because
1054 the encoding can be trivially obtained from the structure.
1057 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1058 not read locks (CRYPTO_r_[un]lock).
1061 *) A first attempt at creating official support for shared
1062 libraries through configuration. I've kept it so the
1063 default is static libraries only, and the OpenSSL programs
1064 are always statically linked for now, but there are
1065 preparations for dynamic linking in place.
1066 This has been tested on Linux and True64.
1069 *) Randomness polling function for Win9x, as described in:
1070 Peter Gutmann, Software Generation of Practically Strong
1074 *) Fix so PRNG is seeded in req if using an already existing
1078 *) New options to smime application. -inform and -outform
1079 allow alternative formats for the S/MIME message including
1080 PEM and DER. The -content option allows the content to be
1081 specified separately. This should allow things like Netscape
1082 form signing output easier to verify.
1085 *) Fix the ASN1 encoding of tags using the 'long form'.
1088 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1089 STRING types. These convert content octets to and from the
1090 underlying type. The actual tag and length octets are
1091 already assumed to have been read in and checked. These
1092 are needed because all other string types have virtually
1093 identical handling apart from the tag. By having versions
1094 of the ASN1 functions that just operate on content octets
1095 IMPLICIT tagging can be handled properly. It also allows
1096 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1097 and ASN1_INTEGER are identical apart from the tag.
1100 *) Change the handling of OID objects as follows:
1102 - New object identifiers are inserted in objects.txt, following
1103 the syntax given in objects.README.
1104 - objects.pl is used to process obj_mac.num and create a new
1106 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1109 This is currently kind of a hack, and the perl code in objects.pl
1110 isn't very elegant, but it works as I intended. The simplest way
1111 to check that it worked correctly is to look in obj_dat.h and
1112 check the array nid_objs and make sure the objects haven't moved
1113 around (this is important!). Additions are OK, as well as
1114 consistent name changes.
1117 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1120 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1121 The given file adds to whatever has already been seeded into the
1122 random pool through the RANDFILE configuration file option or
1123 environment variable, or the default random state file.
1126 *) mkstack.pl now sorts each macro group into lexical order.
1127 Previously the output order depended on the order the files
1128 appeared in the directory, resulting in needless rewriting
1132 *) Patches to make OpenSSL compile under Win32 again. Mostly
1133 work arounds for the VC++ problem that it treats func() as
1134 func(void). Also stripped out the parts of mkdef.pl that
1135 added extra typesafe functions: these no longer exist.
1138 *) Reorganisation of the stack code. The macros are now all
1139 collected in safestack.h . Each macro is defined in terms of
1140 a "stack macro" of the form SKM_<name>(type, a, b). The
1141 DEBUG_SAFESTACK is now handled in terms of function casts,
1142 this has the advantage of retaining type safety without the
1143 use of additional functions. If DEBUG_SAFESTACK is not defined
1144 then the non typesafe macros are used instead. Also modified the
1145 mkstack.pl script to handle the new form. Needs testing to see
1146 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1147 the default if no major problems. Similar behaviour for ASN1_SET_OF
1148 and PKCS12_STACK_OF.
1151 *) When some versions of IIS use the 'NET' form of private key the
1152 key derivation algorithm is different. Normally MD5(password) is
1153 used as a 128 bit RC4 key. In the modified case
1154 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1155 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1156 as the old Netscape_RSA functions except they have an additional
1157 'sgckey' parameter which uses the modified algorithm. Also added
1158 an -sgckey command line option to the rsa utility. Thanks to
1159 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1160 algorithm to openssl-dev.
1163 *) The evp_local.h macros were using 'c.##kname' which resulted in
1164 invalid expansion on some systems (SCO 5.0.5 for example).
1165 Corrected to 'c.kname'.
1166 [Phillip Porch <root@theporch.com>]
1168 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1169 a STACK of email addresses from a certificate or request, these look
1170 in the subject name and the subject alternative name extensions and
1171 omit any duplicate addresses.
1174 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1175 This makes DSA verification about 2 % faster.
1178 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1179 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1180 plus overhead for 1024 bit moduli).
1181 This makes exponentiations about 0.5 % faster for 1024 bit
1182 exponents (as measured by "openssl speed rsa2048").
1185 *) Rename memory handling macros to avoid conflicts with other
1187 Malloc => OPENSSL_malloc
1188 Malloc_locked => OPENSSL_malloc_locked
1189 Realloc => OPENSSL_realloc
1190 Free => OPENSSL_free
1193 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1194 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1197 *) CygWin32 support.
1198 [John Jarvie <jjarvie@newsguy.com>]
1200 *) The type-safe stack code has been rejigged. It is now only compiled
1201 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1202 by default all type-specific stack functions are "#define"d back to
1203 standard stack functions. This results in more streamlined output
1204 but retains the type-safety checking possibilities of the original
1208 *) The STACK code has been cleaned up, and certain type declarations
1209 that didn't make a lot of sense have been brought in line. This has
1210 also involved a cleanup of sorts in safestack.h to more correctly
1211 map type-safe stack functions onto their plain stack counterparts.
1212 This work has also resulted in a variety of "const"ifications of
1213 lots of the code, especially "_cmp" operations which should normally
1214 be prototyped with "const" parameters anyway.
1217 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1218 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1219 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1220 where all of 'md' is used each time the PRNG is used, but 'state'
1221 is used only indexed by a cyclic counter. As entropy may not be
1222 well distributed from the beginning, 'md' is important as a
1223 chaining variable. However, the output function chains only half
1224 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1225 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1226 in all of 'state' being rewritten, with the new values depending
1227 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1230 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1231 the handshake is continued after ssl_verify_cert_chain();
1232 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1233 can lead to 'unexplainable' connection aborts later.
1234 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1236 *) Major EVP API cipher revision.
1237 Add hooks for extra EVP features. This allows various cipher
1238 parameters to be set in the EVP interface. Support added for variable
1239 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1240 setting of RC2 and RC5 parameters.
1242 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1245 Remove lots of duplicated code from the EVP library. For example *every*
1246 cipher init() function handles the 'iv' in the same way according to the
1247 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1248 for CFB and OFB modes they zero ctx->num.
1250 New functionality allows removal of S/MIME code RC2 hack.
1252 Most of the routines have the same form and so can be declared in terms
1255 By shifting this to the top level EVP_CipherInit() it can be removed from
1256 all individual ciphers. If the cipher wants to handle IVs or keys
1257 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1260 Change lots of functions like EVP_EncryptUpdate() to now return a
1261 value: although software versions of the algorithms cannot fail
1262 any installed hardware versions can.
1265 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1266 this option is set, tolerate broken clients that send the negotiated
1267 protocol version number instead of the requested protocol version
1271 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1272 i.e. non-zero for export ciphersuites, zero otherwise.
1273 Previous versions had this flag inverted, inconsistent with
1274 rsa_tmp_cb (..._TMP_RSA_CB).
1275 [Bodo Moeller; problem reported by Amit Chopra]
1277 *) Add missing DSA library text string. Work around for some IIS
1278 key files with invalid SEQUENCE encoding.
1281 *) Add a document (doc/standards.txt) that list all kinds of standards
1282 and so on that are implemented in OpenSSL.
1285 *) Enhance c_rehash script. Old version would mishandle certificates
1286 with the same subject name hash and wouldn't handle CRLs at all.
1287 Added -fingerprint option to crl utility, to support new c_rehash
1291 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1294 *) Fix for SSL server purpose checking. Server checking was
1295 rejecting certificates which had extended key usage present
1296 but no ssl client purpose.
1297 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1299 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1300 is a little unclear about how a blank password is handled.
1301 Since the password in encoded as a BMPString with terminating
1302 double NULL a zero length password would end up as just the
1303 double NULL. However no password at all is different and is
1304 handled differently in the PKCS#12 key generation code. NS
1305 treats a blank password as zero length. MSIE treats it as no
1306 password on export: but it will try both on import. We now do
1307 the same: PKCS12_parse() tries zero length and no password if
1308 the password is set to "" or NULL (NULL is now a valid password:
1309 it wasn't before) as does the pkcs12 application.
1312 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1313 perror when PEM_read_bio_X509_REQ fails, the error message must
1314 be obtained from the error queue.
1317 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1318 it in ERR_remove_state if appropriate, and change ERR_get_state
1319 accordingly to avoid race conditions (this is necessary because
1320 thread_hash is no longer constant once set).
1323 *) Bugfix for linux-elf makefile.one.
1326 *) RSA_get_default_method() will now cause a default
1327 RSA_METHOD to be chosen if one doesn't exist already.
1328 Previously this was only set during a call to RSA_new()
1329 or RSA_new_method(NULL) meaning it was possible for
1330 RSA_get_default_method() to return NULL.
1333 *) Added native name translation to the existing DSO code
1334 that will convert (if the flag to do so is set) filenames
1335 that are sufficiently small and have no path information
1336 into a canonical native form. Eg. "blah" converted to
1337 "libblah.so" or "blah.dll" etc.
1340 *) New function ERR_error_string_n(e, buf, len) which is like
1341 ERR_error_string(e, buf), but writes at most 'len' bytes
1342 including the 0 terminator. For ERR_error_string_n, 'buf'
1344 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1346 *) CONF library reworked to become more general. A new CONF
1347 configuration file reader "class" is implemented as well as a
1348 new functions (NCONF_*, for "New CONF") to handle it. The now
1349 old CONF_* functions are still there, but are reimplemented to
1350 work in terms of the new functions. Also, a set of functions
1351 to handle the internal storage of the configuration data is
1352 provided to make it easier to write new configuration file
1353 reader "classes" (I can definitely see something reading a
1354 configuration file in XML format, for example), called _CONF_*,
1355 or "the configuration storage API"...
1357 The new configuration file reading functions are:
1359 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1360 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1362 NCONF_default, NCONF_WIN32
1364 NCONF_dump_fp, NCONF_dump_bio
1366 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1367 NCONF_new creates a new CONF object. This works in the same way
1368 as other interfaces in OpenSSL, like the BIO interface.
1369 NCONF_dump_* dump the internal storage of the configuration file,
1370 which is useful for debugging. All other functions take the same
1371 arguments as the old CONF_* functions wth the exception of the
1372 first that must be a `CONF *' instead of a `LHASH *'.
1374 To make it easer to use the new classes with the old CONF_* functions,
1375 the function CONF_set_default_method is provided.
1378 *) Add '-tls1' option to 'openssl ciphers', which was already
1379 mentioned in the documentation but had not been implemented.
1380 (This option is not yet really useful because even the additional
1381 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1384 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1385 OpenSSL-based applications) load shared libraries and bind to
1386 them in a portable way.
1387 [Geoff Thorpe, with contributions from Richard Levitte]
1389 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1391 *) Make sure _lrotl and _lrotr are only used with MSVC.
1393 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1394 (the default implementation of RAND_status).
1396 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1397 to '-clrext' (= clear extensions), as intended and documented.
1398 [Bodo Moeller; inconsistency pointed out by Michael Attili
1401 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1402 was larger than the MD block size.
1403 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1405 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1406 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1407 using the passed key: if the passed key was a private key the result
1408 of X509_print(), for example, would be to print out all the private key
1412 *) des_quad_cksum() byte order bug fix.
1413 [Ulf Möller, using the problem description in krb4-0.9.7, where
1414 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1416 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1418 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1420 *) For easily testing in shell scripts whether some command
1421 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1422 returns with exit code 0 iff no command of the given name is available.
1423 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1424 the output goes to stdout and nothing is printed to stderr.
1425 Additional arguments are always ignored.
1427 Since for each cipher there is a command of the same name,
1428 the 'no-cipher' compilation switches can be tested this way.
1430 ('openssl no-XXX' is not able to detect pseudo-commands such
1431 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1434 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1437 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1438 is set; it will be thrown away anyway because each handshake creates
1440 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1441 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1442 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1443 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1446 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1447 'Q' and 'R' lose their special meanings (quit/renegotiate).
1448 This is part of what -quiet does; unlike -quiet, -ign_eof
1449 does not suppress any output.
1452 *) Add compatibility options to the purpose and trust code. The
1453 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1454 accepts a certificate or CA, this was the previous behaviour,
1455 with all the associated security issues.
1457 X509_TRUST_COMPAT is the old trust behaviour: only and
1458 automatically trust self signed roots in certificate store. A
1459 new trust setting X509_TRUST_DEFAULT is used to specify that
1460 a purpose has no associated trust setting and it should instead
1461 use the value in the default purpose.
1464 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1465 and fix a memory leak.
1468 *) In util/mkerr.pl (which implements 'make errors'), preserve
1469 reason strings from the previous version of the .c file, as
1470 the default to have only downcase letters (and digits) in
1471 automatically generated reasons codes is not always appropriate.
1474 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1475 using strerror. Previously, ERR_reason_error_string() returned
1476 library names as reason strings for SYSerr; but SYSerr is a special
1477 case where small numbers are errno values, not library numbers.
1480 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1481 converts DSA parameters into DH parameters. (When creating parameters,
1482 DSA_generate_parameters is used.)
1485 *) Include 'length' (recommended exponent length) in C code generated
1486 by 'openssl dhparam -C'.
1489 *) The second argument to set_label in perlasm was already being used
1490 so couldn't be used as a "file scope" flag. Moved to third argument
1494 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1495 instead of RAND_bytes for encryption IVs and salts.
1498 *) Include RAND_status() into RAND_METHOD instead of implementing
1499 it only for md_rand.c Otherwise replacing the PRNG by calling
1500 RAND_set_rand_method would be impossible.
1503 *) Don't let DSA_generate_key() enter an infinite loop if the random
1504 number generation fails.
1507 *) New 'rand' application for creating pseudo-random output.
1510 *) Added configuration support for Linux/IA64
1511 [Rolf Haberrecker <rolf@suse.de>]
1513 *) Assembler module support for Mingw32.
1516 *) Shared library support for HPUX (in shlib/).
1517 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1519 *) Shared library support for Solaris gcc.
1520 [Lutz Behnke <behnke@trustcenter.de>]
1522 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1524 *) PKCS7_encrypt() was adding text MIME headers twice because they
1525 were added manually and by SMIME_crlf_copy().
1528 *) In bntest.c don't call BN_rand with zero bits argument.
1529 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1531 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1532 case was implemented. This caused BN_div_recp() to fail occasionally.
1535 *) Add an optional second argument to the set_label() in the perl
1536 assembly language builder. If this argument exists and is set
1537 to 1 it signals that the assembler should use a symbol whose
1538 scope is the entire file, not just the current function. This
1539 is needed with MASM which uses the format label:: for this scope.
1540 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1542 *) Change the ASN1 types so they are typedefs by default. Before
1543 almost all types were #define'd to ASN1_STRING which was causing
1544 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1548 *) Change names of new functions to the new get1/get0 naming
1549 convention: After 'get1', the caller owns a reference count
1550 and has to call ..._free; 'get0' returns a pointer to some
1551 data structure without incrementing reference counters.
1552 (Some of the existing 'get' functions increment a reference
1553 counter, some don't.)
1554 Similarly, 'set1' and 'add1' functions increase reference
1555 counters or duplicate objects.
1558 *) Allow for the possibility of temp RSA key generation failure:
1559 the code used to assume it always worked and crashed on failure.
1562 *) Fix potential buffer overrun problem in BIO_printf().
1563 [Ulf Möller, using public domain code by Patrick Powell; problem
1564 pointed out by David Sacerdote <das33@cornell.edu>]
1566 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1567 RAND_egd() and RAND_status(). In the command line application,
1568 the EGD socket can be specified like a seed file using RANDFILE
1572 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1573 Some CAs (e.g. Verisign) distribute certificates in this form.
1576 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1577 list to exclude them. This means that no special compilation option
1578 is needed to use anonymous DH: it just needs to be included in the
1582 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1583 EVP_MD_type. The old functionality is available in a new macro called
1584 EVP_MD_md(). Change code that uses it and update docs.
1587 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1588 where the 'void *' argument is replaced by a function pointer argument.
1589 Previously 'void *' was abused to point to functions, which works on
1590 many platforms, but is not correct. As these functions are usually
1591 called by macros defined in OpenSSL header files, most source code
1592 should work without changes.
1595 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1596 sections with information on -D... compiler switches used for
1597 compiling the library so that applications can see them. To enable
1598 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1599 must be defined. E.g.,
1600 #define OPENSSL_ALGORITHM_DEFINES
1601 #include <openssl/opensslconf.h>
1602 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1603 [Richard Levitte, Ulf and Bodo Möller]
1605 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1609 *) Change the 'other' type in certificate aux info to a STACK_OF
1610 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1611 the required ASN1 format: arbitrary types determined by an OID.
1614 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1615 argument to 'req'. This is not because the function is newer or
1616 better than others it just uses the work 'NEW' in the certificate
1617 request header lines. Some software needs this.
1620 *) Reorganise password command line arguments: now passwords can be
1621 obtained from various sources. Delete the PEM_cb function and make
1622 it the default behaviour: i.e. if the callback is NULL and the
1623 usrdata argument is not NULL interpret it as a null terminated pass
1624 phrase. If usrdata and the callback are NULL then the pass phrase
1625 is prompted for as usual.
1628 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1629 the support is automatically enabled. The resulting binaries will
1630 autodetect the card and use it if present.
1631 [Ben Laurie and Compaq Inc.]
1633 *) Work around for Netscape hang bug. This sends certificate request
1634 and server done in one record. Since this is perfectly legal in the
1635 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1636 the bugs/SSLv3 entry for more info.
1639 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1642 *) Add -rand argument to smime and pkcs12 applications and read/write
1646 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1649 *) Add command line password options to the remaining applications.
1652 *) Bug fix for BN_div_recp() for numerators with an even number of
1656 *) More tests in bntest.c, and changed test_bn output.
1659 *) ./config recognizes MacOS X now.
1662 *) Bug fix for BN_div() when the first words of num and divsor are
1663 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1666 *) Add support for various broken PKCS#8 formats, and command line
1667 options to produce them.
1670 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1671 get temporary BIGNUMs from a BN_CTX.
1674 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1678 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1679 include a #define from the old name to the new. The original intent
1680 was that statically linked binaries could for example just call
1681 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1682 link with digests. This never worked becayse SSLeay_add_all_digests()
1683 and SSLeay_add_all_ciphers() were in the same source file so calling
1684 one would link with the other. They are now in separate source files.
1687 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1690 *) Use a less unusual form of the Miller-Rabin primality test (it used
1691 a binary algorithm for exponentiation integrated into the Miller-Rabin
1692 loop, our standard modexp algorithms are faster).
1695 *) Support for the EBCDIC character set completed.
1696 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1698 *) Source code cleanups: use const where appropriate, eliminate casts,
1699 use void * instead of char * in lhash.
1702 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1703 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1704 this the server could overwrite ephemeral keys that the client
1708 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1709 using 50 iterations of the Rabin-Miller test.
1711 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1712 iterations of the Rabin-Miller test as required by the appendix
1713 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1714 As BN_is_prime_fasttest includes trial division, DSA parameter
1715 generation becomes much faster.
1717 This implies a change for the callback functions in DSA_is_prime
1718 and DSA_generate_parameters: The callback function is called once
1719 for each positive witness in the Rabin-Miller test, not just
1720 occasionally in the inner loop; and the parameters to the
1721 callback function now provide an iteration count for the outer
1722 loop rather than for the current invocation of the inner loop.
1723 DSA_generate_parameters additionally can call the callback
1724 function with an 'iteration count' of -1, meaning that a
1725 candidate has passed the trial division test (when q is generated
1726 from an application-provided seed, trial division is skipped).
1729 *) New function BN_is_prime_fasttest that optionally does trial
1730 division before starting the Rabin-Miller test and has
1731 an additional BN_CTX * argument (whereas BN_is_prime always
1732 has to allocate at least one BN_CTX).
1733 'callback(1, -1, cb_arg)' is called when a number has passed the
1734 trial division stage.
1737 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1741 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1744 *) New function BN_pseudo_rand().
1747 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1748 bignum version of BN_from_montgomery() with the working code from
1749 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1753 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1754 made it impossible to use the same SSL_SESSION data structure in
1755 SSL2 clients in multiple threads.
1758 *) The return value of RAND_load_file() no longer counts bytes obtained
1759 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1760 to seed the PRNG (previously an explicit byte count was required).
1761 [Ulf Möller, Bodo Möller]
1763 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1764 used (char *) instead of (void *) and had casts all over the place.
1767 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1770 *) Retain source code compatibility for BN_prime_checks macro:
1771 BN_is_prime(..., BN_prime_checks, ...) now uses
1772 BN_prime_checks_for_size to determine the appropriate number of
1773 Rabin-Miller iterations.
1776 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1777 DH_CHECK_P_NOT_SAFE_PRIME.
1778 (Check if this is true? OpenPGP calls them "strong".)
1781 *) Merge the functionality of "dh" and "gendh" programs into a new program
1782 "dhparam". The old programs are retained for now but will handle DH keys
1783 (instead of parameters) in future.
1786 *) Make the ciphers, s_server and s_client programs check the return values
1787 when a new cipher list is set.
1790 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1791 ciphers. Before when the 56bit ciphers were enabled the sorting was
1794 The syntax for the cipher sorting has been extended to support sorting by
1795 cipher-strength (using the strength_bits hard coded in the tables).
1796 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1798 Fix a bug in the cipher-command parser: when supplying a cipher command
1799 string with an "undefined" symbol (neither command nor alphanumeric
1800 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1801 an error is flagged.
1803 Due to the strength-sorting extension, the code of the
1804 ssl_create_cipher_list() function was completely rearranged. I hope that
1805 the readability was also increased :-)
1806 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1808 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1809 for the first serial number and places 2 in the serial number file. This
1810 avoids problems when the root CA is created with serial number zero and
1811 the first user certificate has the same issuer name and serial number
1815 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1816 the new code. Add documentation for this stuff.
1819 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1820 X509_*() to X509at_*() on the grounds that they don't handle X509
1821 structures and behave in an analagous way to the X509v3 functions:
1822 they shouldn't be called directly but wrapper functions should be used
1825 So we also now have some wrapper functions that call the X509at functions
1826 when passed certificate requests. (TO DO: similar things can be done with
1827 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1828 things. Some of these need some d2i or i2d and print functionality
1829 because they handle more complex structures.)
1832 *) Add missing #ifndefs that caused missing symbols when building libssl
1833 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1834 NO_RSA in ssl/s2*.c.
1835 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1837 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1838 has a return value which indicates the quality of the random data
1839 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1840 error queue. New function RAND_pseudo_bytes() generates output that is
1841 guaranteed to be unique but not unpredictable. RAND_add is like
1842 RAND_seed, but takes an extra argument for an entropy estimate
1843 (RAND_seed always assumes full entropy).
1846 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1847 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1848 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1849 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1850 false-positive rate of at most 2^-80 for random input.
1853 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1856 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1857 in the 0.9.5 release), this returns the chain
1858 from an X509_CTX structure with a dup of the stack and all
1859 the X509 reference counts upped: so the stack will exist
1860 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1863 Also make SSL_SESSION_print() print out the verify return
1867 *) Add manpage for the pkcs12 command. Also change the default
1868 behaviour so MAC iteration counts are used unless the new
1869 -nomaciter option is used. This improves file security and
1870 only older versions of MSIE (4.0 for example) need it.
1873 *) Honor the no-xxx Configure options when creating .DEF files.
1876 *) Add PKCS#10 attributes to field table: challengePassword,
1877 unstructuredName and unstructuredAddress. These are taken from
1878 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1879 international characters are used.
1881 More changes to X509_ATTRIBUTE code: allow the setting of types
1882 based on strings. Remove the 'loc' parameter when adding
1883 attributes because these will be a SET OF encoding which is sorted
1887 *) Initial changes to the 'req' utility to allow request generation
1888 automation. This will allow an application to just generate a template
1889 file containing all the field values and have req construct the
1892 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1893 used all over the place including certificate requests and PKCS#7
1894 structures. They are currently handled manually where necessary with
1895 some primitive wrappers for PKCS#7. The new functions behave in a
1896 manner analogous to the X509 extension functions: they allow
1897 attributes to be looked up by NID and added.
1899 Later something similar to the X509V3 code would be desirable to
1900 automatically handle the encoding, decoding and printing of the
1901 more complex types. The string types like challengePassword can
1902 be handled by the string table functions.
1904 Also modified the multi byte string table handling. Now there is
1905 a 'global mask' which masks out certain types. The table itself
1906 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1907 is useful when for example there is only one permissible type
1908 (as in countryName) and using the mask might result in no valid
1912 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1913 SSL_get_peer_finished to allow applications to obtain the latest
1914 Finished messages sent to the peer or expected from the peer,
1915 respectively. (SSL_get_peer_finished is usually the Finished message
1916 actually received from the peer, otherwise the protocol will be aborted.)
1918 As the Finished message are message digests of the complete handshake
1919 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1920 be used for external authentication procedures when the authentication
1921 provided by SSL/TLS is not desired or is not enough.
1924 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1925 the host supports BWX extension and if Compaq C is present on the
1926 $PATH. Just exploiting of the BWX extension results in 20-30%
1927 performance kick for some algorithms, e.g. DES and RC4 to mention
1928 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1932 *) Add support for MS "fast SGC". This is arguably a violation of the
1933 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1934 weak crypto and after checking the certificate is SGC a second one
1935 with strong crypto. MS SGC stops the first handshake after receiving
1936 the server certificate message and sends a second client hello. Since
1937 a server will typically do all the time consuming operations before
1938 expecting any further messages from the client (server key exchange
1939 is the most expensive) there is little difference between the two.
1941 To get OpenSSL to support MS SGC we have to permit a second client
1942 hello message after we have sent server done. In addition we have to
1943 reset the MAC if we do get this second client hello.
1946 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1947 if a DER encoded private key is RSA or DSA traditional format. Changed
1948 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1949 format DER encoded private key. Newer code should use PKCS#8 format which
1950 has the key type encoded in the ASN1 structure. Added DER private key
1951 support to pkcs8 application.
1954 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1955 ciphersuites has been selected (as required by the SSL 3/TLS 1
1956 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1957 is set, we interpret this as a request to violate the specification
1958 (the worst that can happen is a handshake failure, and 'correct'
1959 behaviour would result in a handshake failure anyway).
1962 *) In SSL_CTX_add_session, take into account that there might be multiple
1963 SSL_SESSION structures with the same session ID (e.g. when two threads
1964 concurrently obtain them from an external cache).
1965 The internal cache can handle only one SSL_SESSION with a given ID,
1966 so if there's a conflict, we now throw out the old one to achieve
1970 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1971 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1972 some routines that use cipher OIDs: some ciphers do not have OIDs
1973 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1977 *) Simplify the trust setting structure and code. Now we just have
1978 two sequences of OIDs for trusted and rejected settings. These will
1979 typically have values the same as the extended key usage extension
1980 and any application specific purposes.
1982 The trust checking code now has a default behaviour: it will just
1983 check for an object with the same NID as the passed id. Functions can
1984 be provided to override either the default behaviour or the behaviour
1985 for a given id. SSL client, server and email already have functions
1986 in place for compatibility: they check the NID and also return "trusted"
1987 if the certificate is self signed.
1990 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1991 traditional format into an EVP_PKEY structure.
1994 *) Add a password callback function PEM_cb() which either prompts for
1995 a password if usr_data is NULL or otherwise assumes it is a null
1996 terminated password. Allow passwords to be passed on command line
1997 environment or config files in a few more utilities.
2000 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2001 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2002 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2003 Update documentation.
2006 *) Support for ASN1 "NULL" type. This could be handled before by using
2007 ASN1_TYPE but there wasn't any function that would try to read a NULL
2008 and produce an error if it couldn't. For compatibility we also have
2009 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2010 don't allocate anything because they don't need to.
2013 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2015 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2017 *) Rebuild of the memory allocation routines used by OpenSSL code and
2018 possibly others as well. The purpose is to make an interface that
2019 provide hooks so anyone can build a separate set of allocation and
2020 deallocation routines to be used by OpenSSL, for example memory
2021 pool implementations, or something else, which was previously hard
2022 since Malloc(), Realloc() and Free() were defined as macros having
2023 the values malloc, realloc and free, respectively (except for Win32
2024 compilations). The same is provided for memory debugging code.
2025 OpenSSL already comes with functionality to find memory leaks, but
2026 this gives people a chance to debug other memory problems.
2028 With these changes, a new set of functions and macros have appeared:
2030 CRYPTO_set_mem_debug_functions() [F]
2031 CRYPTO_get_mem_debug_functions() [F]
2032 CRYPTO_dbg_set_options() [F]
2033 CRYPTO_dbg_get_options() [F]
2034 CRYPTO_malloc_debug_init() [M]
2036 The memory debug functions are NULL by default, unless the library
2037 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2038 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2039 gives the standard debugging functions that come with OpenSSL) or
2040 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2041 provided by the library user) must be used. When the standard
2042 debugging functions are used, CRYPTO_dbg_set_options can be used to
2043 request additional information:
2044 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2045 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2047 Also, things like CRYPTO_set_mem_functions will always give the
2048 expected result (the new set of functions is used for allocation
2049 and deallocation) at all times, regardless of platform and compiler
2052 To finish it up, some functions that were never use in any other
2053 way than through macros have a new API and new semantic:
2056 CRYPTO_dbg_realloc()
2059 All macros of value have retained their old syntax.
2060 [Richard Levitte and Bodo Moeller]
2062 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2063 ordering of SMIMECapabilities wasn't in "strength order" and there
2064 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2068 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2069 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2070 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2072 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2073 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2074 functionality to handle multipart/signed properly) and a utility
2075 called 'smime' to call all this stuff. This is based on code I
2076 originally wrote for Celo who have kindly allowed it to be
2077 included in OpenSSL.
2080 *) Add variants des_set_key_checked and des_set_key_unchecked of
2081 des_set_key (aka des_key_sched). Global variable des_check_key
2082 decides which of these is called by des_set_key; this way
2083 des_check_key behaves as it always did, but applications and
2084 the library itself, which was buggy for des_check_key == 1,
2085 have a cleaner way to pick the version they need.
2088 *) New function PKCS12_newpass() which changes the password of a
2092 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2093 dynamic mix. In both cases the ids can be used as an index into the
2094 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2095 functions so they accept a list of the field values and the
2096 application doesn't need to directly manipulate the X509_TRUST
2100 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2104 *) Modify the way the V3 extension code looks up extensions. This now
2105 works in a similar way to the object code: we have some "standard"
2106 extensions in a static table which is searched with OBJ_bsearch()
2107 and the application can add dynamic ones if needed. The file
2108 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2109 updated whenever a new extension is added to the core code and kept
2110 in ext_nid order. There is a simple program 'tabtest.c' which checks
2111 this. New extensions are not added too often so this file can readily
2112 be maintained manually.
2114 There are two big advantages in doing things this way. The extensions
2115 can be looked up immediately and no longer need to be "added" using
2116 X509V3_add_standard_extensions(): this function now does nothing.
2117 [Side note: I get *lots* of email saying the extension code doesn't
2118 work because people forget to call this function]
2119 Also no dynamic allocation is done unless new extensions are added:
2120 so if we don't add custom extensions there is no need to call
2121 X509V3_EXT_cleanup().
2124 *) Modify enc utility's salting as follows: make salting the default. Add a
2125 magic header, so unsalted files fail gracefully instead of just decrypting
2126 to garbage. This is because not salting is a big security hole, so people
2127 should be discouraged from doing it.
2130 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2131 digest to be passed on the command line but it only used this
2132 parameter when signing a certificate. Modified so all relevant
2133 operations are affected by the digest parameter including the
2134 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2135 DSA key was used because it didn't fix the digest.
2138 *) Initial certificate chain verify code. Currently tests the untrusted
2139 certificates for consistency with the verify purpose (which is set
2140 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2142 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2143 this is because it will reject chains with invalid extensions whereas
2144 every previous version of OpenSSL and SSLeay made no checks at all.
2146 Trust code: checks the root CA for the relevant trust settings. Trust
2147 settings have an initial value consistent with the verify purpose: e.g.
2148 if the verify purpose is for SSL client use it expects the CA to be
2149 trusted for SSL client use. However the default value can be changed to
2150 permit custom trust settings: one example of this would be to only trust
2151 certificates from a specific "secure" set of CAs.
2153 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2154 which should be used for version portability: especially since the
2155 verify structure is likely to change more often now.
2157 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2158 to set them. If not set then assume SSL clients will verify SSL servers
2161 Two new options to the verify program: -untrusted allows a set of
2162 untrusted certificates to be passed in and -purpose which sets the
2163 intended purpose of the certificate. If a purpose is set then the
2164 new chain verify code is used to check extension consistency.
2167 *) Support for the authority information access extension.
2170 *) Modify RSA and DSA PEM read routines to transparently handle
2171 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2172 public keys in a format compatible with certificate
2173 SubjectPublicKeyInfo structures. Unfortunately there were already
2174 functions called *_PublicKey_* which used various odd formats so
2175 these are retained for compatibility: however the DSA variants were
2176 never in a public release so they have been deleted. Changed dsa/rsa
2177 utilities to handle the new format: note no releases ever handled public
2178 keys so we should be OK.
2180 The primary motivation for this change is to avoid the same fiasco
2181 that dogs private keys: there are several incompatible private key
2182 formats some of which are standard and some OpenSSL specific and
2183 require various evil hacks to allow partial transparent handling and
2184 even then it doesn't work with DER formats. Given the option anything
2185 other than PKCS#8 should be dumped: but the other formats have to
2186 stay in the name of compatibility.
2188 With public keys and the benefit of hindsight one standard format
2189 is used which works with EVP_PKEY, RSA or DSA structures: though
2190 it clearly returns an error if you try to read the wrong kind of key.
2192 Added a -pubkey option to the 'x509' utility to output the public key.
2193 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2194 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2195 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2196 that do the same as the EVP_PKEY_assign_*() except they up the
2197 reference count of the added key (they don't "swallow" the
2201 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2202 CRLs would fail if the file contained no certificates or no CRLs:
2203 added a new function to read in both types and return the number
2204 read: this means that if none are read it will be an error. The
2205 DER versions of the certificate and CRL reader would always fail
2206 because it isn't possible to mix certificates and CRLs in DER format
2207 without choking one or the other routine. Changed this to just read
2208 a certificate: this is the best we can do. Also modified the code
2209 in apps/verify.c to take notice of return codes: it was previously
2210 attempting to read in certificates from NULL pointers and ignoring
2211 any errors: this is one reason why the cert and CRL reader seemed
2212 to work. It doesn't check return codes from the default certificate
2213 routines: these may well fail if the certificates aren't installed.
2216 *) Code to support otherName option in GeneralName.
2219 *) First update to verify code. Change the verify utility
2220 so it warns if it is passed a self signed certificate:
2221 for consistency with the normal behaviour. X509_verify
2222 has been modified to it will now verify a self signed
2223 certificate if *exactly* the same certificate appears
2224 in the store: it was previously impossible to trust a
2225 single self signed certificate. This means that:
2226 openssl verify ss.pem
2227 now gives a warning about a self signed certificate but
2228 openssl verify -CAfile ss.pem ss.pem
2232 *) For servers, store verify_result in SSL_SESSION data structure
2233 (and add it to external session representation).
2234 This is needed when client certificate verifications fails,
2235 but an application-provided verification callback (set by
2236 SSL_CTX_set_cert_verify_callback) allows accepting the session
2237 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2238 but returns 1): When the session is reused, we have to set
2239 ssl->verify_result to the appropriate error code to avoid
2241 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2243 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2244 case in PKCS7_dataInit() where the signed PKCS7 structure
2245 didn't contain any existing data because it was being created.
2246 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2248 *) Add a salt to the key derivation routines in enc.c. This
2249 forms the first 8 bytes of the encrypted file. Also add a
2250 -S option to allow a salt to be input on the command line.
2253 *) New function X509_cmp(). Oddly enough there wasn't a function
2254 to compare two certificates. We do this by working out the SHA1
2255 hash and comparing that. X509_cmp() will be needed by the trust
2259 *) SSL_get1_session() is like SSL_get_session(), but increments
2260 the reference count in the SSL_SESSION returned.
2261 [Geoff Thorpe <geoff@eu.c2.net>]
2263 *) Fix for 'req': it was adding a null to request attributes.
2264 Also change the X509_LOOKUP and X509_INFO code to handle
2265 certificate auxiliary information.
2268 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2272 *) Add the possibility to add extra information to the memory leak
2273 detecting output, to form tracebacks, showing from where each
2274 allocation was originated: CRYPTO_push_info("constant string") adds
2275 the string plus current file name and line number to a per-thread
2276 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2277 is like calling CYRPTO_pop_info() until the stack is empty.
2278 Also updated memory leak detection code to be multi-thread-safe.
2281 *) Add options -text and -noout to pkcs7 utility and delete the
2282 encryption options which never did anything. Update docs.
2285 *) Add options to some of the utilities to allow the pass phrase
2286 to be included on either the command line (not recommended on
2287 OSes like Unix) or read from the environment. Update the
2288 manpages and fix a few bugs.
2291 *) Add a few manpages for some of the openssl commands.
2294 *) Fix the -revoke option in ca. It was freeing up memory twice,
2295 leaking and not finding already revoked certificates.
2298 *) Extensive changes to support certificate auxiliary information.
2299 This involves the use of X509_CERT_AUX structure and X509_AUX
2300 functions. An X509_AUX function such as PEM_read_X509_AUX()
2301 can still read in a certificate file in the usual way but it
2302 will also read in any additional "auxiliary information". By
2303 doing things this way a fair degree of compatibility can be
2304 retained: existing certificates can have this information added
2305 using the new 'x509' options.
2307 Current auxiliary information includes an "alias" and some trust
2308 settings. The trust settings will ultimately be used in enhanced
2309 certificate chain verification routines: currently a certificate
2310 can only be trusted if it is self signed and then it is trusted
2314 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2315 The problem was that one of the replacement routines had not been working
2316 since SSLeay releases. For now the offending routine has been replaced
2317 with non-optimised assembler. Even so, this now gives around 95%
2318 performance improvement for 1024 bit RSA signs.
2321 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2322 handling. Most clients have the effective key size in bits equal to
2323 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2324 A few however don't do this and instead use the size of the decrypted key
2325 to determine the RC2 key length and the AlgorithmIdentifier to determine
2326 the effective key length. In this case the effective key length can still
2327 be 40 bits but the key length can be 168 bits for example. This is fixed
2328 by manually forcing an RC2 key into the EVP_PKEY structure because the
2329 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2330 the key length and effective key length are equal.
2333 *) Add a bunch of functions that should simplify the creation of
2334 X509_NAME structures. Now you should be able to do:
2335 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2336 and have it automatically work out the correct field type and fill in
2337 the structures. The more adventurous can try:
2338 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2339 and it will (hopefully) work out the correct multibyte encoding.
2342 *) Change the 'req' utility to use the new field handling and multibyte
2343 copy routines. Before the DN field creation was handled in an ad hoc
2344 way in req, ca, and x509 which was rather broken and didn't support
2345 BMPStrings or UTF8Strings. Since some software doesn't implement
2346 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2347 using the dirstring_type option. See the new comment in the default
2348 openssl.cnf for more info.
2351 *) Make crypto/rand/md_rand.c more robust:
2352 - Assure unique random numbers after fork().
2353 - Make sure that concurrent threads access the global counter and
2354 md serializably so that we never lose entropy in them
2355 or use exactly the same state in multiple threads.
2356 Access to the large state is not always serializable because
2357 the additional locking could be a performance killer, and
2358 md should be large enough anyway.
2361 *) New file apps/app_rand.c with commonly needed functionality
2362 for handling the random seed file.
2364 Use the random seed file in some applications that previously did not:
2366 dsaparam -genkey (which also ignored its '-rand' option),
2369 x509 (when signing).
2370 Except on systems with /dev/urandom, it is crucial to have a random
2371 seed file at least for key creation, DSA signing, and for DH exchanges;
2372 for RSA signatures we could do without one.
2374 gendh and gendsa (unlike genrsa) used to read only the first byte
2375 of each file listed in the '-rand' option. The function as previously
2376 found in genrsa is now in app_rand.c and is used by all programs
2377 that support '-rand'.
2380 *) In RAND_write_file, use mode 0600 for creating files;
2381 don't just chmod when it may be too late.
2384 *) Report an error from X509_STORE_load_locations
2385 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2388 *) New function ASN1_mbstring_copy() this copies a string in either
2389 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2390 into an ASN1_STRING type. A mask of permissible types is passed
2391 and it chooses the "minimal" type to use or an error if not type
2395 *) Add function equivalents to the various macros in asn1.h. The old
2396 macros are retained with an M_ prefix. Code inside the library can
2397 use the M_ macros. External code (including the openssl utility)
2398 should *NOT* in order to be "shared library friendly".
2401 *) Add various functions that can check a certificate's extensions
2402 to see if it usable for various purposes such as SSL client,
2403 server or S/MIME and CAs of these types. This is currently
2404 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2405 verification. Also added a -purpose flag to x509 utility to
2406 print out all the purposes.
2409 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2413 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2414 for, obtain and decode and extension and obtain its critical flag.
2415 This allows all the necessary extension code to be handled in a
2416 single function call.
2419 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2420 platforms. See crypto/rc4/rc4_enc.c for further details.
2423 *) New -noout option to asn1parse. This causes no output to be produced
2424 its main use is when combined with -strparse and -out to extract data
2425 from a file (which may not be in ASN.1 format).
2428 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2429 when producing the local key id.
2430 [Richard Levitte <levitte@stacken.kth.se>]
2432 *) New option -dhparam in s_server. This allows a DH parameter file to be
2433 stated explicitly. If it is not stated then it tries the first server
2434 certificate file. The previous behaviour hard coded the filename
2438 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2439 a public key to be input or output. For example:
2440 openssl rsa -in key.pem -pubout -out pubkey.pem
2441 Also added necessary DSA public key functions to handle this.
2444 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2445 in the message. This was handled by allowing
2446 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2447 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2449 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2450 to the end of the strings whereas this didn't. This would cause problems
2451 if strings read with d2i_ASN1_bytes() were later modified.
2452 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2454 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2455 data and it contains EOF it will end up returning an error. This is
2456 caused by input 46 bytes long. The cause is due to the way base64
2457 BIOs find the start of base64 encoded data. They do this by trying a
2458 trial decode on each line until they find one that works. When they
2459 do a flag is set and it starts again knowing it can pass all the
2460 data directly through the decoder. Unfortunately it doesn't reset
2461 the context it uses. This means that if EOF is reached an attempt
2462 is made to pass two EOFs through the context and this causes the
2463 resulting error. This can also cause other problems as well. As is
2464 usual with these problems it takes *ages* to find and the fix is
2465 trivial: move one line.
2466 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2468 *) Ugly workaround to get s_client and s_server working under Windows. The
2469 old code wouldn't work because it needed to select() on sockets and the
2470 tty (for keypresses and to see if data could be written). Win32 only
2471 supports select() on sockets so we select() with a 1s timeout on the
2472 sockets and then see if any characters are waiting to be read, if none
2473 are present then we retry, we also assume we can always write data to
2474 the tty. This isn't nice because the code then blocks until we've
2475 received a complete line of data and it is effectively polling the
2476 keyboard at 1s intervals: however it's quite a bit better than not
2477 working at all :-) A dedicated Windows application might handle this
2478 with an event loop for example.
2481 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2482 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2483 will be called when RSA_sign() and RSA_verify() are used. This is useful
2484 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2485 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2486 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2487 This necessitated the support of an extra signature type NID_md5_sha1
2488 for SSL signatures and modifications to the SSL library to use it instead
2489 of calling RSA_public_decrypt() and RSA_private_encrypt().
2492 *) Add new -verify -CAfile and -CApath options to the crl program, these
2493 will lookup a CRL issuers certificate and verify the signature in a
2494 similar way to the verify program. Tidy up the crl program so it
2495 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2496 less strict. It will now permit CRL extensions even if it is not
2497 a V2 CRL: this will allow it to tolerate some broken CRLs.
2500 *) Initialize all non-automatic variables each time one of the openssl
2501 sub-programs is started (this is necessary as they may be started
2502 multiple times from the "OpenSSL>" prompt).
2503 [Lennart Bang, Bodo Moeller]
2505 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2506 removing all other RSA functionality (this is what NO_RSA does). This
2507 is so (for example) those in the US can disable those operations covered
2508 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2512 *) Non-copying interface to BIO pairs.
2513 (still largely untested)
2516 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2517 ASCII string. This was handled independently in various places before.
2520 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2521 UTF8 strings a character at a time.
2524 *) Use client_version from client hello to select the protocol
2525 (s23_srvr.c) and for RSA client key exchange verification
2526 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2529 *) Add various utility functions to handle SPKACs, these were previously
2530 handled by poking round in the structure internals. Added new function
2531 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2532 print, verify and generate SPKACs. Based on an original idea from
2533 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2536 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2539 *) Allow the config file extension section to be overwritten on the
2540 command line. Based on an original idea from Massimiliano Pala
2541 <madwolf@comune.modena.it>. The new option is called -extensions
2542 and can be applied to ca, req and x509. Also -reqexts to override
2543 the request extensions in req and -crlexts to override the crl extensions
2547 *) Add new feature to the SPKAC handling in ca. Now you can include
2548 the same field multiple times by preceding it by "XXXX." for example:
2551 this is the same syntax as used in the req config file.
2554 *) Allow certificate extensions to be added to certificate requests. These
2555 are specified in a 'req_extensions' option of the req section of the
2556 config file. They can be printed out with the -text option to req but
2557 are otherwise ignored at present.
2560 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2561 data read consists of only the final block it would not decrypted because
2562 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2563 A misplaced 'break' also meant the decrypted final block might not be
2564 copied until the next read.
2567 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2568 a few extra parameters to the DH structure: these will be useful if
2569 for example we want the value of 'q' or implement X9.42 DH.
2572 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2573 provides hooks that allow the default DSA functions or functions on a
2574 "per key" basis to be replaced. This allows hardware acceleration and
2575 hardware key storage to be handled without major modification to the
2576 library. Also added low level modexp hooks and CRYPTO_EX structure and
2577 associated functions.
2580 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2581 as "read only": it can't be written to and the buffer it points to will
2582 not be freed. Reading from a read only BIO is much more efficient than
2583 a normal memory BIO. This was added because there are several times when
2584 an area of memory needs to be read from a BIO. The previous method was
2585 to create a memory BIO and write the data to it, this results in two
2586 copies of the data and an O(n^2) reading algorithm. There is a new
2587 function BIO_new_mem_buf() which creates a read only memory BIO from
2588 an area of memory. Also modified the PKCS#7 routines to use read only
2592 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2593 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2594 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2595 but a retry condition occured while trying to read the rest.
2598 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2599 NID_pkcs7_encrypted by default: this was wrong since this should almost
2600 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2601 the encrypted data type: this is a more sensible place to put it and it
2602 allows the PKCS#12 code to be tidied up that duplicated this
2606 *) Changed obj_dat.pl script so it takes its input and output files on
2607 the command line. This should avoid shell escape redirection problems
2611 *) Initial support for certificate extension requests, these are included
2612 in things like Xenroll certificate requests. Included functions to allow
2613 extensions to be obtained and added.
2616 *) -crlf option to s_client and s_server for sending newlines as
2617 CRLF (as required by many protocols).
2620 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2622 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2623 [Ralf S. Engelschall]
2625 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2626 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2628 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2632 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2633 DH parameters/keys (q is lost during that conversion, but the resulting
2634 DH parameters contain its length).
2636 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2637 much faster than DH_generate_parameters (which creates parameters
2638 where p = 2*q + 1), and also the smaller q makes DH computations
2639 much more efficient (160-bit exponentiation instead of 1024-bit
2640 exponentiation); so this provides a convenient way to support DHE
2641 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2642 utter importance to use
2643 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2645 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2646 when such DH parameters are used, because otherwise small subgroup
2647 attacks may become possible!
2650 *) Avoid memory leak in i2d_DHparams.
2653 *) Allow the -k option to be used more than once in the enc program:
2654 this allows the same encrypted message to be read by multiple recipients.
2657 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2658 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2659 it will always use the numerical form of the OID, even if it has a short
2663 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2664 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2665 otherwise bn_mod_exp was called. In the case of hardware keys for example
2666 no private key components need be present and it might store extra data
2667 in the RSA structure, which cannot be accessed from bn_mod_exp.
2668 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2669 private key operations.
2672 *) Added support for SPARC Linux.
2675 *) pem_password_cb function type incompatibly changed from
2676 typedef int pem_password_cb(char *buf, int size, int rwflag);
2678 ....(char *buf, int size, int rwflag, void *userdata);
2679 so that applications can pass data to their callbacks:
2680 The PEM[_ASN1]_{read,write}... functions and macros now take an
2681 additional void * argument, which is just handed through whenever
2682 the password callback is called.
2683 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2685 New function SSL_CTX_set_default_passwd_cb_userdata.
2687 Compatibility note: As many C implementations push function arguments
2688 onto the stack in reverse order, the new library version is likely to
2689 interoperate with programs that have been compiled with the old
2690 pem_password_cb definition (PEM_whatever takes some data that
2691 happens to be on the stack as its last argument, and the callback
2692 just ignores this garbage); but there is no guarantee whatsoever that
2695 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2696 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2697 problems not only on Windows, but also on some Unix platforms.
2698 To avoid problematic command lines, these definitions are now in an
2699 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2700 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2703 *) MIPS III/IV assembler module is reimplemented.
2706 *) More DES library cleanups: remove references to srand/rand and
2707 delete an unused file.
2710 *) Add support for the the free Netwide assembler (NASM) under Win32,
2711 since not many people have MASM (ml) and it can be hard to obtain.
2712 This is currently experimental but it seems to work OK and pass all
2713 the tests. Check out INSTALL.W32 for info.
2716 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2717 without temporary keys kept an extra copy of the server key,
2718 and connections with temporary keys did not free everything in case
2722 *) New function RSA_check_key and new openssl rsa option -check
2723 for verifying the consistency of RSA keys.
2724 [Ulf Moeller, Bodo Moeller]
2726 *) Various changes to make Win32 compile work:
2727 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2728 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2729 comparison" warnings.
2730 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2733 *) Add a debugging option to PKCS#5 v2 key generation function: when
2734 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2735 derived keys are printed to stderr.
2738 *) Copy the flags in ASN1_STRING_dup().
2739 [Roman E. Pavlov <pre@mo.msk.ru>]
2741 *) The x509 application mishandled signing requests containing DSA
2742 keys when the signing key was also DSA and the parameters didn't match.
2744 It was supposed to omit the parameters when they matched the signing key:
2745 the verifying software was then supposed to automatically use the CA's
2746 parameters if they were absent from the end user certificate.
2748 Omitting parameters is no longer recommended. The test was also
2749 the wrong way round! This was probably due to unusual behaviour in
2750 EVP_cmp_parameters() which returns 1 if the parameters match.
2751 This meant that parameters were omitted when they *didn't* match and
2752 the certificate was useless. Certificates signed with 'ca' didn't have
2754 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2756 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2757 The interface is as follows:
2758 Applications can use
2759 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2760 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2761 "off" is now the default.
2762 The library internally uses
2763 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2764 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2765 to disable memory-checking temporarily.
2767 Some inconsistent states that previously were possible (and were
2768 even the default) are now avoided.
2770 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2771 with each memory chunk allocated; this is occasionally more helpful
2772 than just having a counter.
2774 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2776 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2780 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2781 which largely parallels "options", but is for changing API behaviour,
2782 whereas "options" are about protocol behaviour.
2783 Initial "mode" flags are:
2785 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2786 a single record has been written.
2787 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2788 retries use the same buffer location.
2789 (But all of the contents must be
2793 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2796 *) Fix problems with no-hmac etc.
2797 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2799 *) New functions RSA_get_default_method(), RSA_set_method() and
2800 RSA_get_method(). These allows replacement of RSA_METHODs without having
2801 to mess around with the internals of an RSA structure.
2804 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2805 Also really enable memory leak checks in openssl.c and in some
2807 [Chad C. Mulligan, Bodo Moeller]
2809 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2810 up the length of negative integers. This has now been simplified to just
2811 store the length when it is first determined and use it later, rather
2812 than trying to keep track of where data is copied and updating it to
2814 [Steve Henson, reported by Brien Wheeler
2815 <bwheeler@authentica-security.com>]
2817 *) Add a new function PKCS7_signatureVerify. This allows the verification
2818 of a PKCS#7 signature but with the signing certificate passed to the
2819 function itself. This contrasts with PKCS7_dataVerify which assumes the
2820 certificate is present in the PKCS#7 structure. This isn't always the
2821 case: certificates can be omitted from a PKCS#7 structure and be
2822 distributed by "out of band" means (such as a certificate database).
2825 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2826 function prototypes in pem.h, also change util/mkdef.pl to add the
2827 necessary function names.
2830 *) mk1mf.pl (used by Windows builds) did not properly read the
2831 options set by Configure in the top level Makefile, and Configure
2832 was not even able to write more than one option correctly.
2833 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2836 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2837 file to be loaded from a BIO or FILE pointer. The BIO version will
2838 for example allow memory BIOs to contain config info.
2841 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2842 Whoever hopes to achieve shared-library compatibility across versions
2843 must use this, not the compile-time macro.
2844 (Exercise 0.9.4: Which is the minimum library version required by
2846 Note: All this applies only to multi-threaded programs, others don't
2850 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2851 through a BIO pair triggered the default case, i.e.
2852 SSLerr(...,SSL_R_UNKNOWN_STATE).
2855 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2856 can use the SSL library even if none of the specific BIOs is
2860 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2861 for the encoded length.
2862 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2864 *) Add initial documentation of the X509V3 functions.
2867 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2868 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2869 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2870 secure PKCS#8 private key format with a high iteration count.
2873 *) Fix determination of Perl interpreter: A perl or perl5
2874 _directory_ in $PATH was also accepted as the interpreter.
2875 [Ralf S. Engelschall]
2877 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2878 wrong with it but it was very old and did things like calling
2879 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2883 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2884 to use the new extension code.
2887 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2888 with macros. This should make it easier to change their form, add extra
2889 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2893 *) Add to configuration table a new entry that can specify an alternative
2894 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2895 according to Mark Crispin <MRC@Panda.COM>.
2899 *) DES CBC did not update the IV. Weird.
2902 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2903 Changing the behaviour of the former might break existing programs --
2904 where IV updating is needed, des_ncbc_encrypt can be used.
2907 *) When bntest is run from "make test" it drives bc to check its
2908 calculations, as well as internally checking them. If an internal check
2909 fails, it needs to cause bc to give a non-zero result or make test carries
2910 on without noticing the failure. Fixed.
2913 *) DES library cleanups.
2916 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2917 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2918 ciphers. NOTE: although the key derivation function has been verified
2919 against some published test vectors it has not been extensively tested
2920 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2924 *) Instead of "mkdir -p", which is not fully portable, use new
2925 Perl script "util/mkdir-p.pl".
2928 *) Rewrite the way password based encryption (PBE) is handled. It used to
2929 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2930 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2931 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2932 the 'parameter' field of the AlgorithmIdentifier is passed to the
2933 underlying key generation function so it must do its own ASN1 parsing.
2934 This has also changed the EVP_PBE_CipherInit() function which now has a
2935 'parameter' argument instead of literal salt and iteration count values
2936 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2939 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2940 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2941 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2942 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2943 value was just used as a "magic string" and not used directly its
2944 value doesn't matter.
2947 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2951 *) "linux-sparc64" configuration (ultrapenguin).
2952 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2953 "linux-sparc" configuration.
2954 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2956 *) config now generates no-xxx options for missing ciphers.
2959 *) Support the EBCDIC character set (work in progress).
2960 File ebcdic.c not yet included because it has a different license.
2961 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2963 *) Support BS2000/OSD-POSIX.
2964 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2966 *) Make callbacks for key generation use void * instead of char *.
2969 *) Make S/MIME samples compile (not yet tested).
2972 *) Additional typesafe stacks.
2975 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2979 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2981 *) New configuration variant "sco5-gcc".
2983 *) Updated some demos.
2984 [Sean O Riordain, Wade Scholine]
2986 *) Add missing BIO_free at exit of pkcs12 application.
2989 *) Fix memory leak in conf.c.
2992 *) Updates for Win32 to assembler version of MD5.
2995 *) Set #! path to perl in apps/der_chop to where we found it
2996 instead of using a fixed path.
2999 *) SHA library changes for irix64-mips4-cc.
3002 *) Improvements for VMS support.
3006 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3008 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3009 This also avoids the problems with SC4.2 and unpatched SC5.
3010 [Andy Polyakov <appro@fy.chalmers.se>]
3012 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3013 These are required because of the typesafe stack would otherwise break
3014 existing code. If old code used a structure member which used to be STACK
3015 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3016 sk_num or sk_value it would produce an error because the num, data members
3017 are not present in STACK_OF. Now it just produces a warning. sk_set
3018 replaces the old method of assigning a value to sk_value
3019 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3020 that does this will no longer work (and should use sk_set instead) but
3021 this could be regarded as a "questionable" behaviour anyway.
3024 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3025 correctly handle encrypted S/MIME data.
3028 *) Change type of various DES function arguments from des_cblock
3029 (which means, in function argument declarations, pointer to char)
3030 to des_cblock * (meaning pointer to array with 8 char elements),
3031 which allows the compiler to do more typechecking; it was like
3032 that back in SSLeay, but with lots of ugly casts.
3034 Introduce new type const_des_cblock.
3037 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3038 problems: find RecipientInfo structure that matches recipient certificate
3039 and initialise the ASN1 structures properly based on passed cipher.
3042 *) Belatedly make the BN tests actually check the results.
3045 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3046 to and from BNs: it was completely broken. New compilation option
3047 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3048 key elements as negative integers.
3051 *) Reorganize and speed up MD5.
3052 [Andy Polyakov <appro@fy.chalmers.se>]
3055 [Richard Levitte <richard@levitte.org>]
3057 *) New option -out to asn1parse to allow the parsed structure to be
3058 output to a file. This is most useful when combined with the -strparse
3059 option to examine the output of things like OCTET STRINGS.
3062 *) Make SSL library a little more fool-proof by not requiring any longer
3063 that SSL_set_{accept,connect}_state be called before
3064 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3065 in many applications because usually everything *appeared* to work as
3066 intended anyway -- now it really works as intended).
3069 *) Move openssl.cnf out of lib/.
3072 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3073 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3074 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3075 [Ralf S. Engelschall]
3077 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3078 handle PKCS#7 enveloped data properly.
3079 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3081 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3082 copying pointers. The cert_st handling is changed by this in
3083 various ways (and thus what used to be known as ctx->default_cert
3084 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3085 any longer when s->cert does not give us what we need).
3086 ssl_cert_instantiate becomes obsolete by this change.
3087 As soon as we've got the new code right (possibly it already is?),
3088 we have solved a couple of bugs of the earlier code where s->cert
3089 was used as if it could not have been shared with other SSL structures.
3091 Note that using the SSL API in certain dirty ways now will result
3092 in different behaviour than observed with earlier library versions:
3093 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3094 does not influence s as it used to.
3096 In order to clean up things more thoroughly, inside SSL_SESSION
3097 we don't use CERT any longer, but a new structure SESS_CERT
3098 that holds per-session data (if available); currently, this is
3099 the peer's certificate chain and, for clients, the server's certificate
3100 and temporary key. CERT holds only those values that can have
3101 meaningful defaults in an SSL_CTX.
3104 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3105 from the internal representation. Various PKCS#7 fixes: remove some
3106 evil casts and set the enc_dig_alg field properly based on the signing
3110 *) Allow PKCS#12 password to be set from the command line or the
3111 environment. Let 'ca' get its config file name from the environment
3112 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3116 *) Allow certificate policies extension to use an IA5STRING for the
3117 organization field. This is contrary to the PKIX definition but
3118 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3122 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3123 without disallowing inline assembler and the like for non-pedantic builds.
3126 *) Support Borland C++ builder.
3127 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3132 *) SHA-1 cleanups and performance enhancements.
3133 [Andy Polyakov <appro@fy.chalmers.se>]
3135 *) Sparc v8plus assembler for the bignum library.
3136 [Andy Polyakov <appro@fy.chalmers.se>]
3138 *) Accept any -xxx and +xxx compiler options in Configure.
3141 *) Update HPUX configuration.
3144 *) Add missing sk_<type>_unshift() function to safestack.h
3145 [Ralf S. Engelschall]
3147 *) New function SSL_CTX_use_certificate_chain_file that sets the
3148 "extra_cert"s in addition to the certificate. (This makes sense
3149 only for "PEM" format files, as chains as a whole are not
3153 *) Support verify_depth from the SSL API.
3154 x509_vfy.c had what can be considered an off-by-one-error:
3155 Its depth (which was not part of the external interface)
3156 was actually counting the number of certificates in a chain;
3157 now it really counts the depth.
3160 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3161 instead of X509err, which often resulted in confusing error
3162 messages since the error codes are not globally unique
3163 (e.g. an alleged error in ssl3_accept when a certificate
3164 didn't match the private key).
3166 *) New function SSL_CTX_set_session_id_context that allows to set a default
3167 value (so that you don't need SSL_set_session_id_context for each
3168 connection using the SSL_CTX).
3171 *) OAEP decoding bug fix.
3174 *) Support INSTALL_PREFIX for package builders, as proposed by
3178 *) New Configure options "threads" and "no-threads". For systems
3179 where the proper compiler options are known (currently Solaris
3180 and Linux), "threads" is the default.
3183 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3186 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3187 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3188 such as /usr/local/bin.
3191 *) "make linux-shared" to build shared libraries.
3192 [Niels Poppe <niels@netbox.org>]
3194 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3197 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3198 extension adding in x509 utility.
3201 *) Remove NOPROTO sections and error code comments.
3204 *) Partial rewrite of the DEF file generator to now parse the ANSI
3208 *) New Configure options --prefix=DIR and --openssldir=DIR.
3211 *) Complete rewrite of the error code script(s). It is all now handled
3212 by one script at the top level which handles error code gathering,
3213 header rewriting and C source file generation. It should be much better
3214 than the old method: it now uses a modified version of Ulf's parser to
3215 read the ANSI prototypes in all header files (thus the old K&R definitions
3216 aren't needed for error creation any more) and do a better job of
3217 translating function codes into names. The old 'ASN1 error code imbedded
3218 in a comment' is no longer necessary and it doesn't use .err files which
3219 have now been deleted. Also the error code call doesn't have to appear all
3220 on one line (which resulted in some large lines...).
3223 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3226 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3227 0 (which usually indicates a closed connection), but continue reading.
3230 *) Fix some race conditions.
3233 *) Add support for CRL distribution points extension. Add Certificate
3234 Policies and CRL distribution points documentation.
3237 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3240 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3241 8 of keying material. Merlin has also confirmed interop with this fix
3242 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3243 [Merlin Hughes <merlin@baltimore.ie>]
3245 *) Fix lots of warnings.
3246 [Richard Levitte <levitte@stacken.kth.se>]
3248 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3249 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3250 [Richard Levitte <levitte@stacken.kth.se>]
3252 *) Fix problems with sizeof(long) == 8.
3253 [Andy Polyakov <appro@fy.chalmers.se>]
3255 *) Change functions to ANSI C.
3258 *) Fix typos in error codes.
3259 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3261 *) Remove defunct assembler files from Configure.
3264 *) SPARC v8 assembler BIGNUM implementation.
3265 [Andy Polyakov <appro@fy.chalmers.se>]
3267 *) Support for Certificate Policies extension: both print and set.
3268 Various additions to support the r2i method this uses.
3271 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3272 return a const string when you are expecting an allocated buffer.
3275 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3276 types DirectoryString and DisplayText.
3279 *) Add code to allow r2i extensions to access the configuration database,
3280 add an LHASH database driver and add several ctx helper functions.
3283 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3284 fail when they extended the size of a BIGNUM.
3287 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3288 support typesafe stack.
3291 *) Fix typo in SSL_[gs]et_options().
3292 [Nils Frostberg <nils@medcom.se>]
3294 *) Delete various functions and files that belonged to the (now obsolete)
3295 old X509V3 handling code.
3298 *) New Configure option "rsaref".
3301 *) Don't auto-generate pem.h.
3304 *) Introduce type-safe ASN.1 SETs.
3307 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3308 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3310 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3311 that links with OpenSSL (well at least cause lots of warnings), but fear
3312 not: the conversion is trivial, and it eliminates loads of evil casts. A
3313 few STACKed things have been converted already. Feel free to convert more.
3314 In the fullness of time, I'll do away with the STACK type altogether.
3317 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3318 specified in <certfile> by updating the entry in the index.txt file.
3319 This way one no longer has to edit the index.txt file manually for
3320 revoking a certificate. The -revoke option does the gory details now.
3321 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3323 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3324 `-text' option at all and this way the `-noout -text' combination was
3325 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3326 [Ralf S. Engelschall]
3328 *) Make sure a corresponding plain text error message exists for the
3329 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3330 verify callback function determined that a certificate was revoked.
3331 [Ralf S. Engelschall]
3333 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3334 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3335 all available cipers including rc5, which was forgotten until now.
3336 In order to let the testing shell script know which algorithms
3337 are available, a new (up to now undocumented) command
3338 "openssl list-cipher-commands" is used.
3341 *) Bugfix: s_client occasionally would sleep in select() when
3342 it should have checked SSL_pending() first.
3345 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3346 the raw DSA values prior to ASN.1 encoding.
3349 *) Tweaks to Configure
3350 [Niels Poppe <niels@netbox.org>]
3352 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3356 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3359 *) New config option to avoid instructions that are illegal on the 80386.
3360 The default code is faster, but requires at least a 486.
3363 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3364 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3365 same as SSL2_VERSION anyway.
3368 *) New "-showcerts" option for s_client.
3371 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3372 application. Various cleanups and fixes.
3375 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3376 modify error routines to work internally. Add error codes and PBE init
3377 to library startup routines.
3380 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3381 packing functions to asn1 and evp. Changed function names and error
3382 codes along the way.
3385 *) PKCS12 integration: and so it begins... First of several patches to
3386 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3387 objects to objects.h
3390 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3391 and display support for Thawte strong extranet extension.
3394 *) Add LinuxPPC support.
3395 [Jeff Dubrule <igor@pobox.org>]
3397 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3398 bn_div_words in alpha.s.
3399 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3401 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3402 OAEP isn't supported when OpenSSL is built with RSAref.
3403 [Ulf Moeller <ulf@fitug.de>]
3405 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3406 so they no longer are missing under -DNOPROTO.
3407 [Soren S. Jorvang <soren@t.dk>]
3410 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3412 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3413 doesn't work when the session is reused. Coming soon!
3416 *) Fix a security hole, that allows sessions to be reused in the wrong
3417 context thus bypassing client cert protection! All software that uses
3418 client certs and session caches in multiple contexts NEEDS PATCHING to
3419 allow session reuse! A fuller solution is in the works.
3420 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3422 *) Some more source tree cleanups (removed obsolete files
3423 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3424 permission on "config" script to be executable) and a fix for the INSTALL
3426 [Ulf Moeller <ulf@fitug.de>]
3428 *) Remove some legacy and erroneous uses of malloc, free instead of
3430 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3432 *) Make rsa_oaep_test return non-zero on error.
3433 [Ulf Moeller <ulf@fitug.de>]
3435 *) Add support for native Solaris shared libraries. Configure
3436 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3437 if someone would make that last step automatic.
3438 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3440 *) ctx_size was not built with the right compiler during "make links". Fixed.
3443 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3444 except NULL ciphers". This means the default cipher list will no longer
3445 enable NULL ciphers. They need to be specifically enabled e.g. with
3446 the string "DEFAULT:eNULL".
3449 *) Fix to RSA private encryption routines: if p < q then it would
3450 occasionally produce an invalid result. This will only happen with
3451 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3454 *) Be less restrictive and allow also `perl util/perlpath.pl
3455 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3456 because this way one can also use an interpreter named `perl5' (which is
3457 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3458 installed as `perl').
3459 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3461 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3462 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3464 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3465 advapi32.lib to Win32 build and change the pem test comparision
3466 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3467 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3468 and crypto/des/ede_cbcm_enc.c.
3471 *) DES quad checksum was broken on big-endian architectures. Fixed.
3474 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3475 Win32 test batch file so it (might) work again. The Win32 test batch file
3476 is horrible: I feel ill....
3479 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3480 in e_os.h. Audit of header files to check ANSI and non ANSI
3481 sections: 10 functions were absent from non ANSI section and not exported
3482 from Windows DLLs. Fixed up libeay.num for new functions.
3485 *) Make `openssl version' output lines consistent.
3486 [Ralf S. Engelschall]
3488 *) Fix Win32 symbol export lists for BIO functions: Added
3489 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3490 to ms/libeay{16,32}.def.
3491 [Ralf S. Engelschall]
3493 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3494 fine under Unix and passes some trivial tests I've now added. But the
3495 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3496 added to make sure no one expects that this stuff really works in the
3497 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3498 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3500 [Ralf S. Engelschall]
3502 *) Fix the generation of two part addresses in perl.
3503 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3505 *) Add config entry for Linux on MIPS.
3506 [John Tobey <jtobey@channel1.com>]
3508 *) Make links whenever Configure is run, unless we are on Windoze.
3511 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3512 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3516 *) Add a useful kludge to allow package maintainers to specify compiler and
3517 other platforms details on the command line without having to patch the
3518 Configure script everytime: One now can use ``perl Configure
3519 <id>:<details>'', i.e. platform ids are allowed to have details appended
3520 to them (seperated by colons). This is treated as there would be a static
3521 pre-configured entry in Configure's %table under key <id> with value
3522 <details> and ``perl Configure <id>'' is called. So, when you want to
3523 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3524 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3525 now, which overrides the FreeBSD-elf entry on-the-fly.
3526 [Ralf S. Engelschall]
3528 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3531 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3532 on the `perl Configure ...' command line. This way one can compile
3533 OpenSSL libraries with Position Independent Code (PIC) which is needed
3534 for linking it into DSOs.
3535 [Ralf S. Engelschall]
3537 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3541 *) Cleaned up the LICENSE document: The official contact for any license
3542 questions now is the OpenSSL core team under openssl-core@openssl.org.
3543 And add a paragraph about the dual-license situation to make sure people
3544 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3545 to the OpenSSL toolkit.
3546 [Ralf S. Engelschall]
3548 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3549 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3550 Additonally cleaned up the `make links' target: Remove unnecessary
3551 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3552 to speed processing and no longer clutter the display with confusing
3553 stuff. Instead only the actually done links are displayed.
3554 [Ralf S. Engelschall]
3556 *) Permit null encryption ciphersuites, used for authentication only. It used
3557 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3558 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3562 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3563 signed attributes when verifying signatures (this would break them),
3564 the detached data encoding was wrong and public keys obtained using
3565 X509_get_pubkey() weren't freed.
3568 *) Add text documentation for the BUFFER functions. Also added a work around
3569 to a Win95 console bug. This was triggered by the password read stuff: the
3570 last character typed gets carried over to the next fread(). If you were
3571 generating a new cert request using 'req' for example then the last
3572 character of the passphrase would be CR which would then enter the first
3576 *) Added the new `Includes OpenSSL Cryptography Software' button as
3577 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3578 button and can be used by applications based on OpenSSL to show the
3579 relationship to the OpenSSL project.
3580 [Ralf S. Engelschall]
3582 *) Remove confusing variables in function signatures in files
3583 ssl/ssl_lib.c and ssl/ssl.h.
3584 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3586 *) Don't install bss_file.c under PREFIX/include/
3587 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3589 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3590 functions that return function pointers and has support for NT specific
3591 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3592 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3593 unsigned to signed types: this was killing the Win32 compile.
3596 *) Add new certificate file to stack functions,
3597 SSL_add_dir_cert_subjects_to_stack() and
3598 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3599 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3600 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3601 This means that Apache-SSL and similar packages don't have to mess around
3602 to add as many CAs as they want to the preferred list.
3605 *) Experiment with doxygen documentation. Currently only partially applied to
3607 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3608 openssl.doxy as the configuration file.
3611 *) Get rid of remaining C++-style comments which strict C compilers hate.
3612 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3614 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3615 compiled in by default: it has problems with large keys.
3618 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3619 DH private keys and/or callback functions which directly correspond to
3620 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3621 is needed for applications which have to configure certificates on a
3622 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3624 For the RSA certificate situation is makes no difference, but
3625 for the DSA certificate situation this fixes the "no shared cipher"
3626 problem where the OpenSSL cipher selection procedure failed because the
3627 temporary keys were not overtaken from the context and the API provided
3628 no way to reconfigure them.
3629 The new functions now let applications reconfigure the stuff and they
3630 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3631 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3632 non-public-API function ssl_cert_instantiate() is used as a helper
3633 function and also to reduce code redundancy inside ssl_rsa.c.
3634 [Ralf S. Engelschall]
3636 *) Move s_server -dcert and -dkey options out of the undocumented feature
3637 area because they are useful for the DSA situation and should be
3638 recognized by the users.
3639 [Ralf S. Engelschall]
3641 *) Fix the cipher decision scheme for export ciphers: the export bits are
3642 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3643 SSL_EXP_MASK. So, the original variable has to be used instead of the
3644 already masked variable.
3645 [Richard Levitte <levitte@stacken.kth.se>]
3647 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3648 [Richard Levitte <levitte@stacken.kth.se>]
3650 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3651 from `int' to `unsigned int' because it's a length and initialized by
3652 EVP_DigestFinal() which expects an `unsigned int *'.
3653 [Richard Levitte <levitte@stacken.kth.se>]
3655 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3656 script. Instead use the usual Shell->Perl transition trick.
3657 [Ralf S. Engelschall]
3659 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3660 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3661 -noout -modulus' as it's already the case for `openssl rsa -noout
3662 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3663 currently the public key is printed (a decision which was already done by
3664 `openssl dsa -modulus' in the past) which serves a similar purpose.
3665 Additionally the NO_RSA no longer completely removes the whole -modulus
3666 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3668 [Ralf S. Engelschall]
3670 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3671 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3672 [Arne Ansper <arne@ats.cyber.ee>]
3674 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3675 to be added. Now both 'req' and 'ca' can use new objects defined in the
3679 *) Add cool BIO that does syslog (or event log on NT).
3680 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3682 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3683 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3684 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3685 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3688 *) Add preliminary config info for new extension code.
3691 *) Make RSA_NO_PADDING really use no padding.
3692 [Ulf Moeller <ulf@fitug.de>]
3694 *) Generate errors when private/public key check is done.
3697 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3698 for some CRL extensions and new objects added.
3701 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3702 key usage extension and fuller support for authority key id.
3705 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3706 padding method for RSA, which is recommended for new applications in PKCS
3707 #1 v2.0 (RFC 2437, October 1998).
3708 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3709 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3710 against Bleichbacher's attack on RSA.
3711 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3714 *) Updates to the new SSL compression code
3715 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3717 *) Fix so that the version number in the master secret, when passed
3718 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3719 (because the server will not accept higher), that the version number
3720 is 0x03,0x01, not 0x03,0x00
3721 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3723 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3724 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3725 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3728 *) Support for RAW extensions where an arbitrary extension can be
3729 created by including its DER encoding. See apps/openssl.cnf for
3733 *) Make sure latest Perl versions don't interpret some generated C array
3734 code as Perl array code in the crypto/err/err_genc.pl script.
3735 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3737 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3738 not many people have the assembler. Various Win32 compilation fixes and
3739 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3743 *) Modify configure script 'Configure' to automatically create crypto/date.h
3744 file under Win32 and also build pem.h from pem.org. New script
3745 util/mkfiles.pl to create the MINFO file on environments that can't do a
3746 'make files': perl util/mkfiles.pl >MINFO should work.
3749 *) Major rework of DES function declarations, in the pursuit of correctness
3750 and purity. As a result, many evil casts evaporated, and some weirdness,
3751 too. You may find this causes warnings in your code. Zapping your evil
3752 casts will probably fix them. Mostly.
3755 *) Fix for a typo in asn1.h. Bug fix to object creation script
3756 obj_dat.pl. It considered a zero in an object definition to mean
3757 "end of object": none of the objects in objects.h have any zeros
3758 so it wasn't spotted.
3759 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3761 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3762 Masking (CBCM). In the absence of test vectors, the best I have been able
3763 to do is check that the decrypt undoes the encrypt, so far. Send me test
3764 vectors if you have them.
3767 *) Correct calculation of key length for export ciphers (too much space was
3768 allocated for null ciphers). This has not been tested!
3771 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3772 message is now correct (it understands "crypto" and "ssl" on its
3773 command line). There is also now an "update" option. This will update
3774 the util/ssleay.num and util/libeay.num files with any new functions.
3776 perl util/mkdef.pl crypto ssl update
3777 it will update them.
3780 *) Overhauled the Perl interface (perl/*):
3781 - ported BN stuff to OpenSSL's different BN library
3782 - made the perl/ source tree CVS-aware
3783 - renamed the package from SSLeay to OpenSSL (the files still contain
3784 their history because I've copied them in the repository)
3785 - removed obsolete files (the test scripts will be replaced
3786 by better Test::Harness variants in the future)
3787 [Ralf S. Engelschall]
3789 *) First cut for a very conservative source tree cleanup:
3790 1. merge various obsolete readme texts into doc/ssleay.txt
3791 where we collect the old documents and readme texts.
3792 2. remove the first part of files where I'm already sure that we no
3793 longer need them because of three reasons: either they are just temporary
3794 files which were left by Eric or they are preserved original files where
3795 I've verified that the diff is also available in the CVS via "cvs diff
3796 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3797 the crypto/md/ stuff).
3798 [Ralf S. Engelschall]
3800 *) More extension code. Incomplete support for subject and issuer alt
3801 name, issuer and authority key id. Change the i2v function parameters
3802 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3803 what that's for :-) Fix to ASN1 macro which messed up
3804 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3807 *) Preliminary support for ENUMERATED type. This is largely copied from the
3811 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3812 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3814 *) Make sure `make rehash' target really finds the `openssl' program.
3815 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3817 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3818 like to hear about it if this slows down other processors.
3821 *) Add CygWin32 platform information to Configure script.
3822 [Alan Batie <batie@aahz.jf.intel.com>]
3824 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3825 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3827 *) New program nseq to manipulate netscape certificate sequences
3830 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3834 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3835 but the BN code had some problems that would cause failures when
3836 doing certificate verification and some other functions.
3837 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3839 *) Add ASN1 and PEM code to support netscape certificate sequences.
3842 *) Add ASN1 and PEM code to support netscape certificate sequences.
3845 *) Add several PKIX and private extended key usage OIDs.
3848 *) Modify the 'ca' program to handle the new extension code. Modify
3849 openssl.cnf for new extension format, add comments.
3852 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3853 and add a sample to openssl.cnf so req -x509 now adds appropriate
3857 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3858 error code, add initial support to X509_print() and x509 application.
3861 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3862 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3863 stuff is currently isolated and isn't even compiled yet.
3866 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3867 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3868 Removed the versions check from X509 routines when loading extensions:
3869 this allows certain broken certificates that don't set the version
3870 properly to be processed.
3873 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3874 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3875 can still be regenerated with "make depend".
3878 *) Spelling mistake in C version of CAST-128.
3879 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3881 *) Changes to the error generation code. The perl script err-code.pl
3882 now reads in the old error codes and retains the old numbers, only
3883 adding new ones if necessary. It also only changes the .err files if new
3884 codes are added. The makefiles have been modified to only insert errors
3885 when needed (to avoid needlessly modifying header files). This is done
3886 by only inserting errors if the .err file is newer than the auto generated
3887 C file. To rebuild all the error codes from scratch (the old behaviour)
3888 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3889 or delete all the .err files.
3892 *) CAST-128 was incorrectly implemented for short keys. The C version has
3893 been fixed, but is untested. The assembler versions are also fixed, but
3894 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3895 to regenerate it if needed.
3896 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3897 Hagino <itojun@kame.net>]
3899 *) File was opened incorrectly in randfile.c.
3900 [Ulf Möller <ulf@fitug.de>]
3902 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3903 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3904 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3905 al: it's just almost always a UTCTime. Note this patch adds new error
3906 codes so do a "make errors" if there are problems.
3909 *) Correct Linux 1 recognition in config.
3910 [Ulf Möller <ulf@fitug.de>]
3912 *) Remove pointless MD5 hash when using DSA keys in ca.
3913 [Anonymous <nobody@replay.com>]
3915 *) Generate an error if given an empty string as a cert directory. Also
3916 generate an error if handed NULL (previously returned 0 to indicate an
3917 error, but didn't set one).
3918 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3920 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3923 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3924 parameters. This was causing a warning which killed off the Win32 compile.
3927 *) Remove C++ style comments from crypto/bn/bn_local.h.
3928 [Neil Costigan <neil.costigan@celocom.com>]
3930 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3931 based on a text string, looking up short and long names and finally
3932 "dot" format. The "dot" format stuff didn't work. Added new function
3933 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3934 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3935 OID is not part of the table.
3938 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3939 X509_LOOKUP_by_alias().
3942 *) Sort openssl functions by name.
3945 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3946 encryption from sample DSA keys (in case anyone is interested the password
3950 *) Make _all_ *_free functions accept a NULL pointer.
3951 [Frans Heymans <fheymans@isaserver.be>]
3953 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3955 [Anonymous <nobody@replay.com>]
3957 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3958 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3960 *) Don't blow it for numeric -newkey arguments to apps/req.
3961 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3963 *) Temp key "for export" tests were wrong in s3_srvr.c.
3964 [Anonymous <nobody@replay.com>]
3966 *) Add prototype for temp key callback functions
3967 SSL_CTX_set_tmp_{rsa,dh}_callback().
3970 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3971 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3974 *) X509_name_add_entry() freed the wrong thing after an error.
3975 [Arne Ansper <arne@ats.cyber.ee>]
3977 *) rsa_eay.c would attempt to free a NULL context.
3978 [Arne Ansper <arne@ats.cyber.ee>]
3980 *) BIO_s_socket() had a broken should_retry() on Windoze.
3981 [Arne Ansper <arne@ats.cyber.ee>]
3983 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3984 [Arne Ansper <arne@ats.cyber.ee>]
3986 *) Make sure the already existing X509_STORE->depth variable is initialized
3987 in X509_STORE_new(), but document the fact that this variable is still
3988 unused in the certificate verification process.
3989 [Ralf S. Engelschall]
3991 *) Fix the various library and apps files to free up pkeys obtained from
3992 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3995 *) Fix reference counting in X509_PUBKEY_get(). This makes
3996 demos/maurice/example2.c work, amongst others, probably.
3997 [Steve Henson and Ben Laurie]
3999 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4000 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4001 are no longer created. This way we have a single and consistent command
4002 line interface `openssl <command>', similar to `cvs <command>'.
4003 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4005 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4006 BIT STRING wrapper always have zero unused bits.
4009 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4012 *) Make the top-level INSTALL documentation easier to understand.
4015 *) Makefiles updated to exit if an error occurs in a sub-directory
4016 make (including if user presses ^C) [Paul Sutton]
4018 *) Make Montgomery context stuff explicit in RSA data structure.
4021 *) Fix build order of pem and err to allow for generated pem.h.
4024 *) Fix renumbering bug in X509_NAME_delete_entry().
4027 *) Enhanced the err-ins.pl script so it makes the error library number
4028 global and can add a library name. This is needed for external ASN1 and
4029 other error libraries.
4032 *) Fixed sk_insert which never worked properly.
4035 *) Fix ASN1 macros so they can handle indefinite length construted
4036 EXPLICIT tags. Some non standard certificates use these: they can now
4040 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4041 into a single doc/ssleay.txt bundle. This way the information is still
4042 preserved but no longer messes up this directory. Now it's new room for
4043 the new set of documenation files.
4044 [Ralf S. Engelschall]
4046 *) SETs were incorrectly DER encoded. This was a major pain, because they
4047 shared code with SEQUENCEs, which aren't coded the same. This means that
4048 almost everything to do with SETs or SEQUENCEs has either changed name or
4049 number of arguments.
4050 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4052 *) Fix test data to work with the above.
4055 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4056 was already fixed by Eric for 0.9.1 it seems.
4057 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4059 *) Autodetect FreeBSD3.
4062 *) Fix various bugs in Configure. This affects the following platforms:
4066 unixware-2.0-pentium
4070 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4071 before they are needed.
4074 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4078 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4080 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4081 changed SSLeay to OpenSSL in version strings.
4082 [Ralf S. Engelschall]
4084 *) Some fixups to the top-level documents.
4087 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4088 because the symlink to include/ was missing.
4089 [Ralf S. Engelschall]
4091 *) Incorporated the popular no-RSA/DSA-only patches
4092 which allow to compile a RSA-free SSLeay.
4093 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4095 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4096 when "ssleay" is still not found.
4097 [Ralf S. Engelschall]
4099 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4100 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4102 *) Updated the README file.
4103 [Ralf S. Engelschall]
4105 *) Added various .cvsignore files in the CVS repository subdirs
4106 to make a "cvs update" really silent.
4107 [Ralf S. Engelschall]
4109 *) Recompiled the error-definition header files and added
4110 missing symbols to the Win32 linker tables.
4111 [Ralf S. Engelschall]
4113 *) Cleaned up the top-level documents;
4114 o new files: CHANGES and LICENSE
4115 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4116 o merged COPYRIGHT into LICENSE
4117 o removed obsolete TODO file
4118 o renamed MICROSOFT to INSTALL.W32
4119 [Ralf S. Engelschall]
4121 *) Removed dummy files from the 0.9.1b source tree:
4122 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4123 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4124 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4125 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4126 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4127 [Ralf S. Engelschall]
4129 *) Added various platform portability fixes.
4132 *) The Genesis of the OpenSSL rpject:
4133 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4134 Young and Tim J. Hudson created while they were working for C2Net until
4136 [The OpenSSL Project]
4139 Changes between 0.9.0b and 0.9.1b [not released]
4141 *) Updated a few CA certificates under certs/
4144 *) Changed some BIGNUM api stuff.
4147 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4148 DGUX x86, Linux Alpha, etc.
4151 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4152 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4156 *) Add -strparse option to asn1pars program which parses nested
4158 [Dr Stephen Henson <shenson@bigfoot.com>]
4160 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4163 *) DSA fix for "ca" program.
4166 *) Added "-genkey" option to "dsaparam" program.
4169 *) Added RIPE MD160 (rmd160) message digest.
4172 *) Added -a (all) option to "ssleay version" command.
4175 *) Added PLATFORM define which is the id given to Configure.
4178 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4181 *) Extended the ASN.1 parser routines.
4184 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4187 *) Added a BN_CTX to the BN library.
4190 *) Fixed the weak key values in DES library
4193 *) Changed API in EVP library for cipher aliases.
4196 *) Added support for RC2/64bit cipher.
4199 *) Converted the lhash library to the crypto/mem.c functions.
4202 *) Added more recognized ASN.1 object ids.
4205 *) Added more RSA padding checks for SSL/TLS.
4208 *) Added BIO proxy/filter functionality.
4211 *) Added extra_certs to SSL_CTX which can be used
4212 send extra CA certificates to the client in the CA cert chain sending
4213 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4216 *) Now Fortezza is denied in the authentication phase because
4217 this is key exchange mechanism is not supported by SSLeay at all.
4220 *) Additional PKCS1 checks.
4223 *) Support the string "TLSv1" for all TLS v1 ciphers.
4226 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4227 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4230 *) Fixed a few memory leaks.
4233 *) Fixed various code and comment typos.
4236 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4237 bytes sent in the client random.
4238 [Edward Bishop <ebishop@spyglass.com>]