4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) In BN_div() keep a copy of the sign of 'num' before writing the
7 result to 'rm' because if rm==num the value will be overwritten
8 and produce the wrong result if 'num' is negative: this caused
9 problems with BN_mod() and BN_nnmod().
12 *) Function OCSP_request_verify(). This checks the signature on an
13 OCSP request and verifies the signer certificate. The signer
14 certificate is just checked for a generic purpose and OCSP request
18 *) Add OCSP_check_validity() function to check the validity of OCSP
19 responses. OCSP responses are prepared in real time and may only
20 be a few seconds old. Simply checking that the current time lies
21 between thisUpdate and nextUpdate max reject otherwise valid responses
22 caused by either OCSP responder or client clock innacuracy. Instead
23 we allow thisUpdate and nextUpdate to fall within a certain period of
24 the current time. The age of the response can also optionally be
25 checked. Two new options -validity_period and -status_age added to
29 *) If signature or public key algorithm is unrecognized print out its
30 OID rather that just UNKOWN.
33 *) Avoid coredump with unsupported or invalid public keys by checking if
34 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
35 PKCS7_verify() fails with non detached data.
38 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
39 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
40 ID to be generated from the issuer certificate alone which can then be
41 passed to OCSP_id_issuer_cmp().
44 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
45 ASN1 modules to export functions returning ASN1_ITEM pointers
46 instead of the ASN1_ITEM structures themselves. This adds several
47 new macros which allow the underlying ASN1 function/structure to
48 be accessed transparently. As a result code should not use ASN1_ITEM
49 references directly (such as &X509_it) but instead use the relevant
50 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
51 use of the new ASN1 code on platforms where exporting structures
52 is problematical (for example in shared libraries) but exporting
53 functions returning pointers to structures is not.
56 *) Add support for overriding the generation of SSL/TLS session IDs.
57 These callbacks can be registered either in an SSL_CTX or per SSL.
58 The purpose of this is to allow applications to control, if they wish,
59 the arbitrary values chosen for use as session IDs, particularly as it
60 can be useful for session caching in multiple-server environments. A
61 command-line switch for testing this (and any client code that wishes
62 to use such a feature) has been added to "s_server".
63 [Geoff Thorpe, Lutz Jaenicke]
65 *) Modify mkdef.pl to recognise and parse prprocessor conditionals
66 of the form '#if defined(...) || defined(...) || ...' and
67 '#if !defined(...) && !defined(...) && ...'. This also avoids
68 the growing number of special cases it was previously handling.
71 *) Don't use getenv in library functions when run as setuid/setgid.
72 New function OPENSSL_issetugid().
75 *) Make all configuration macros available for application by making
76 sure they are available in opensslconf.h, by giving them names starting
77 with "OPENSSL_" to avoid conflicts with other packages and by making
78 sure e_os2.h will cover all platform-specific cases together with
80 Additionally, it is now possible to define configuration/platform-
81 specific names (called "system identities"). In the C code, these
82 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
83 macro with the name beginning with "OPENSSL_SYS_", which is determined
84 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
88 *) New option -set_serial to 'req' and 'x509' this allows the serial
89 number to use to be specified on the command line. Previously self
90 signed certificates were hard coded with serial number 0 and the
91 CA options of 'x509' had to use a serial number in a file which was
95 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
96 due to incorrect handling of multi-threading:
98 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
100 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
102 3. Count how many times MemCheck_off() has been called so that
103 nested use can be treated correctly. This also avoids
104 inband-signalling in the previous code (which relied on the
105 assumption that thread ID 0 is impossible).
108 *) New options to 'ca' utility to support V2 CRL entry extensions.
109 Currently CRL reason, invalidity date and hold instruction are
110 supported. Add new CRL extensions to V3 code and some new objects.
113 *) Add "-rand" option also to s_client and s_server.
116 *) New function EVP_CIPHER_CTX_set_padding() this is used to
117 disable standard block padding (aka PKCS#5 padding) in the EVP
118 API, which was previously mandatory. This means that the data is
119 not padded in any way and so the total length much be a multiple
120 of the block size, otherwise an error occurs.
123 *) Initial (incomplete) OCSP SSL support.
126 *) Fix CPU detection on Irix 6.x.
127 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
128 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
130 *) New function OCSP_parse_url(). This splits up a URL into its host,
131 port and path components: primarily to parse OCSP URLs. New -url
132 option to ocsp utility.
135 *) New nonce behavior. The return value of OCSP_check_nonce() now
136 reflects the various checks performed. Applications can decide
137 whether to tolerate certain situations such as an absent nonce
138 in a response when one was present in a request: the ocsp application
139 just prints out a warning. New function OCSP_add1_basic_nonce()
140 this is to allow responders to include a nonce in a response even if
141 the request is nonce-less.
144 *) Use the cached encoding of an X509_NAME structure rather than
145 copying it. This is apparently the reason for the libsafe "errors"
146 but the code is actually correct.
149 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
150 skipped when using openssl x509 multiple times on a single input file,
151 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
154 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
155 set string type: to handle setting ASN1_TIME structures. Fix ca
156 utility to correctly initialize revocation date of CRLs.
159 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
160 the clients preferred ciphersuites and rather use its own preferences.
161 Should help to work around M$ SGC (Server Gated Cryptography) bug in
162 Internet Explorer by ensuring unchanged hash method during stepup.
165 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
166 to aes and add a new 'exist' option to print out symbols that don't
170 *) Additional options to ocsp utility to allow flags to be set and
171 additional certificates supplied.
174 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
175 OCSP client a number of certificate to only verify the response
179 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
180 Bleichenbacher's DSA attack.
181 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
182 to be set and top=0 forces the highest bit to be set; top=-1 is new
183 and leaves the highest bit random.
184 [Ulf Moeller, Bodo Moeller]
186 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
187 handle the new API. Currently only ECB, CBC modes supported. Add new
188 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
189 for TLS" draft-ietf-tls-ciphersuite-03.txt.
190 [Ben Laurie, Steve Henson]
192 *) In the NCONF_...-based implementations for CONF_... queries
193 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
194 a temporary CONF structure with the data component set to NULL
195 (which gives segmentation faults in lh_retrieve).
196 Instead, use NULL for the CONF pointer in CONF_get_string and
197 CONF_get_number (which may use environment variables) and directly
198 return NULL from CONF_get_section.
201 *) Fix potential buffer overrun for EBCDIC.
204 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
208 *) Functions for OCSP responders. OCSP_request_onereq_count(),
209 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
210 extract information from a certificate request. OCSP_response_create()
211 creates a response and optionally adds a basic response structure.
212 OCSP_basic_add1_status() adds a complete single response to a basic
213 reponse and returns the OCSP_SINGLERESP structure just added (to allow
214 extensions to be included for example). OCSP_basic_add1_cert() adds a
215 certificate to a basic response and OCSP_basic_sign() signs a basic
216 response with various flags. New helper functions ASN1_TIME_check()
217 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
218 (converts ASN1_TIME to GeneralizedTime).
221 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
222 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
223 structure from a certificate. X509_pubkey_digest() digests tha public_key
224 contents: this is used in various key identifiers.
227 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
228 keyUsage if basicConstraints absent for a CA.
231 *) Make SMIME_write_PKCS7() write mail header values with a format that
232 is more generally accepted (no spaces before the semicolon), since
233 some programs can't parse those values properly otherwise. Also make
234 sure BIO's that break lines after each write do not create invalid
238 *) Make sk_sort() tolerate a NULL argument.
239 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
241 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
242 passed by the function are trusted implicitly. If any of them signed the
243 reponse then it is assumed to be valid and is not verified.
246 *) Zero the premaster secret after deriving the master secret in
250 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
251 to data. This was previously part of the PKCS7 ASN1 code. This
252 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
253 [Steve Henson, reported by Kenneth R. Robinette
254 <support@securenetterm.com>]
256 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
257 routines: without these tracing memory leaks is very painful.
258 Fix leaks in PKCS12 and PKCS7 routines.
261 *) Fix for Irix with NO_ASM.
262 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
264 *) Add some EVP_add_digest_alias registrations (as found in
265 OpenSSL_add_all_digests()), to SSL_library_init()
266 aka OpenSSL_add_ssl_algorithms(). This provides improved
267 compatibility with peers using X.509 certificates
268 with unconventional AlgorithmIdentifier OIDs.
271 *) ./config script fixes.
272 [Ulf Moeller, Richard Levitte]
274 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
275 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
276 effectively meant GeneralizedTime would never be used. Now it
277 is initialised to -1 but X509_time_adj() now has to check the value
278 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
279 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
280 [Steve Henson, reported by Kenneth R. Robinette
281 <support@securenetterm.com>]
283 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
284 result in a zero length in the ASN1_INTEGER structure which was
285 not consistent with the structure when d2i_ASN1_INTEGER() was used
286 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
287 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
288 where it did not print out a minus for negative ASN1_INTEGER.
291 *) Fix 'openssl passwd -1'.
294 *) Add summary printout to ocsp utility. The various functions which
295 convert status values to strings have been renamed to:
296 OCSP_response_status_str(), OCSP_cert_status_str() and
297 OCSP_crl_reason_str() and are no longer static. New options
298 to verify nonce values and to disable verification. OCSP response
299 printout format cleaned up.
302 *) Add additional OCSP certificate checks. These are those specified
303 in RFC2560. This consists of two separate checks: the CA of the
304 certificate being checked must either be the OCSP signer certificate
305 or the issuer of the OCSP signer certificate. In the latter case the
306 OCSP signer certificate must contain the OCSP signing extended key
307 usage. This check is performed by attempting to match the OCSP
308 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
309 in the OCSP_CERTID structures of the response.
312 *) Initial OCSP certificate verification added to OCSP_basic_verify()
313 and related routines. This uses the standard OpenSSL certificate
314 verify routines to perform initial checks (just CA validity) and
315 to obtain the certificate chain. Then additional checks will be
316 performed on the chain. Currently the root CA is checked to see
317 if it is explicitly trusted for OCSP signing. This is used to set
318 a root CA as a global signing root: that is any certificate that
319 chains to that CA is an acceptable OCSP signing certificate.
322 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
323 extensions from a separate configuration file.
324 As when reading extensions from the main configuration file,
325 the '-extensions ...' option may be used for specifying the
327 [Massimiliano Pala <madwolf@comune.modena.it>]
329 *) Change PKCS12_key_gen_asc() so it can cope with non null
330 terminated strings whose length is passed in the passlen
331 parameter, for example from PEM callbacks. This was done
332 by adding an extra length parameter to asc2uni().
333 [Steve Henson, reported by <oddissey@samsung.co.kr>]
335 *) New OCSP utility. Allows OCSP requests to be generated or
336 read. The request can be sent to a responder and the output
337 parsed, outputed or printed in text form. Not complete yet:
338 still needs to check the OCSP response validity.
341 *) New subcommands for 'openssl ca':
342 'openssl ca -status <serial>' prints the status of the cert with
343 the given serial number (according to the index file).
344 'openssl ca -updatedb' updates the expiry status of certificates
346 [Massimiliano Pala <madwolf@comune.modena.it>]
348 *) New '-newreq-nodes' command option to CA.pl. This is like
349 '-newreq', but calls 'openssl req' with the '-nodes' option
350 so that the resulting key is not encrypted.
351 [Damien Miller <djm@mindrot.org>]
353 *) New configuration for the GNU Hurd.
354 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
356 *) Initial code to implement OCSP basic response verify. This
357 is currently incomplete. Currently just finds the signer's
358 certificate and verifies the signature on the response.
361 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
362 value of OPENSSLDIR. This is available via the new '-d' option
363 to 'openssl version', and is also included in 'openssl version -a'.
366 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
367 call failed, free the DSA structure.
370 *) Allowing defining memory allocation callbacks that will be given
371 file name and line number information in additional arguments
372 (a const char* and an int). The basic functionality remains, as
373 well as the original possibility to just replace malloc(),
374 realloc() and free() by functions that do not know about these
375 additional arguments. To register and find out the current
376 settings for extended allocation functions, the following
377 functions are provided:
379 CRYPTO_set_mem_ex_functions
380 CRYPTO_set_locked_mem_ex_functions
381 CRYPTO_get_mem_ex_functions
382 CRYPTO_get_locked_mem_ex_functions
384 These work the same way as CRYPTO_set_mem_functions and friends.
385 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
386 extended allocation function is enabled.
387 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
388 a conventional allocation function is enabled.
389 [Richard Levitte, Bodo Moeller]
391 *) Fix to uni2asc() to cope with zero length Unicode strings.
392 These are present in some PKCS#12 files.
395 *) Finish off removing the remaining LHASH function pointer casts.
396 There should no longer be any prototype-casting required when using
397 the LHASH abstraction, and any casts that remain are "bugs". See
398 the callback types and macros at the head of lhash.h for details
399 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
402 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
403 If an EGD or PRNGD is running and enough entropy is returned, automatic
404 seeding like with /dev/[u]random will be performed.
405 Positions tried are: /etc/entropy, /var/run/egd-pool.
408 *) Change the Unix RAND_poll() variant to be able to poll several
409 random devices, as specified by DEVRANDOM, until a sufficient amount
410 of data has been collected. We spend at most 10 ms on each file
411 (select timeout) and read in non-blocking mode. DEVRANDOM now
412 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
413 (previously it was just the string "/dev/urandom"), so on typical
414 platforms the 10 ms delay will never occur.
415 Also separate out the Unix variant to its own file, rand_unix.c.
416 For VMS, there's a currently-empty rand_vms.c.
419 *) Move OCSP client related routines to ocsp_cl.c. These
420 provide utility functions which an application needing
421 to issue a request to an OCSP responder and analyse the
422 response will typically need: as opposed to those which an
423 OCSP responder itself would need which will be added later.
425 OCSP_request_sign() signs an OCSP request with an API similar
426 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
427 response. OCSP_response_get1_basic() extracts basic response
428 from response. OCSP_resp_find_status(): finds and extracts status
429 information from an OCSP_CERTID structure (which will be created
430 when the request structure is built). These are built from lower
431 level functions which work on OCSP_SINGLERESP structures but
432 wont normally be used unless the application wishes to examine
433 extensions in the OCSP response for example.
435 Replace nonce routines with a pair of functions.
436 OCSP_request_add1_nonce() adds a nonce value and optionally
437 generates a random value. OCSP_check_nonce() checks the
438 validity of the nonce in an OCSP response.
441 *) Change function OCSP_request_add() to OCSP_request_add0_id().
442 This doesn't copy the supplied OCSP_CERTID and avoids the
443 need to free up the newly created id. Change return type
444 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
445 This can then be used to add extensions to the request.
446 Deleted OCSP_request_new(), since most of its functionality
447 is now in OCSP_REQUEST_new() (and the case insensitive name
448 clash) apart from the ability to set the request name which
449 will be added elsewhere.
452 *) Update OCSP API. Remove obsolete extensions argument from
453 various functions. Extensions are now handled using the new
454 OCSP extension code. New simple OCSP HTTP function which
455 can be used to send requests and parse the response.
458 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
459 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
460 uses the special reorder version of SET OF to sort the attributes
461 and reorder them to match the encoded order. This resolves a long
462 standing problem: a verify on a PKCS7 structure just after signing
463 it used to fail because the attribute order did not match the
464 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
465 it uses the received order. This is necessary to tolerate some broken
466 software that does not order SET OF. This is handled by encoding
467 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
468 to produce the required SET OF.
471 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
472 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
473 files to get correct declarations of the ASN.1 item variables.
476 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
477 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
478 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
479 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
480 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
481 ASN1_ITEM and no wrapper functions.
484 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
485 replace the old function pointer based I/O routines. Change most of
486 the *_d2i_bio() and *_d2i_fp() functions to use these.
489 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
490 lines, recognice more "algorithms" that can be deselected, and make
491 it complain about algorithm deselection that isn't recognised.
494 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
495 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
496 to use new functions. Add NO_ASN1_OLD which can be set to remove
497 some old style ASN1 functions: this can be used to determine if old
498 code will still work when these eventually go away.
501 *) New extension functions for OCSP structures, these follow the
502 same conventions as certificates and CRLs.
505 *) New function X509V3_add1_i2d(). This automatically encodes and
506 adds an extension. Its behaviour can be customised with various
507 flags to append, replace or delete. Various wrappers added for
508 certifcates and CRLs.
511 *) Fix to avoid calling the underlying ASN1 print routine when
512 an extension cannot be parsed. Correct a typo in the
513 OCSP_SERVICELOC extension. Tidy up print OCSP format.
516 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
517 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
518 when writing a 32767 byte record.
519 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
521 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
522 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
524 (RSA objects have a reference count access to which is protected
525 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
526 so they are meant to be shared between threads.)
527 [Bodo Moeller, Geoff Thorpe; original patch submitted by
528 "Reddie, Steven" <Steven.Reddie@ca.com>]
530 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
531 entries for variables.
534 *) Fix a deadlock in CRYPTO_mem_leaks().
537 *) Add functionality to apps/openssl.c for detecting locking
538 problems: As the program is single-threaded, all we have
539 to do is register a locking callback using an array for
540 storing which locks are currently held by the program.
543 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
544 SSL_get_ex_data_X509_STORE_idx(), which is used in
545 ssl_verify_cert_chain() and thus can be called at any time
546 during TLS/SSL handshakes so that thread-safety is essential.
547 Unfortunately, the ex_data design is not at all suited
548 for multi-threaded use, so it probably should be abolished.
551 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
552 [Broadcom, tweaked and integrated by Geoff Thorpe]
554 *) Move common extension printing code to new function
555 X509V3_print_extensions(). Reorganise OCSP print routines and
556 implement some needed OCSP ASN1 functions. Add OCSP extensions.
559 *) New function X509_signature_print() to remove duplication in some
563 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
564 set (this was treated exactly the same as SET OF previously). This
565 is used to reorder the STACK representing the structure to match the
566 encoding. This will be used to get round a problem where a PKCS7
567 structure which was signed could not be verified because the STACK
568 order did not reflect the encoded order.
571 *) Reimplement the OCSP ASN1 module using the new code.
574 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
575 for its ASN1 operations. The old style function pointers still exist
576 for now but they will eventually go away.
579 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
580 completely replaces the old ASN1 functionality with a table driven
581 encoder and decoder which interprets an ASN1_ITEM structure describing
582 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
583 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
584 has also been converted to the new form.
587 *) Change BN_mod_exp_recp so that negative moduli are tolerated
588 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
589 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
593 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
594 of not touching the result's sign bit.
597 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
601 *) Changed the LHASH code to use prototypes for callbacks, and created
602 macros to declare and implement thin (optionally static) functions
603 that provide type-safety and avoid function pointer casting for the
604 type-specific callbacks.
607 *) Use better test patterns in bntest.
610 *) Added Kerberos Cipher Suites to be used with TLS, as written in
612 [Veers Staats <staatsvr@asc.hpc.mil>,
613 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
615 *) rand_win.c fix for Borland C.
618 *) BN_rshift bugfix for n == 0.
621 *) Reformat the FAQ so the different questions and answers can be divided
622 in sections depending on the subject.
625 *) Have the zlib compression code load ZLIB.DLL dynamically under
629 *) New function BN_mod_sqrt for computing square roots modulo a prime
630 (using the probabilistic Tonelli-Shanks algorithm unless
631 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
632 be handled deterministically).
633 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
635 *) Store verify_result within SSL_SESSION also for client side to
636 avoid potential security hole. (Re-used sessions on the client side
637 always resulted in verify_result==X509_V_OK, not using the original
638 result of the server certificate verification.)
641 *) Make BN_mod_inverse faster by explicitly handling small quotients
642 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
643 512 bits], about 30% for larger ones [1024 or 2048 bits].)
646 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
647 SSL3_RT_APPLICATION_DATA, return 0.
648 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
652 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
653 releases, have been re-implemented by renaming the previous
654 implementations of ssl2_read and ssl3_read to ssl2_read_internal
655 and ssl3_read_internal, respectively, and adding 'peek' parameters
656 to them. The new ssl[23]_{read,peek} functions are calls to
657 ssl[23]_read_internal with the 'peek' flag set appropriately.
658 A 'peek' parameter has also been added to ssl3_read_bytes, which
659 does the actual work for ssl3_read_internal.
662 *) New function BN_kronecker.
665 *) Fix BN_gcd so that it works on negative inputs; the result is
666 positive unless both parameters are zero.
667 Previously something reasonably close to an infinite loop was
668 possible because numbers could be growing instead of shrinking
669 in the implementation of Euclid's algorithm.
672 *) Fix BN_is_word() and BN_is_one() macros to take into account the
673 sign of the number in question.
675 Fix BN_is_word(a,w) to work correctly for w == 0.
677 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
678 because its test if the absolute value of 'a' equals 'w'.
679 Note that BN_abs_is_word does *not* handle w == 0 reliably;
680 it exists mostly for use in the implementations of BN_is_zero(),
681 BN_is_one(), and BN_is_word().
684 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
685 the method-specific "init()" handler. Also clean up ex_data after
686 calling the method-specific "finish()" handler. Previously, this was
687 happening the other way round.
690 *) New function BN_swap.
693 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
694 the exponentiation functions are more likely to produce reasonable
695 results on negative inputs.
698 *) Change BN_mod_mul so that the result is always non-negative.
699 Previously, it could be negative if one of the factors was negative;
700 I don't think anyone really wanted that behaviour.
703 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
704 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
705 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
706 and add new functions:
719 These functions always generate non-negative results.
721 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
722 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
724 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
725 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
727 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
729 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
730 was actually never needed) and in BN_mul(). The removal in BN_mul()
731 required a small change in bn_mul_part_recursive() and the addition
732 of the functions bn_cmp_part_words(), bn_sub_part_words() and
733 bn_add_part_words(), which do the same thing as bn_cmp_words(),
734 bn_sub_words() and bn_add_words() except they take arrays with
738 *) In 'openssl passwd', verify passwords read from the terminal
739 unless the '-salt' option is used (which usually means that
740 verification would just waste user's time since the resulting
741 hash is going to be compared with some given password hash)
742 or the new '-noverify' option is used.
744 This is an incompatible change, but it does not affect
745 non-interactive use of 'openssl passwd' (passwords on the command
746 line, '-stdin' option, '-in ...' option) and thus should not
750 *) Remove all references to RSAref, since there's no more need for it.
753 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
754 The previous value, 12, was not always sufficient for BN_mod_exp().
757 *) Make DSO load along a path given through an environment variable
758 (SHLIB_PATH) with shl_load().
761 *) Constify the ENGINE code as a result of BIGNUM constification.
762 Also constify the RSA code and most things related to it. In a
763 few places, most notable in the depth of the ASN.1 code, ugly
764 casts back to non-const were required (to be solved at a later
768 *) Make it so the openssl application has all engines loaded by default.
771 *) Constify the BIGNUM routines a little more.
774 *) Make sure that shared libraries get the internal name engine with
775 the full version number and not just 0. This should mark the
776 shared libraries as not backward compatible. Of course, this should
777 be changed again when we can guarantee backward binary compatibility.
780 *) Add the following functions:
786 ENGINE_load_builtin_engines()
788 That way, an application can itself choose if external engines that
789 are built-in in OpenSSL shall ever be used or not. The benefit is
790 that applications won't have to be linked with libdl or other dso
791 libraries unless it's really needed.
793 Changed 'openssl engine' to load all engines on demand.
794 Changed the engine header files to avoid the duplication of some
795 declarations (they differed!).
798 *) 'openssl engine' can now list capabilities.
801 *) Better error reporting in 'openssl engine'.
804 *) Never call load_dh_param(NULL) in s_server.
807 *) Add engine application. It can currently list engines by name and
808 identity, and test if they are actually available.
811 *) Add support for shared libraries under Irix.
812 [Albert Chin-A-Young <china@thewrittenword.com>]
814 *) Improve RPM specification file by forcing symbolic linking and making
815 sure the installed documentation is also owned by root.root.
816 [Damien Miller <djm@mindrot.org>]
818 *) Add configuration option to build on Linux on both big-endian and
820 [Ralf Baechle <ralf@uni-koblenz.de>]
822 *) Give the OpenSSL applications more possibilities to make use of
823 keys (public as well as private) handled by engines.
826 *) Add OCSP code that comes from CertCo.
829 *) Add VMS support for the Rijndael code.
832 *) Added untested support for Nuron crypto accelerator.
835 *) Add support for external cryptographic devices. This code was
836 previously distributed separately as the "engine" branch.
837 [Geoff Thorpe, Richard Levitte]
839 *) Rework the filename-translation in the DSO code. It is now possible to
840 have far greater control over how a "name" is turned into a filename
841 depending on the operating environment and any oddities about the
842 different shared library filenames on each system.
845 *) Support threads on FreeBSD-elf in Configure.
848 *) Add the possibility to create shared libraries on HP-UX
851 *) Fix for SHA1 assembly problem with MASM: it produces
852 warnings about corrupt line number information when assembling
853 with debugging information. This is caused by the overlapping
855 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
858 NCONF_get_number() has no error checking at all. As a replacement,
859 NCONF_get_number_e() is defined (_e for "error checking") and is
860 promoted strongly. The old NCONF_get_number is kept around for
861 binary backward compatibility.
862 Make it possible for methods to load from something other than a BIO,
863 by providing a function pointer that is given a name instead of a BIO.
864 For example, this could be used to load configuration data from an
868 *) Fix typo in get_cert_by_subject() in by_dir.c
869 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
871 *) Rework the system to generate shared libraries:
873 - Make note of the expected extension for the shared libraries and
874 if there is a need for symbolic links from for example libcrypto.so.0
875 to libcrypto.so.0.9.7. There is extended info in Configure for
878 - Make as few rebuilds of the shared libraries as possible.
880 - Still avoid linking the OpenSSL programs with the shared libraries.
882 - When installing, install the shared libraries separately from the
886 *) Fix for non blocking accept BIOs. Added new I/O special reason
887 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
888 with non blocking I/O was not possible because no retry code was
889 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
893 *) Added the beginnings of Rijndael support.
896 *) Fix for bug in DirectoryString mask setting. Add support for
897 X509_NAME_print_ex() in 'req' and X509_print_ex() function
898 to allow certificate printing to more controllable, additional
899 'certopt' option to 'x509' to allow new printing options to be
903 *) Clean old EAY MD5 hack from e_os.h.
906 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
908 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
909 and not in SSL_clear because the latter is also used by the
910 accept/connect functions; previously, the settings made by
911 SSL_set_read_ahead would be lost during the handshake.
912 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
914 *) Correct util/mkdef.pl to be selective about disabled algorithms.
915 Previously, it would create entries for disableed algorithms no
919 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
921 *) In ssl23_get_client_hello, generate an error message when faced
922 with an initial SSL 3.0/TLS record that is too small to contain the
923 first two bytes of the ClientHello message, i.e. client_version.
924 (Note that this is a pathologic case that probably has never happened
925 in real life.) The previous approach was to use the version number
926 from the record header as a substitute; but our protocol choice
927 should not depend on that one because it is not authenticated
928 by the Finished messages.
931 *) More robust randomness gathering functions for Windows.
932 [Jeffrey Altman <jaltman@columbia.edu>]
934 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
935 not set then we don't setup the error code for issuer check errors
936 to avoid possibly overwriting other errors which the callback does
937 handle. If an application does set the flag then we assume it knows
938 what it is doing and can handle the new informational codes
942 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
943 a general "ANY" type, as such it should be able to decode anything
944 including tagged types. However it didn't check the class so it would
945 wrongly interpret tagged types in the same way as their universal
946 counterpart and unknown types were just rejected. Changed so that the
947 tagged and unknown types are handled in the same way as a SEQUENCE:
948 that is the encoding is stored intact. There is also a new type
949 "V_ASN1_OTHER" which is used when the class is not universal, in this
950 case we have no idea what the actual type is so we just lump them all
954 *) On VMS, stdout may very well lead to a file that is written to
955 in a record-oriented fashion. That means that every write() will
956 write a separate record, which will be read separately by the
957 programs trying to read from it. This can be very confusing.
959 The solution is to put a BIO filter in the way that will buffer
960 text until a linefeed is reached, and then write everything a
961 line at a time, so every record written will be an actual line,
962 not chunks of lines and not (usually doesn't happen, but I've
963 seen it once) several lines in one record. BIO_f_linebuffer() is
966 Currently, it's a VMS-only method, because that's where it has
967 been tested well enough.
970 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
971 it can return incorrect results.
972 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
973 but it was in 0.9.6-beta[12].)
976 *) Disable the check for content being present when verifying detached
977 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
978 include zero length content when signing messages.
981 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
982 BIO_ctrl (for BIO pairs).
985 *) Add DSO method for VMS.
988 *) Bug fix: Montgomery multiplication could produce results with the
992 *) Add RPM specification openssl.spec and modify it to build three
993 packages. The default package contains applications, application
994 documentation and run-time libraries. The devel package contains
995 include files, static libraries and function documentation. The
996 doc package contains the contents of the doc directory. The original
997 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
1000 *) Add a large number of documentation files for many SSL routines.
1001 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1003 *) Add a configuration entry for Sony News 4.
1004 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1006 *) Don't set the two most significant bits to one when generating a
1007 random number < q in the DSA library.
1010 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1011 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1012 the underlying transport is blocking) if a handshake took place.
1013 (The default behaviour is needed by applications such as s_client
1014 and s_server that use select() to determine when to use SSL_read;
1015 but for applications that know in advance when to expect data, it
1016 just makes things more complicated.)
1019 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1023 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1024 work better on such systems.
1025 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1027 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1028 Update PKCS12_parse() so it copies the friendlyName and the
1029 keyid to the certificates aux info.
1032 *) Fix bug in PKCS7_verify() which caused an infinite loop
1033 if there was more than one signature.
1034 [Sven Uszpelkat <su@celocom.de>]
1036 *) Major change in util/mkdef.pl to include extra information
1037 about each symbol, as well as presentig variables as well
1038 as functions. This change means that there's n more need
1039 to rebuild the .num files when some algorithms are excluded.
1042 *) Allow the verify time to be set by an application,
1043 rather than always using the current time.
1046 *) Phase 2 verify code reorganisation. The certificate
1047 verify code now looks up an issuer certificate by a
1048 number of criteria: subject name, authority key id
1049 and key usage. It also verifies self signed certificates
1050 by the same criteria. The main comparison function is
1051 X509_check_issued() which performs these checks.
1053 Lot of changes were necessary in order to support this
1054 without completely rewriting the lookup code.
1056 Authority and subject key identifier are now cached.
1058 The LHASH 'certs' is X509_STORE has now been replaced
1059 by a STACK_OF(X509_OBJECT). This is mainly because an
1060 LHASH can't store or retrieve multiple objects with
1061 the same hash value.
1063 As a result various functions (which were all internal
1064 use only) have changed to handle the new X509_STORE
1065 structure. This will break anything that messed round
1066 with X509_STORE internally.
1068 The functions X509_STORE_add_cert() now checks for an
1069 exact match, rather than just subject name.
1071 The X509_STORE API doesn't directly support the retrieval
1072 of multiple certificates matching a given criteria, however
1073 this can be worked round by performing a lookup first
1074 (which will fill the cache with candidate certificates)
1075 and then examining the cache for matches. This is probably
1076 the best we can do without throwing out X509_LOOKUP
1077 entirely (maybe later...).
1079 The X509_VERIFY_CTX structure has been enhanced considerably.
1081 All certificate lookup operations now go via a get_issuer()
1082 callback. Although this currently uses an X509_STORE it
1083 can be replaced by custom lookups. This is a simple way
1084 to bypass the X509_STORE hackery necessary to make this
1085 work and makes it possible to use more efficient techniques
1086 in future. A very simple version which uses a simple
1087 STACK for its trusted certificate store is also provided
1088 using X509_STORE_CTX_trusted_stack().
1090 The verify_cb() and verify() callbacks now have equivalents
1091 in the X509_STORE_CTX structure.
1093 X509_STORE_CTX also has a 'flags' field which can be used
1094 to customise the verify behaviour.
1097 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1098 excludes S/MIME capabilities.
1101 *) When a certificate request is read in keep a copy of the
1102 original encoding of the signed data and use it when outputing
1103 again. Signatures then use the original encoding rather than
1104 a decoded, encoded version which may cause problems if the
1105 request is improperly encoded.
1108 *) For consistency with other BIO_puts implementations, call
1109 buffer_write(b, ...) directly in buffer_puts instead of calling
1112 In BIO_puts, increment b->num_write as in BIO_write.
1113 [Peter.Sylvester@EdelWeb.fr]
1115 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1116 BN_zero, we may not return a BIGNUM with an array consisting of
1120 *) Avoid calling abort() from within the library when problems are
1121 detected, except if preprocessor symbols have been defined
1122 (such as REF_CHECK, BN_DEBUG etc.).
1125 *) New openssl application 'rsautl'. This utility can be
1126 used for low level RSA operations. DER public key
1127 BIO/fp routines also added.
1130 *) New Configure entry and patches for compiling on QNX 4.
1131 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1133 *) A demo state-machine implementation was sponsored by
1134 Nuron (http://www.nuron.com/) and is now available in
1135 demos/state_machine.
1138 *) New options added to the 'dgst' utility for signature
1139 generation and verification.
1142 *) Unrecognized PKCS#7 content types are now handled via a
1143 catch all ASN1_TYPE structure. This allows unsupported
1144 types to be stored as a "blob" and an application can
1145 encode and decode it manually.
1148 *) Fix various signed/unsigned issues to make a_strex.c
1150 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1152 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1153 length if passed a buffer. ASN1_INTEGER_to_BN failed
1154 if passed a NULL BN and its argument was negative.
1155 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1157 *) Modification to PKCS#7 encoding routines to output definite
1158 length encoding. Since currently the whole structures are in
1159 memory there's not real point in using indefinite length
1160 constructed encoding. However if OpenSSL is compiled with
1161 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1164 *) Added BIO_vprintf() and BIO_vsnprintf().
1167 *) Added more prefixes to parse for in the the strings written
1168 through a logging bio, to cover all the levels that are available
1169 through syslog. The prefixes are now:
1171 PANIC, EMERG, EMR => LOG_EMERG
1172 ALERT, ALR => LOG_ALERT
1173 CRIT, CRI => LOG_CRIT
1174 ERROR, ERR => LOG_ERR
1175 WARNING, WARN, WAR => LOG_WARNING
1176 NOTICE, NOTE, NOT => LOG_NOTICE
1177 INFO, INF => LOG_INFO
1178 DEBUG, DBG => LOG_DEBUG
1180 and as before, if none of those prefixes are present at the
1181 beginning of the string, LOG_ERR is chosen.
1183 On Win32, the LOG_* levels are mapped according to this:
1185 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1186 LOG_WARNING => EVENTLOG_WARNING_TYPE
1187 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1191 *) Made it possible to reconfigure with just the configuration
1192 argument "reconf" or "reconfigure". The command line arguments
1193 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1194 and are retrieved from there when reconfiguring.
1198 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1200 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1203 *) The obj_dat.pl script was messing up the sorting of object
1204 names. The reason was that it compared the quoted version
1205 of strings as a result "OCSP" > "OCSP Signing" because
1206 " > SPACE. Changed script to store unquoted versions of
1207 names and add quotes on output. It was also omitting some
1208 names from the lookup table if they were given a default
1209 value (that is if SN is missing it is given the same
1210 value as LN and vice versa), these are now added on the
1211 grounds that if an object has a name we should be able to
1212 look it up. Finally added warning output when duplicate
1213 short or long names are found.
1216 *) Changes needed for Tandem NSK.
1217 [Scott Uroff <scott@xypro.com>]
1219 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1220 RSA_padding_check_SSLv23(), special padding was never detected
1221 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1222 version rollback attacks was not effective.
1224 In s23_clnt.c, don't use special rollback-attack detection padding
1225 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1226 client; similarly, in s23_srvr.c, don't do the rollback check if
1227 SSL 2.0 is the only protocol enabled in the server.
1230 *) Make it possible to get hexdumps of unprintable data with 'openssl
1231 asn1parse'. By implication, the functions ASN1_parse_dump() and
1232 BIO_dump_indent() are added.
1235 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1236 these print out strings and name structures based on various
1237 flags including RFC2253 support and proper handling of
1238 multibyte characters. Added options to the 'x509' utility
1239 to allow the various flags to be set.
1242 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1243 Also change the functions X509_cmp_current_time() and
1244 X509_gmtime_adj() work with an ASN1_TIME structure,
1245 this will enable certificates using GeneralizedTime in validity
1246 dates to be checked.
1249 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1250 negative public key encodings) on by default,
1251 NO_NEG_PUBKEY_BUG can be set to disable it.
1254 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1255 content octets. An i2c_ASN1_OBJECT is unnecessary because
1256 the encoding can be trivially obtained from the structure.
1259 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1260 not read locks (CRYPTO_r_[un]lock).
1263 *) A first attempt at creating official support for shared
1264 libraries through configuration. I've kept it so the
1265 default is static libraries only, and the OpenSSL programs
1266 are always statically linked for now, but there are
1267 preparations for dynamic linking in place.
1268 This has been tested on Linux and True64.
1271 *) Randomness polling function for Win9x, as described in:
1272 Peter Gutmann, Software Generation of Practically Strong
1276 *) Fix so PRNG is seeded in req if using an already existing
1280 *) New options to smime application. -inform and -outform
1281 allow alternative formats for the S/MIME message including
1282 PEM and DER. The -content option allows the content to be
1283 specified separately. This should allow things like Netscape
1284 form signing output easier to verify.
1287 *) Fix the ASN1 encoding of tags using the 'long form'.
1290 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1291 STRING types. These convert content octets to and from the
1292 underlying type. The actual tag and length octets are
1293 already assumed to have been read in and checked. These
1294 are needed because all other string types have virtually
1295 identical handling apart from the tag. By having versions
1296 of the ASN1 functions that just operate on content octets
1297 IMPLICIT tagging can be handled properly. It also allows
1298 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1299 and ASN1_INTEGER are identical apart from the tag.
1302 *) Change the handling of OID objects as follows:
1304 - New object identifiers are inserted in objects.txt, following
1305 the syntax given in objects.README.
1306 - objects.pl is used to process obj_mac.num and create a new
1308 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1311 This is currently kind of a hack, and the perl code in objects.pl
1312 isn't very elegant, but it works as I intended. The simplest way
1313 to check that it worked correctly is to look in obj_dat.h and
1314 check the array nid_objs and make sure the objects haven't moved
1315 around (this is important!). Additions are OK, as well as
1316 consistent name changes.
1319 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1322 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1323 The given file adds to whatever has already been seeded into the
1324 random pool through the RANDFILE configuration file option or
1325 environment variable, or the default random state file.
1328 *) mkstack.pl now sorts each macro group into lexical order.
1329 Previously the output order depended on the order the files
1330 appeared in the directory, resulting in needless rewriting
1334 *) Patches to make OpenSSL compile under Win32 again. Mostly
1335 work arounds for the VC++ problem that it treats func() as
1336 func(void). Also stripped out the parts of mkdef.pl that
1337 added extra typesafe functions: these no longer exist.
1340 *) Reorganisation of the stack code. The macros are now all
1341 collected in safestack.h . Each macro is defined in terms of
1342 a "stack macro" of the form SKM_<name>(type, a, b). The
1343 DEBUG_SAFESTACK is now handled in terms of function casts,
1344 this has the advantage of retaining type safety without the
1345 use of additional functions. If DEBUG_SAFESTACK is not defined
1346 then the non typesafe macros are used instead. Also modified the
1347 mkstack.pl script to handle the new form. Needs testing to see
1348 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1349 the default if no major problems. Similar behaviour for ASN1_SET_OF
1350 and PKCS12_STACK_OF.
1353 *) When some versions of IIS use the 'NET' form of private key the
1354 key derivation algorithm is different. Normally MD5(password) is
1355 used as a 128 bit RC4 key. In the modified case
1356 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1357 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1358 as the old Netscape_RSA functions except they have an additional
1359 'sgckey' parameter which uses the modified algorithm. Also added
1360 an -sgckey command line option to the rsa utility. Thanks to
1361 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1362 algorithm to openssl-dev.
1365 *) The evp_local.h macros were using 'c.##kname' which resulted in
1366 invalid expansion on some systems (SCO 5.0.5 for example).
1367 Corrected to 'c.kname'.
1368 [Phillip Porch <root@theporch.com>]
1370 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1371 a STACK of email addresses from a certificate or request, these look
1372 in the subject name and the subject alternative name extensions and
1373 omit any duplicate addresses.
1376 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1377 This makes DSA verification about 2 % faster.
1380 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1381 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1382 plus overhead for 1024 bit moduli).
1383 This makes exponentiations about 0.5 % faster for 1024 bit
1384 exponents (as measured by "openssl speed rsa2048").
1387 *) Rename memory handling macros to avoid conflicts with other
1389 Malloc => OPENSSL_malloc
1390 Malloc_locked => OPENSSL_malloc_locked
1391 Realloc => OPENSSL_realloc
1392 Free => OPENSSL_free
1395 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1396 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1399 *) CygWin32 support.
1400 [John Jarvie <jjarvie@newsguy.com>]
1402 *) The type-safe stack code has been rejigged. It is now only compiled
1403 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1404 by default all type-specific stack functions are "#define"d back to
1405 standard stack functions. This results in more streamlined output
1406 but retains the type-safety checking possibilities of the original
1410 *) The STACK code has been cleaned up, and certain type declarations
1411 that didn't make a lot of sense have been brought in line. This has
1412 also involved a cleanup of sorts in safestack.h to more correctly
1413 map type-safe stack functions onto their plain stack counterparts.
1414 This work has also resulted in a variety of "const"ifications of
1415 lots of the code, especially "_cmp" operations which should normally
1416 be prototyped with "const" parameters anyway.
1419 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1420 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1421 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1422 where all of 'md' is used each time the PRNG is used, but 'state'
1423 is used only indexed by a cyclic counter. As entropy may not be
1424 well distributed from the beginning, 'md' is important as a
1425 chaining variable. However, the output function chains only half
1426 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1427 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1428 in all of 'state' being rewritten, with the new values depending
1429 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1432 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1433 the handshake is continued after ssl_verify_cert_chain();
1434 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1435 can lead to 'unexplainable' connection aborts later.
1436 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1438 *) Major EVP API cipher revision.
1439 Add hooks for extra EVP features. This allows various cipher
1440 parameters to be set in the EVP interface. Support added for variable
1441 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1442 setting of RC2 and RC5 parameters.
1444 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1447 Remove lots of duplicated code from the EVP library. For example *every*
1448 cipher init() function handles the 'iv' in the same way according to the
1449 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1450 for CFB and OFB modes they zero ctx->num.
1452 New functionality allows removal of S/MIME code RC2 hack.
1454 Most of the routines have the same form and so can be declared in terms
1457 By shifting this to the top level EVP_CipherInit() it can be removed from
1458 all individual ciphers. If the cipher wants to handle IVs or keys
1459 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1462 Change lots of functions like EVP_EncryptUpdate() to now return a
1463 value: although software versions of the algorithms cannot fail
1464 any installed hardware versions can.
1467 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1468 this option is set, tolerate broken clients that send the negotiated
1469 protocol version number instead of the requested protocol version
1473 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1474 i.e. non-zero for export ciphersuites, zero otherwise.
1475 Previous versions had this flag inverted, inconsistent with
1476 rsa_tmp_cb (..._TMP_RSA_CB).
1477 [Bodo Moeller; problem reported by Amit Chopra]
1479 *) Add missing DSA library text string. Work around for some IIS
1480 key files with invalid SEQUENCE encoding.
1483 *) Add a document (doc/standards.txt) that list all kinds of standards
1484 and so on that are implemented in OpenSSL.
1487 *) Enhance c_rehash script. Old version would mishandle certificates
1488 with the same subject name hash and wouldn't handle CRLs at all.
1489 Added -fingerprint option to crl utility, to support new c_rehash
1493 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1496 *) Fix for SSL server purpose checking. Server checking was
1497 rejecting certificates which had extended key usage present
1498 but no ssl client purpose.
1499 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1501 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1502 is a little unclear about how a blank password is handled.
1503 Since the password in encoded as a BMPString with terminating
1504 double NULL a zero length password would end up as just the
1505 double NULL. However no password at all is different and is
1506 handled differently in the PKCS#12 key generation code. NS
1507 treats a blank password as zero length. MSIE treats it as no
1508 password on export: but it will try both on import. We now do
1509 the same: PKCS12_parse() tries zero length and no password if
1510 the password is set to "" or NULL (NULL is now a valid password:
1511 it wasn't before) as does the pkcs12 application.
1514 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1515 perror when PEM_read_bio_X509_REQ fails, the error message must
1516 be obtained from the error queue.
1519 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1520 it in ERR_remove_state if appropriate, and change ERR_get_state
1521 accordingly to avoid race conditions (this is necessary because
1522 thread_hash is no longer constant once set).
1525 *) Bugfix for linux-elf makefile.one.
1528 *) RSA_get_default_method() will now cause a default
1529 RSA_METHOD to be chosen if one doesn't exist already.
1530 Previously this was only set during a call to RSA_new()
1531 or RSA_new_method(NULL) meaning it was possible for
1532 RSA_get_default_method() to return NULL.
1535 *) Added native name translation to the existing DSO code
1536 that will convert (if the flag to do so is set) filenames
1537 that are sufficiently small and have no path information
1538 into a canonical native form. Eg. "blah" converted to
1539 "libblah.so" or "blah.dll" etc.
1542 *) New function ERR_error_string_n(e, buf, len) which is like
1543 ERR_error_string(e, buf), but writes at most 'len' bytes
1544 including the 0 terminator. For ERR_error_string_n, 'buf'
1546 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1548 *) CONF library reworked to become more general. A new CONF
1549 configuration file reader "class" is implemented as well as a
1550 new functions (NCONF_*, for "New CONF") to handle it. The now
1551 old CONF_* functions are still there, but are reimplemented to
1552 work in terms of the new functions. Also, a set of functions
1553 to handle the internal storage of the configuration data is
1554 provided to make it easier to write new configuration file
1555 reader "classes" (I can definitely see something reading a
1556 configuration file in XML format, for example), called _CONF_*,
1557 or "the configuration storage API"...
1559 The new configuration file reading functions are:
1561 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1562 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1564 NCONF_default, NCONF_WIN32
1566 NCONF_dump_fp, NCONF_dump_bio
1568 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1569 NCONF_new creates a new CONF object. This works in the same way
1570 as other interfaces in OpenSSL, like the BIO interface.
1571 NCONF_dump_* dump the internal storage of the configuration file,
1572 which is useful for debugging. All other functions take the same
1573 arguments as the old CONF_* functions wth the exception of the
1574 first that must be a `CONF *' instead of a `LHASH *'.
1576 To make it easer to use the new classes with the old CONF_* functions,
1577 the function CONF_set_default_method is provided.
1580 *) Add '-tls1' option to 'openssl ciphers', which was already
1581 mentioned in the documentation but had not been implemented.
1582 (This option is not yet really useful because even the additional
1583 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1586 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1587 OpenSSL-based applications) load shared libraries and bind to
1588 them in a portable way.
1589 [Geoff Thorpe, with contributions from Richard Levitte]
1591 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1593 *) Make sure _lrotl and _lrotr are only used with MSVC.
1595 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1596 (the default implementation of RAND_status).
1598 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1599 to '-clrext' (= clear extensions), as intended and documented.
1600 [Bodo Moeller; inconsistency pointed out by Michael Attili
1603 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1604 was larger than the MD block size.
1605 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1607 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1608 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1609 using the passed key: if the passed key was a private key the result
1610 of X509_print(), for example, would be to print out all the private key
1614 *) des_quad_cksum() byte order bug fix.
1615 [Ulf Möller, using the problem description in krb4-0.9.7, where
1616 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1618 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1620 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1622 *) For easily testing in shell scripts whether some command
1623 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1624 returns with exit code 0 iff no command of the given name is available.
1625 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1626 the output goes to stdout and nothing is printed to stderr.
1627 Additional arguments are always ignored.
1629 Since for each cipher there is a command of the same name,
1630 the 'no-cipher' compilation switches can be tested this way.
1632 ('openssl no-XXX' is not able to detect pseudo-commands such
1633 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1636 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1639 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1640 is set; it will be thrown away anyway because each handshake creates
1642 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1643 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1644 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1645 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1648 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1649 'Q' and 'R' lose their special meanings (quit/renegotiate).
1650 This is part of what -quiet does; unlike -quiet, -ign_eof
1651 does not suppress any output.
1654 *) Add compatibility options to the purpose and trust code. The
1655 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1656 accepts a certificate or CA, this was the previous behaviour,
1657 with all the associated security issues.
1659 X509_TRUST_COMPAT is the old trust behaviour: only and
1660 automatically trust self signed roots in certificate store. A
1661 new trust setting X509_TRUST_DEFAULT is used to specify that
1662 a purpose has no associated trust setting and it should instead
1663 use the value in the default purpose.
1666 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1667 and fix a memory leak.
1670 *) In util/mkerr.pl (which implements 'make errors'), preserve
1671 reason strings from the previous version of the .c file, as
1672 the default to have only downcase letters (and digits) in
1673 automatically generated reasons codes is not always appropriate.
1676 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1677 using strerror. Previously, ERR_reason_error_string() returned
1678 library names as reason strings for SYSerr; but SYSerr is a special
1679 case where small numbers are errno values, not library numbers.
1682 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1683 converts DSA parameters into DH parameters. (When creating parameters,
1684 DSA_generate_parameters is used.)
1687 *) Include 'length' (recommended exponent length) in C code generated
1688 by 'openssl dhparam -C'.
1691 *) The second argument to set_label in perlasm was already being used
1692 so couldn't be used as a "file scope" flag. Moved to third argument
1696 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1697 instead of RAND_bytes for encryption IVs and salts.
1700 *) Include RAND_status() into RAND_METHOD instead of implementing
1701 it only for md_rand.c Otherwise replacing the PRNG by calling
1702 RAND_set_rand_method would be impossible.
1705 *) Don't let DSA_generate_key() enter an infinite loop if the random
1706 number generation fails.
1709 *) New 'rand' application for creating pseudo-random output.
1712 *) Added configuration support for Linux/IA64
1713 [Rolf Haberrecker <rolf@suse.de>]
1715 *) Assembler module support for Mingw32.
1718 *) Shared library support for HPUX (in shlib/).
1719 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1721 *) Shared library support for Solaris gcc.
1722 [Lutz Behnke <behnke@trustcenter.de>]
1724 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1726 *) PKCS7_encrypt() was adding text MIME headers twice because they
1727 were added manually and by SMIME_crlf_copy().
1730 *) In bntest.c don't call BN_rand with zero bits argument.
1731 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1733 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1734 case was implemented. This caused BN_div_recp() to fail occasionally.
1737 *) Add an optional second argument to the set_label() in the perl
1738 assembly language builder. If this argument exists and is set
1739 to 1 it signals that the assembler should use a symbol whose
1740 scope is the entire file, not just the current function. This
1741 is needed with MASM which uses the format label:: for this scope.
1742 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1744 *) Change the ASN1 types so they are typedefs by default. Before
1745 almost all types were #define'd to ASN1_STRING which was causing
1746 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1750 *) Change names of new functions to the new get1/get0 naming
1751 convention: After 'get1', the caller owns a reference count
1752 and has to call ..._free; 'get0' returns a pointer to some
1753 data structure without incrementing reference counters.
1754 (Some of the existing 'get' functions increment a reference
1755 counter, some don't.)
1756 Similarly, 'set1' and 'add1' functions increase reference
1757 counters or duplicate objects.
1760 *) Allow for the possibility of temp RSA key generation failure:
1761 the code used to assume it always worked and crashed on failure.
1764 *) Fix potential buffer overrun problem in BIO_printf().
1765 [Ulf Möller, using public domain code by Patrick Powell; problem
1766 pointed out by David Sacerdote <das33@cornell.edu>]
1768 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1769 RAND_egd() and RAND_status(). In the command line application,
1770 the EGD socket can be specified like a seed file using RANDFILE
1774 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1775 Some CAs (e.g. Verisign) distribute certificates in this form.
1778 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1779 list to exclude them. This means that no special compilation option
1780 is needed to use anonymous DH: it just needs to be included in the
1784 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1785 EVP_MD_type. The old functionality is available in a new macro called
1786 EVP_MD_md(). Change code that uses it and update docs.
1789 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1790 where the 'void *' argument is replaced by a function pointer argument.
1791 Previously 'void *' was abused to point to functions, which works on
1792 many platforms, but is not correct. As these functions are usually
1793 called by macros defined in OpenSSL header files, most source code
1794 should work without changes.
1797 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1798 sections with information on -D... compiler switches used for
1799 compiling the library so that applications can see them. To enable
1800 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1801 must be defined. E.g.,
1802 #define OPENSSL_ALGORITHM_DEFINES
1803 #include <openssl/opensslconf.h>
1804 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1805 [Richard Levitte, Ulf and Bodo Möller]
1807 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1811 *) Change the 'other' type in certificate aux info to a STACK_OF
1812 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1813 the required ASN1 format: arbitrary types determined by an OID.
1816 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1817 argument to 'req'. This is not because the function is newer or
1818 better than others it just uses the work 'NEW' in the certificate
1819 request header lines. Some software needs this.
1822 *) Reorganise password command line arguments: now passwords can be
1823 obtained from various sources. Delete the PEM_cb function and make
1824 it the default behaviour: i.e. if the callback is NULL and the
1825 usrdata argument is not NULL interpret it as a null terminated pass
1826 phrase. If usrdata and the callback are NULL then the pass phrase
1827 is prompted for as usual.
1830 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1831 the support is automatically enabled. The resulting binaries will
1832 autodetect the card and use it if present.
1833 [Ben Laurie and Compaq Inc.]
1835 *) Work around for Netscape hang bug. This sends certificate request
1836 and server done in one record. Since this is perfectly legal in the
1837 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1838 the bugs/SSLv3 entry for more info.
1841 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1844 *) Add -rand argument to smime and pkcs12 applications and read/write
1848 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1851 *) Add command line password options to the remaining applications.
1854 *) Bug fix for BN_div_recp() for numerators with an even number of
1858 *) More tests in bntest.c, and changed test_bn output.
1861 *) ./config recognizes MacOS X now.
1864 *) Bug fix for BN_div() when the first words of num and divsor are
1865 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1868 *) Add support for various broken PKCS#8 formats, and command line
1869 options to produce them.
1872 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1873 get temporary BIGNUMs from a BN_CTX.
1876 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1880 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1881 include a #define from the old name to the new. The original intent
1882 was that statically linked binaries could for example just call
1883 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1884 link with digests. This never worked becayse SSLeay_add_all_digests()
1885 and SSLeay_add_all_ciphers() were in the same source file so calling
1886 one would link with the other. They are now in separate source files.
1889 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1892 *) Use a less unusual form of the Miller-Rabin primality test (it used
1893 a binary algorithm for exponentiation integrated into the Miller-Rabin
1894 loop, our standard modexp algorithms are faster).
1897 *) Support for the EBCDIC character set completed.
1898 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1900 *) Source code cleanups: use const where appropriate, eliminate casts,
1901 use void * instead of char * in lhash.
1904 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1905 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1906 this the server could overwrite ephemeral keys that the client
1910 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1911 using 50 iterations of the Rabin-Miller test.
1913 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1914 iterations of the Rabin-Miller test as required by the appendix
1915 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1916 As BN_is_prime_fasttest includes trial division, DSA parameter
1917 generation becomes much faster.
1919 This implies a change for the callback functions in DSA_is_prime
1920 and DSA_generate_parameters: The callback function is called once
1921 for each positive witness in the Rabin-Miller test, not just
1922 occasionally in the inner loop; and the parameters to the
1923 callback function now provide an iteration count for the outer
1924 loop rather than for the current invocation of the inner loop.
1925 DSA_generate_parameters additionally can call the callback
1926 function with an 'iteration count' of -1, meaning that a
1927 candidate has passed the trial division test (when q is generated
1928 from an application-provided seed, trial division is skipped).
1931 *) New function BN_is_prime_fasttest that optionally does trial
1932 division before starting the Rabin-Miller test and has
1933 an additional BN_CTX * argument (whereas BN_is_prime always
1934 has to allocate at least one BN_CTX).
1935 'callback(1, -1, cb_arg)' is called when a number has passed the
1936 trial division stage.
1939 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1943 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1946 *) New function BN_pseudo_rand().
1949 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1950 bignum version of BN_from_montgomery() with the working code from
1951 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1955 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1956 made it impossible to use the same SSL_SESSION data structure in
1957 SSL2 clients in multiple threads.
1960 *) The return value of RAND_load_file() no longer counts bytes obtained
1961 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1962 to seed the PRNG (previously an explicit byte count was required).
1963 [Ulf Möller, Bodo Möller]
1965 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1966 used (char *) instead of (void *) and had casts all over the place.
1969 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1972 *) Retain source code compatibility for BN_prime_checks macro:
1973 BN_is_prime(..., BN_prime_checks, ...) now uses
1974 BN_prime_checks_for_size to determine the appropriate number of
1975 Rabin-Miller iterations.
1978 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1979 DH_CHECK_P_NOT_SAFE_PRIME.
1980 (Check if this is true? OpenPGP calls them "strong".)
1983 *) Merge the functionality of "dh" and "gendh" programs into a new program
1984 "dhparam". The old programs are retained for now but will handle DH keys
1985 (instead of parameters) in future.
1988 *) Make the ciphers, s_server and s_client programs check the return values
1989 when a new cipher list is set.
1992 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1993 ciphers. Before when the 56bit ciphers were enabled the sorting was
1996 The syntax for the cipher sorting has been extended to support sorting by
1997 cipher-strength (using the strength_bits hard coded in the tables).
1998 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
2000 Fix a bug in the cipher-command parser: when supplying a cipher command
2001 string with an "undefined" symbol (neither command nor alphanumeric
2002 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
2003 an error is flagged.
2005 Due to the strength-sorting extension, the code of the
2006 ssl_create_cipher_list() function was completely rearranged. I hope that
2007 the readability was also increased :-)
2008 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2010 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2011 for the first serial number and places 2 in the serial number file. This
2012 avoids problems when the root CA is created with serial number zero and
2013 the first user certificate has the same issuer name and serial number
2017 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2018 the new code. Add documentation for this stuff.
2021 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2022 X509_*() to X509at_*() on the grounds that they don't handle X509
2023 structures and behave in an analagous way to the X509v3 functions:
2024 they shouldn't be called directly but wrapper functions should be used
2027 So we also now have some wrapper functions that call the X509at functions
2028 when passed certificate requests. (TO DO: similar things can be done with
2029 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2030 things. Some of these need some d2i or i2d and print functionality
2031 because they handle more complex structures.)
2034 *) Add missing #ifndefs that caused missing symbols when building libssl
2035 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2036 NO_RSA in ssl/s2*.c.
2037 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2039 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2040 has a return value which indicates the quality of the random data
2041 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2042 error queue. New function RAND_pseudo_bytes() generates output that is
2043 guaranteed to be unique but not unpredictable. RAND_add is like
2044 RAND_seed, but takes an extra argument for an entropy estimate
2045 (RAND_seed always assumes full entropy).
2048 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2049 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2050 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2051 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2052 false-positive rate of at most 2^-80 for random input.
2055 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2058 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2059 in the 0.9.5 release), this returns the chain
2060 from an X509_CTX structure with a dup of the stack and all
2061 the X509 reference counts upped: so the stack will exist
2062 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2065 Also make SSL_SESSION_print() print out the verify return
2069 *) Add manpage for the pkcs12 command. Also change the default
2070 behaviour so MAC iteration counts are used unless the new
2071 -nomaciter option is used. This improves file security and
2072 only older versions of MSIE (4.0 for example) need it.
2075 *) Honor the no-xxx Configure options when creating .DEF files.
2078 *) Add PKCS#10 attributes to field table: challengePassword,
2079 unstructuredName and unstructuredAddress. These are taken from
2080 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2081 international characters are used.
2083 More changes to X509_ATTRIBUTE code: allow the setting of types
2084 based on strings. Remove the 'loc' parameter when adding
2085 attributes because these will be a SET OF encoding which is sorted
2089 *) Initial changes to the 'req' utility to allow request generation
2090 automation. This will allow an application to just generate a template
2091 file containing all the field values and have req construct the
2094 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2095 used all over the place including certificate requests and PKCS#7
2096 structures. They are currently handled manually where necessary with
2097 some primitive wrappers for PKCS#7. The new functions behave in a
2098 manner analogous to the X509 extension functions: they allow
2099 attributes to be looked up by NID and added.
2101 Later something similar to the X509V3 code would be desirable to
2102 automatically handle the encoding, decoding and printing of the
2103 more complex types. The string types like challengePassword can
2104 be handled by the string table functions.
2106 Also modified the multi byte string table handling. Now there is
2107 a 'global mask' which masks out certain types. The table itself
2108 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2109 is useful when for example there is only one permissible type
2110 (as in countryName) and using the mask might result in no valid
2114 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2115 SSL_get_peer_finished to allow applications to obtain the latest
2116 Finished messages sent to the peer or expected from the peer,
2117 respectively. (SSL_get_peer_finished is usually the Finished message
2118 actually received from the peer, otherwise the protocol will be aborted.)
2120 As the Finished message are message digests of the complete handshake
2121 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2122 be used for external authentication procedures when the authentication
2123 provided by SSL/TLS is not desired or is not enough.
2126 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2127 the host supports BWX extension and if Compaq C is present on the
2128 $PATH. Just exploiting of the BWX extension results in 20-30%
2129 performance kick for some algorithms, e.g. DES and RC4 to mention
2130 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2134 *) Add support for MS "fast SGC". This is arguably a violation of the
2135 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2136 weak crypto and after checking the certificate is SGC a second one
2137 with strong crypto. MS SGC stops the first handshake after receiving
2138 the server certificate message and sends a second client hello. Since
2139 a server will typically do all the time consuming operations before
2140 expecting any further messages from the client (server key exchange
2141 is the most expensive) there is little difference between the two.
2143 To get OpenSSL to support MS SGC we have to permit a second client
2144 hello message after we have sent server done. In addition we have to
2145 reset the MAC if we do get this second client hello.
2148 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2149 if a DER encoded private key is RSA or DSA traditional format. Changed
2150 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2151 format DER encoded private key. Newer code should use PKCS#8 format which
2152 has the key type encoded in the ASN1 structure. Added DER private key
2153 support to pkcs8 application.
2156 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2157 ciphersuites has been selected (as required by the SSL 3/TLS 1
2158 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2159 is set, we interpret this as a request to violate the specification
2160 (the worst that can happen is a handshake failure, and 'correct'
2161 behaviour would result in a handshake failure anyway).
2164 *) In SSL_CTX_add_session, take into account that there might be multiple
2165 SSL_SESSION structures with the same session ID (e.g. when two threads
2166 concurrently obtain them from an external cache).
2167 The internal cache can handle only one SSL_SESSION with a given ID,
2168 so if there's a conflict, we now throw out the old one to achieve
2172 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2173 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2174 some routines that use cipher OIDs: some ciphers do not have OIDs
2175 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2179 *) Simplify the trust setting structure and code. Now we just have
2180 two sequences of OIDs for trusted and rejected settings. These will
2181 typically have values the same as the extended key usage extension
2182 and any application specific purposes.
2184 The trust checking code now has a default behaviour: it will just
2185 check for an object with the same NID as the passed id. Functions can
2186 be provided to override either the default behaviour or the behaviour
2187 for a given id. SSL client, server and email already have functions
2188 in place for compatibility: they check the NID and also return "trusted"
2189 if the certificate is self signed.
2192 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2193 traditional format into an EVP_PKEY structure.
2196 *) Add a password callback function PEM_cb() which either prompts for
2197 a password if usr_data is NULL or otherwise assumes it is a null
2198 terminated password. Allow passwords to be passed on command line
2199 environment or config files in a few more utilities.
2202 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2203 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2204 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2205 Update documentation.
2208 *) Support for ASN1 "NULL" type. This could be handled before by using
2209 ASN1_TYPE but there wasn't any function that would try to read a NULL
2210 and produce an error if it couldn't. For compatibility we also have
2211 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2212 don't allocate anything because they don't need to.
2215 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2217 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2219 *) Rebuild of the memory allocation routines used by OpenSSL code and
2220 possibly others as well. The purpose is to make an interface that
2221 provide hooks so anyone can build a separate set of allocation and
2222 deallocation routines to be used by OpenSSL, for example memory
2223 pool implementations, or something else, which was previously hard
2224 since Malloc(), Realloc() and Free() were defined as macros having
2225 the values malloc, realloc and free, respectively (except for Win32
2226 compilations). The same is provided for memory debugging code.
2227 OpenSSL already comes with functionality to find memory leaks, but
2228 this gives people a chance to debug other memory problems.
2230 With these changes, a new set of functions and macros have appeared:
2232 CRYPTO_set_mem_debug_functions() [F]
2233 CRYPTO_get_mem_debug_functions() [F]
2234 CRYPTO_dbg_set_options() [F]
2235 CRYPTO_dbg_get_options() [F]
2236 CRYPTO_malloc_debug_init() [M]
2238 The memory debug functions are NULL by default, unless the library
2239 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2240 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2241 gives the standard debugging functions that come with OpenSSL) or
2242 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2243 provided by the library user) must be used. When the standard
2244 debugging functions are used, CRYPTO_dbg_set_options can be used to
2245 request additional information:
2246 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2247 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2249 Also, things like CRYPTO_set_mem_functions will always give the
2250 expected result (the new set of functions is used for allocation
2251 and deallocation) at all times, regardless of platform and compiler
2254 To finish it up, some functions that were never use in any other
2255 way than through macros have a new API and new semantic:
2258 CRYPTO_dbg_realloc()
2261 All macros of value have retained their old syntax.
2262 [Richard Levitte and Bodo Moeller]
2264 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2265 ordering of SMIMECapabilities wasn't in "strength order" and there
2266 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2270 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2271 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2272 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2274 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2275 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2276 functionality to handle multipart/signed properly) and a utility
2277 called 'smime' to call all this stuff. This is based on code I
2278 originally wrote for Celo who have kindly allowed it to be
2279 included in OpenSSL.
2282 *) Add variants des_set_key_checked and des_set_key_unchecked of
2283 des_set_key (aka des_key_sched). Global variable des_check_key
2284 decides which of these is called by des_set_key; this way
2285 des_check_key behaves as it always did, but applications and
2286 the library itself, which was buggy for des_check_key == 1,
2287 have a cleaner way to pick the version they need.
2290 *) New function PKCS12_newpass() which changes the password of a
2294 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2295 dynamic mix. In both cases the ids can be used as an index into the
2296 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2297 functions so they accept a list of the field values and the
2298 application doesn't need to directly manipulate the X509_TRUST
2302 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2306 *) Modify the way the V3 extension code looks up extensions. This now
2307 works in a similar way to the object code: we have some "standard"
2308 extensions in a static table which is searched with OBJ_bsearch()
2309 and the application can add dynamic ones if needed. The file
2310 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2311 updated whenever a new extension is added to the core code and kept
2312 in ext_nid order. There is a simple program 'tabtest.c' which checks
2313 this. New extensions are not added too often so this file can readily
2314 be maintained manually.
2316 There are two big advantages in doing things this way. The extensions
2317 can be looked up immediately and no longer need to be "added" using
2318 X509V3_add_standard_extensions(): this function now does nothing.
2319 [Side note: I get *lots* of email saying the extension code doesn't
2320 work because people forget to call this function]
2321 Also no dynamic allocation is done unless new extensions are added:
2322 so if we don't add custom extensions there is no need to call
2323 X509V3_EXT_cleanup().
2326 *) Modify enc utility's salting as follows: make salting the default. Add a
2327 magic header, so unsalted files fail gracefully instead of just decrypting
2328 to garbage. This is because not salting is a big security hole, so people
2329 should be discouraged from doing it.
2332 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2333 digest to be passed on the command line but it only used this
2334 parameter when signing a certificate. Modified so all relevant
2335 operations are affected by the digest parameter including the
2336 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2337 DSA key was used because it didn't fix the digest.
2340 *) Initial certificate chain verify code. Currently tests the untrusted
2341 certificates for consistency with the verify purpose (which is set
2342 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2344 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2345 this is because it will reject chains with invalid extensions whereas
2346 every previous version of OpenSSL and SSLeay made no checks at all.
2348 Trust code: checks the root CA for the relevant trust settings. Trust
2349 settings have an initial value consistent with the verify purpose: e.g.
2350 if the verify purpose is for SSL client use it expects the CA to be
2351 trusted for SSL client use. However the default value can be changed to
2352 permit custom trust settings: one example of this would be to only trust
2353 certificates from a specific "secure" set of CAs.
2355 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2356 which should be used for version portability: especially since the
2357 verify structure is likely to change more often now.
2359 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2360 to set them. If not set then assume SSL clients will verify SSL servers
2363 Two new options to the verify program: -untrusted allows a set of
2364 untrusted certificates to be passed in and -purpose which sets the
2365 intended purpose of the certificate. If a purpose is set then the
2366 new chain verify code is used to check extension consistency.
2369 *) Support for the authority information access extension.
2372 *) Modify RSA and DSA PEM read routines to transparently handle
2373 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2374 public keys in a format compatible with certificate
2375 SubjectPublicKeyInfo structures. Unfortunately there were already
2376 functions called *_PublicKey_* which used various odd formats so
2377 these are retained for compatibility: however the DSA variants were
2378 never in a public release so they have been deleted. Changed dsa/rsa
2379 utilities to handle the new format: note no releases ever handled public
2380 keys so we should be OK.
2382 The primary motivation for this change is to avoid the same fiasco
2383 that dogs private keys: there are several incompatible private key
2384 formats some of which are standard and some OpenSSL specific and
2385 require various evil hacks to allow partial transparent handling and
2386 even then it doesn't work with DER formats. Given the option anything
2387 other than PKCS#8 should be dumped: but the other formats have to
2388 stay in the name of compatibility.
2390 With public keys and the benefit of hindsight one standard format
2391 is used which works with EVP_PKEY, RSA or DSA structures: though
2392 it clearly returns an error if you try to read the wrong kind of key.
2394 Added a -pubkey option to the 'x509' utility to output the public key.
2395 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2396 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2397 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2398 that do the same as the EVP_PKEY_assign_*() except they up the
2399 reference count of the added key (they don't "swallow" the
2403 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2404 CRLs would fail if the file contained no certificates or no CRLs:
2405 added a new function to read in both types and return the number
2406 read: this means that if none are read it will be an error. The
2407 DER versions of the certificate and CRL reader would always fail
2408 because it isn't possible to mix certificates and CRLs in DER format
2409 without choking one or the other routine. Changed this to just read
2410 a certificate: this is the best we can do. Also modified the code
2411 in apps/verify.c to take notice of return codes: it was previously
2412 attempting to read in certificates from NULL pointers and ignoring
2413 any errors: this is one reason why the cert and CRL reader seemed
2414 to work. It doesn't check return codes from the default certificate
2415 routines: these may well fail if the certificates aren't installed.
2418 *) Code to support otherName option in GeneralName.
2421 *) First update to verify code. Change the verify utility
2422 so it warns if it is passed a self signed certificate:
2423 for consistency with the normal behaviour. X509_verify
2424 has been modified to it will now verify a self signed
2425 certificate if *exactly* the same certificate appears
2426 in the store: it was previously impossible to trust a
2427 single self signed certificate. This means that:
2428 openssl verify ss.pem
2429 now gives a warning about a self signed certificate but
2430 openssl verify -CAfile ss.pem ss.pem
2434 *) For servers, store verify_result in SSL_SESSION data structure
2435 (and add it to external session representation).
2436 This is needed when client certificate verifications fails,
2437 but an application-provided verification callback (set by
2438 SSL_CTX_set_cert_verify_callback) allows accepting the session
2439 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2440 but returns 1): When the session is reused, we have to set
2441 ssl->verify_result to the appropriate error code to avoid
2443 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2445 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2446 case in PKCS7_dataInit() where the signed PKCS7 structure
2447 didn't contain any existing data because it was being created.
2448 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2450 *) Add a salt to the key derivation routines in enc.c. This
2451 forms the first 8 bytes of the encrypted file. Also add a
2452 -S option to allow a salt to be input on the command line.
2455 *) New function X509_cmp(). Oddly enough there wasn't a function
2456 to compare two certificates. We do this by working out the SHA1
2457 hash and comparing that. X509_cmp() will be needed by the trust
2461 *) SSL_get1_session() is like SSL_get_session(), but increments
2462 the reference count in the SSL_SESSION returned.
2463 [Geoff Thorpe <geoff@eu.c2.net>]
2465 *) Fix for 'req': it was adding a null to request attributes.
2466 Also change the X509_LOOKUP and X509_INFO code to handle
2467 certificate auxiliary information.
2470 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2474 *) Add the possibility to add extra information to the memory leak
2475 detecting output, to form tracebacks, showing from where each
2476 allocation was originated: CRYPTO_push_info("constant string") adds
2477 the string plus current file name and line number to a per-thread
2478 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2479 is like calling CYRPTO_pop_info() until the stack is empty.
2480 Also updated memory leak detection code to be multi-thread-safe.
2483 *) Add options -text and -noout to pkcs7 utility and delete the
2484 encryption options which never did anything. Update docs.
2487 *) Add options to some of the utilities to allow the pass phrase
2488 to be included on either the command line (not recommended on
2489 OSes like Unix) or read from the environment. Update the
2490 manpages and fix a few bugs.
2493 *) Add a few manpages for some of the openssl commands.
2496 *) Fix the -revoke option in ca. It was freeing up memory twice,
2497 leaking and not finding already revoked certificates.
2500 *) Extensive changes to support certificate auxiliary information.
2501 This involves the use of X509_CERT_AUX structure and X509_AUX
2502 functions. An X509_AUX function such as PEM_read_X509_AUX()
2503 can still read in a certificate file in the usual way but it
2504 will also read in any additional "auxiliary information". By
2505 doing things this way a fair degree of compatibility can be
2506 retained: existing certificates can have this information added
2507 using the new 'x509' options.
2509 Current auxiliary information includes an "alias" and some trust
2510 settings. The trust settings will ultimately be used in enhanced
2511 certificate chain verification routines: currently a certificate
2512 can only be trusted if it is self signed and then it is trusted
2516 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2517 The problem was that one of the replacement routines had not been working
2518 since SSLeay releases. For now the offending routine has been replaced
2519 with non-optimised assembler. Even so, this now gives around 95%
2520 performance improvement for 1024 bit RSA signs.
2523 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2524 handling. Most clients have the effective key size in bits equal to
2525 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2526 A few however don't do this and instead use the size of the decrypted key
2527 to determine the RC2 key length and the AlgorithmIdentifier to determine
2528 the effective key length. In this case the effective key length can still
2529 be 40 bits but the key length can be 168 bits for example. This is fixed
2530 by manually forcing an RC2 key into the EVP_PKEY structure because the
2531 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2532 the key length and effective key length are equal.
2535 *) Add a bunch of functions that should simplify the creation of
2536 X509_NAME structures. Now you should be able to do:
2537 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2538 and have it automatically work out the correct field type and fill in
2539 the structures. The more adventurous can try:
2540 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2541 and it will (hopefully) work out the correct multibyte encoding.
2544 *) Change the 'req' utility to use the new field handling and multibyte
2545 copy routines. Before the DN field creation was handled in an ad hoc
2546 way in req, ca, and x509 which was rather broken and didn't support
2547 BMPStrings or UTF8Strings. Since some software doesn't implement
2548 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2549 using the dirstring_type option. See the new comment in the default
2550 openssl.cnf for more info.
2553 *) Make crypto/rand/md_rand.c more robust:
2554 - Assure unique random numbers after fork().
2555 - Make sure that concurrent threads access the global counter and
2556 md serializably so that we never lose entropy in them
2557 or use exactly the same state in multiple threads.
2558 Access to the large state is not always serializable because
2559 the additional locking could be a performance killer, and
2560 md should be large enough anyway.
2563 *) New file apps/app_rand.c with commonly needed functionality
2564 for handling the random seed file.
2566 Use the random seed file in some applications that previously did not:
2568 dsaparam -genkey (which also ignored its '-rand' option),
2571 x509 (when signing).
2572 Except on systems with /dev/urandom, it is crucial to have a random
2573 seed file at least for key creation, DSA signing, and for DH exchanges;
2574 for RSA signatures we could do without one.
2576 gendh and gendsa (unlike genrsa) used to read only the first byte
2577 of each file listed in the '-rand' option. The function as previously
2578 found in genrsa is now in app_rand.c and is used by all programs
2579 that support '-rand'.
2582 *) In RAND_write_file, use mode 0600 for creating files;
2583 don't just chmod when it may be too late.
2586 *) Report an error from X509_STORE_load_locations
2587 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2590 *) New function ASN1_mbstring_copy() this copies a string in either
2591 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2592 into an ASN1_STRING type. A mask of permissible types is passed
2593 and it chooses the "minimal" type to use or an error if not type
2597 *) Add function equivalents to the various macros in asn1.h. The old
2598 macros are retained with an M_ prefix. Code inside the library can
2599 use the M_ macros. External code (including the openssl utility)
2600 should *NOT* in order to be "shared library friendly".
2603 *) Add various functions that can check a certificate's extensions
2604 to see if it usable for various purposes such as SSL client,
2605 server or S/MIME and CAs of these types. This is currently
2606 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2607 verification. Also added a -purpose flag to x509 utility to
2608 print out all the purposes.
2611 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2615 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2616 for, obtain and decode and extension and obtain its critical flag.
2617 This allows all the necessary extension code to be handled in a
2618 single function call.
2621 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2622 platforms. See crypto/rc4/rc4_enc.c for further details.
2625 *) New -noout option to asn1parse. This causes no output to be produced
2626 its main use is when combined with -strparse and -out to extract data
2627 from a file (which may not be in ASN.1 format).
2630 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2631 when producing the local key id.
2632 [Richard Levitte <levitte@stacken.kth.se>]
2634 *) New option -dhparam in s_server. This allows a DH parameter file to be
2635 stated explicitly. If it is not stated then it tries the first server
2636 certificate file. The previous behaviour hard coded the filename
2640 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2641 a public key to be input or output. For example:
2642 openssl rsa -in key.pem -pubout -out pubkey.pem
2643 Also added necessary DSA public key functions to handle this.
2646 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2647 in the message. This was handled by allowing
2648 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2649 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2651 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2652 to the end of the strings whereas this didn't. This would cause problems
2653 if strings read with d2i_ASN1_bytes() were later modified.
2654 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2656 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2657 data and it contains EOF it will end up returning an error. This is
2658 caused by input 46 bytes long. The cause is due to the way base64
2659 BIOs find the start of base64 encoded data. They do this by trying a
2660 trial decode on each line until they find one that works. When they
2661 do a flag is set and it starts again knowing it can pass all the
2662 data directly through the decoder. Unfortunately it doesn't reset
2663 the context it uses. This means that if EOF is reached an attempt
2664 is made to pass two EOFs through the context and this causes the
2665 resulting error. This can also cause other problems as well. As is
2666 usual with these problems it takes *ages* to find and the fix is
2667 trivial: move one line.
2668 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2670 *) Ugly workaround to get s_client and s_server working under Windows. The
2671 old code wouldn't work because it needed to select() on sockets and the
2672 tty (for keypresses and to see if data could be written). Win32 only
2673 supports select() on sockets so we select() with a 1s timeout on the
2674 sockets and then see if any characters are waiting to be read, if none
2675 are present then we retry, we also assume we can always write data to
2676 the tty. This isn't nice because the code then blocks until we've
2677 received a complete line of data and it is effectively polling the
2678 keyboard at 1s intervals: however it's quite a bit better than not
2679 working at all :-) A dedicated Windows application might handle this
2680 with an event loop for example.
2683 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2684 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2685 will be called when RSA_sign() and RSA_verify() are used. This is useful
2686 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2687 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2688 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2689 This necessitated the support of an extra signature type NID_md5_sha1
2690 for SSL signatures and modifications to the SSL library to use it instead
2691 of calling RSA_public_decrypt() and RSA_private_encrypt().
2694 *) Add new -verify -CAfile and -CApath options to the crl program, these
2695 will lookup a CRL issuers certificate and verify the signature in a
2696 similar way to the verify program. Tidy up the crl program so it
2697 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2698 less strict. It will now permit CRL extensions even if it is not
2699 a V2 CRL: this will allow it to tolerate some broken CRLs.
2702 *) Initialize all non-automatic variables each time one of the openssl
2703 sub-programs is started (this is necessary as they may be started
2704 multiple times from the "OpenSSL>" prompt).
2705 [Lennart Bang, Bodo Moeller]
2707 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2708 removing all other RSA functionality (this is what NO_RSA does). This
2709 is so (for example) those in the US can disable those operations covered
2710 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2714 *) Non-copying interface to BIO pairs.
2715 (still largely untested)
2718 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2719 ASCII string. This was handled independently in various places before.
2722 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2723 UTF8 strings a character at a time.
2726 *) Use client_version from client hello to select the protocol
2727 (s23_srvr.c) and for RSA client key exchange verification
2728 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2731 *) Add various utility functions to handle SPKACs, these were previously
2732 handled by poking round in the structure internals. Added new function
2733 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2734 print, verify and generate SPKACs. Based on an original idea from
2735 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2738 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2741 *) Allow the config file extension section to be overwritten on the
2742 command line. Based on an original idea from Massimiliano Pala
2743 <madwolf@comune.modena.it>. The new option is called -extensions
2744 and can be applied to ca, req and x509. Also -reqexts to override
2745 the request extensions in req and -crlexts to override the crl extensions
2749 *) Add new feature to the SPKAC handling in ca. Now you can include
2750 the same field multiple times by preceding it by "XXXX." for example:
2753 this is the same syntax as used in the req config file.
2756 *) Allow certificate extensions to be added to certificate requests. These
2757 are specified in a 'req_extensions' option of the req section of the
2758 config file. They can be printed out with the -text option to req but
2759 are otherwise ignored at present.
2762 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2763 data read consists of only the final block it would not decrypted because
2764 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2765 A misplaced 'break' also meant the decrypted final block might not be
2766 copied until the next read.
2769 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2770 a few extra parameters to the DH structure: these will be useful if
2771 for example we want the value of 'q' or implement X9.42 DH.
2774 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2775 provides hooks that allow the default DSA functions or functions on a
2776 "per key" basis to be replaced. This allows hardware acceleration and
2777 hardware key storage to be handled without major modification to the
2778 library. Also added low level modexp hooks and CRYPTO_EX structure and
2779 associated functions.
2782 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2783 as "read only": it can't be written to and the buffer it points to will
2784 not be freed. Reading from a read only BIO is much more efficient than
2785 a normal memory BIO. This was added because there are several times when
2786 an area of memory needs to be read from a BIO. The previous method was
2787 to create a memory BIO and write the data to it, this results in two
2788 copies of the data and an O(n^2) reading algorithm. There is a new
2789 function BIO_new_mem_buf() which creates a read only memory BIO from
2790 an area of memory. Also modified the PKCS#7 routines to use read only
2794 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2795 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2796 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2797 but a retry condition occured while trying to read the rest.
2800 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2801 NID_pkcs7_encrypted by default: this was wrong since this should almost
2802 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2803 the encrypted data type: this is a more sensible place to put it and it
2804 allows the PKCS#12 code to be tidied up that duplicated this
2808 *) Changed obj_dat.pl script so it takes its input and output files on
2809 the command line. This should avoid shell escape redirection problems
2813 *) Initial support for certificate extension requests, these are included
2814 in things like Xenroll certificate requests. Included functions to allow
2815 extensions to be obtained and added.
2818 *) -crlf option to s_client and s_server for sending newlines as
2819 CRLF (as required by many protocols).
2822 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2824 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2825 [Ralf S. Engelschall]
2827 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2828 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2830 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2834 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2835 DH parameters/keys (q is lost during that conversion, but the resulting
2836 DH parameters contain its length).
2838 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2839 much faster than DH_generate_parameters (which creates parameters
2840 where p = 2*q + 1), and also the smaller q makes DH computations
2841 much more efficient (160-bit exponentiation instead of 1024-bit
2842 exponentiation); so this provides a convenient way to support DHE
2843 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2844 utter importance to use
2845 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2847 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2848 when such DH parameters are used, because otherwise small subgroup
2849 attacks may become possible!
2852 *) Avoid memory leak in i2d_DHparams.
2855 *) Allow the -k option to be used more than once in the enc program:
2856 this allows the same encrypted message to be read by multiple recipients.
2859 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2860 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2861 it will always use the numerical form of the OID, even if it has a short
2865 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2866 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2867 otherwise bn_mod_exp was called. In the case of hardware keys for example
2868 no private key components need be present and it might store extra data
2869 in the RSA structure, which cannot be accessed from bn_mod_exp.
2870 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2871 private key operations.
2874 *) Added support for SPARC Linux.
2877 *) pem_password_cb function type incompatibly changed from
2878 typedef int pem_password_cb(char *buf, int size, int rwflag);
2880 ....(char *buf, int size, int rwflag, void *userdata);
2881 so that applications can pass data to their callbacks:
2882 The PEM[_ASN1]_{read,write}... functions and macros now take an
2883 additional void * argument, which is just handed through whenever
2884 the password callback is called.
2885 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2887 New function SSL_CTX_set_default_passwd_cb_userdata.
2889 Compatibility note: As many C implementations push function arguments
2890 onto the stack in reverse order, the new library version is likely to
2891 interoperate with programs that have been compiled with the old
2892 pem_password_cb definition (PEM_whatever takes some data that
2893 happens to be on the stack as its last argument, and the callback
2894 just ignores this garbage); but there is no guarantee whatsoever that
2897 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2898 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2899 problems not only on Windows, but also on some Unix platforms.
2900 To avoid problematic command lines, these definitions are now in an
2901 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2902 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2905 *) MIPS III/IV assembler module is reimplemented.
2908 *) More DES library cleanups: remove references to srand/rand and
2909 delete an unused file.
2912 *) Add support for the the free Netwide assembler (NASM) under Win32,
2913 since not many people have MASM (ml) and it can be hard to obtain.
2914 This is currently experimental but it seems to work OK and pass all
2915 the tests. Check out INSTALL.W32 for info.
2918 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2919 without temporary keys kept an extra copy of the server key,
2920 and connections with temporary keys did not free everything in case
2924 *) New function RSA_check_key and new openssl rsa option -check
2925 for verifying the consistency of RSA keys.
2926 [Ulf Moeller, Bodo Moeller]
2928 *) Various changes to make Win32 compile work:
2929 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2930 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2931 comparison" warnings.
2932 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2935 *) Add a debugging option to PKCS#5 v2 key generation function: when
2936 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2937 derived keys are printed to stderr.
2940 *) Copy the flags in ASN1_STRING_dup().
2941 [Roman E. Pavlov <pre@mo.msk.ru>]
2943 *) The x509 application mishandled signing requests containing DSA
2944 keys when the signing key was also DSA and the parameters didn't match.
2946 It was supposed to omit the parameters when they matched the signing key:
2947 the verifying software was then supposed to automatically use the CA's
2948 parameters if they were absent from the end user certificate.
2950 Omitting parameters is no longer recommended. The test was also
2951 the wrong way round! This was probably due to unusual behaviour in
2952 EVP_cmp_parameters() which returns 1 if the parameters match.
2953 This meant that parameters were omitted when they *didn't* match and
2954 the certificate was useless. Certificates signed with 'ca' didn't have
2956 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2958 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2959 The interface is as follows:
2960 Applications can use
2961 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2962 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2963 "off" is now the default.
2964 The library internally uses
2965 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2966 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2967 to disable memory-checking temporarily.
2969 Some inconsistent states that previously were possible (and were
2970 even the default) are now avoided.
2972 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2973 with each memory chunk allocated; this is occasionally more helpful
2974 than just having a counter.
2976 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2978 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2982 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2983 which largely parallels "options", but is for changing API behaviour,
2984 whereas "options" are about protocol behaviour.
2985 Initial "mode" flags are:
2987 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2988 a single record has been written.
2989 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2990 retries use the same buffer location.
2991 (But all of the contents must be
2995 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2998 *) Fix problems with no-hmac etc.
2999 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
3001 *) New functions RSA_get_default_method(), RSA_set_method() and
3002 RSA_get_method(). These allows replacement of RSA_METHODs without having
3003 to mess around with the internals of an RSA structure.
3006 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3007 Also really enable memory leak checks in openssl.c and in some
3009 [Chad C. Mulligan, Bodo Moeller]
3011 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3012 up the length of negative integers. This has now been simplified to just
3013 store the length when it is first determined and use it later, rather
3014 than trying to keep track of where data is copied and updating it to
3016 [Steve Henson, reported by Brien Wheeler
3017 <bwheeler@authentica-security.com>]
3019 *) Add a new function PKCS7_signatureVerify. This allows the verification
3020 of a PKCS#7 signature but with the signing certificate passed to the
3021 function itself. This contrasts with PKCS7_dataVerify which assumes the
3022 certificate is present in the PKCS#7 structure. This isn't always the
3023 case: certificates can be omitted from a PKCS#7 structure and be
3024 distributed by "out of band" means (such as a certificate database).
3027 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3028 function prototypes in pem.h, also change util/mkdef.pl to add the
3029 necessary function names.
3032 *) mk1mf.pl (used by Windows builds) did not properly read the
3033 options set by Configure in the top level Makefile, and Configure
3034 was not even able to write more than one option correctly.
3035 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3038 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3039 file to be loaded from a BIO or FILE pointer. The BIO version will
3040 for example allow memory BIOs to contain config info.
3043 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3044 Whoever hopes to achieve shared-library compatibility across versions
3045 must use this, not the compile-time macro.
3046 (Exercise 0.9.4: Which is the minimum library version required by
3048 Note: All this applies only to multi-threaded programs, others don't
3052 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3053 through a BIO pair triggered the default case, i.e.
3054 SSLerr(...,SSL_R_UNKNOWN_STATE).
3057 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3058 can use the SSL library even if none of the specific BIOs is
3062 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3063 for the encoded length.
3064 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3066 *) Add initial documentation of the X509V3 functions.
3069 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3070 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3071 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3072 secure PKCS#8 private key format with a high iteration count.
3075 *) Fix determination of Perl interpreter: A perl or perl5
3076 _directory_ in $PATH was also accepted as the interpreter.
3077 [Ralf S. Engelschall]
3079 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3080 wrong with it but it was very old and did things like calling
3081 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3085 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3086 to use the new extension code.
3089 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3090 with macros. This should make it easier to change their form, add extra
3091 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3095 *) Add to configuration table a new entry that can specify an alternative
3096 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3097 according to Mark Crispin <MRC@Panda.COM>.
3101 *) DES CBC did not update the IV. Weird.
3104 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3105 Changing the behaviour of the former might break existing programs --
3106 where IV updating is needed, des_ncbc_encrypt can be used.
3109 *) When bntest is run from "make test" it drives bc to check its
3110 calculations, as well as internally checking them. If an internal check
3111 fails, it needs to cause bc to give a non-zero result or make test carries
3112 on without noticing the failure. Fixed.
3115 *) DES library cleanups.
3118 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3119 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3120 ciphers. NOTE: although the key derivation function has been verified
3121 against some published test vectors it has not been extensively tested
3122 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3126 *) Instead of "mkdir -p", which is not fully portable, use new
3127 Perl script "util/mkdir-p.pl".
3130 *) Rewrite the way password based encryption (PBE) is handled. It used to
3131 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3132 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3133 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3134 the 'parameter' field of the AlgorithmIdentifier is passed to the
3135 underlying key generation function so it must do its own ASN1 parsing.
3136 This has also changed the EVP_PBE_CipherInit() function which now has a
3137 'parameter' argument instead of literal salt and iteration count values
3138 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3141 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3142 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3143 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3144 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3145 value was just used as a "magic string" and not used directly its
3146 value doesn't matter.
3149 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3153 *) "linux-sparc64" configuration (ultrapenguin).
3154 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3155 "linux-sparc" configuration.
3156 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3158 *) config now generates no-xxx options for missing ciphers.
3161 *) Support the EBCDIC character set (work in progress).
3162 File ebcdic.c not yet included because it has a different license.
3163 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3165 *) Support BS2000/OSD-POSIX.
3166 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3168 *) Make callbacks for key generation use void * instead of char *.
3171 *) Make S/MIME samples compile (not yet tested).
3174 *) Additional typesafe stacks.
3177 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3181 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3183 *) New configuration variant "sco5-gcc".
3185 *) Updated some demos.
3186 [Sean O Riordain, Wade Scholine]
3188 *) Add missing BIO_free at exit of pkcs12 application.
3191 *) Fix memory leak in conf.c.
3194 *) Updates for Win32 to assembler version of MD5.
3197 *) Set #! path to perl in apps/der_chop to where we found it
3198 instead of using a fixed path.
3201 *) SHA library changes for irix64-mips4-cc.
3204 *) Improvements for VMS support.
3208 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3210 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3211 This also avoids the problems with SC4.2 and unpatched SC5.
3212 [Andy Polyakov <appro@fy.chalmers.se>]
3214 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3215 These are required because of the typesafe stack would otherwise break
3216 existing code. If old code used a structure member which used to be STACK
3217 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3218 sk_num or sk_value it would produce an error because the num, data members
3219 are not present in STACK_OF. Now it just produces a warning. sk_set
3220 replaces the old method of assigning a value to sk_value
3221 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3222 that does this will no longer work (and should use sk_set instead) but
3223 this could be regarded as a "questionable" behaviour anyway.
3226 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3227 correctly handle encrypted S/MIME data.
3230 *) Change type of various DES function arguments from des_cblock
3231 (which means, in function argument declarations, pointer to char)
3232 to des_cblock * (meaning pointer to array with 8 char elements),
3233 which allows the compiler to do more typechecking; it was like
3234 that back in SSLeay, but with lots of ugly casts.
3236 Introduce new type const_des_cblock.
3239 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3240 problems: find RecipientInfo structure that matches recipient certificate
3241 and initialise the ASN1 structures properly based on passed cipher.
3244 *) Belatedly make the BN tests actually check the results.
3247 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3248 to and from BNs: it was completely broken. New compilation option
3249 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3250 key elements as negative integers.
3253 *) Reorganize and speed up MD5.
3254 [Andy Polyakov <appro@fy.chalmers.se>]
3257 [Richard Levitte <richard@levitte.org>]
3259 *) New option -out to asn1parse to allow the parsed structure to be
3260 output to a file. This is most useful when combined with the -strparse
3261 option to examine the output of things like OCTET STRINGS.
3264 *) Make SSL library a little more fool-proof by not requiring any longer
3265 that SSL_set_{accept,connect}_state be called before
3266 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3267 in many applications because usually everything *appeared* to work as
3268 intended anyway -- now it really works as intended).
3271 *) Move openssl.cnf out of lib/.
3274 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3275 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3276 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3277 [Ralf S. Engelschall]
3279 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3280 handle PKCS#7 enveloped data properly.
3281 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3283 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3284 copying pointers. The cert_st handling is changed by this in
3285 various ways (and thus what used to be known as ctx->default_cert
3286 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3287 any longer when s->cert does not give us what we need).
3288 ssl_cert_instantiate becomes obsolete by this change.
3289 As soon as we've got the new code right (possibly it already is?),
3290 we have solved a couple of bugs of the earlier code where s->cert
3291 was used as if it could not have been shared with other SSL structures.
3293 Note that using the SSL API in certain dirty ways now will result
3294 in different behaviour than observed with earlier library versions:
3295 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3296 does not influence s as it used to.
3298 In order to clean up things more thoroughly, inside SSL_SESSION
3299 we don't use CERT any longer, but a new structure SESS_CERT
3300 that holds per-session data (if available); currently, this is
3301 the peer's certificate chain and, for clients, the server's certificate
3302 and temporary key. CERT holds only those values that can have
3303 meaningful defaults in an SSL_CTX.
3306 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3307 from the internal representation. Various PKCS#7 fixes: remove some
3308 evil casts and set the enc_dig_alg field properly based on the signing
3312 *) Allow PKCS#12 password to be set from the command line or the
3313 environment. Let 'ca' get its config file name from the environment
3314 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3318 *) Allow certificate policies extension to use an IA5STRING for the
3319 organization field. This is contrary to the PKIX definition but
3320 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3324 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3325 without disallowing inline assembler and the like for non-pedantic builds.
3328 *) Support Borland C++ builder.
3329 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3334 *) SHA-1 cleanups and performance enhancements.
3335 [Andy Polyakov <appro@fy.chalmers.se>]
3337 *) Sparc v8plus assembler for the bignum library.
3338 [Andy Polyakov <appro@fy.chalmers.se>]
3340 *) Accept any -xxx and +xxx compiler options in Configure.
3343 *) Update HPUX configuration.
3346 *) Add missing sk_<type>_unshift() function to safestack.h
3347 [Ralf S. Engelschall]
3349 *) New function SSL_CTX_use_certificate_chain_file that sets the
3350 "extra_cert"s in addition to the certificate. (This makes sense
3351 only for "PEM" format files, as chains as a whole are not
3355 *) Support verify_depth from the SSL API.
3356 x509_vfy.c had what can be considered an off-by-one-error:
3357 Its depth (which was not part of the external interface)
3358 was actually counting the number of certificates in a chain;
3359 now it really counts the depth.
3362 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3363 instead of X509err, which often resulted in confusing error
3364 messages since the error codes are not globally unique
3365 (e.g. an alleged error in ssl3_accept when a certificate
3366 didn't match the private key).
3368 *) New function SSL_CTX_set_session_id_context that allows to set a default
3369 value (so that you don't need SSL_set_session_id_context for each
3370 connection using the SSL_CTX).
3373 *) OAEP decoding bug fix.
3376 *) Support INSTALL_PREFIX for package builders, as proposed by
3380 *) New Configure options "threads" and "no-threads". For systems
3381 where the proper compiler options are known (currently Solaris
3382 and Linux), "threads" is the default.
3385 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3388 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3389 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3390 such as /usr/local/bin.
3393 *) "make linux-shared" to build shared libraries.
3394 [Niels Poppe <niels@netbox.org>]
3396 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3399 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3400 extension adding in x509 utility.
3403 *) Remove NOPROTO sections and error code comments.
3406 *) Partial rewrite of the DEF file generator to now parse the ANSI
3410 *) New Configure options --prefix=DIR and --openssldir=DIR.
3413 *) Complete rewrite of the error code script(s). It is all now handled
3414 by one script at the top level which handles error code gathering,
3415 header rewriting and C source file generation. It should be much better
3416 than the old method: it now uses a modified version of Ulf's parser to
3417 read the ANSI prototypes in all header files (thus the old K&R definitions
3418 aren't needed for error creation any more) and do a better job of
3419 translating function codes into names. The old 'ASN1 error code imbedded
3420 in a comment' is no longer necessary and it doesn't use .err files which
3421 have now been deleted. Also the error code call doesn't have to appear all
3422 on one line (which resulted in some large lines...).
3425 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3428 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3429 0 (which usually indicates a closed connection), but continue reading.
3432 *) Fix some race conditions.
3435 *) Add support for CRL distribution points extension. Add Certificate
3436 Policies and CRL distribution points documentation.
3439 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3442 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3443 8 of keying material. Merlin has also confirmed interop with this fix
3444 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3445 [Merlin Hughes <merlin@baltimore.ie>]
3447 *) Fix lots of warnings.
3448 [Richard Levitte <levitte@stacken.kth.se>]
3450 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3451 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3452 [Richard Levitte <levitte@stacken.kth.se>]
3454 *) Fix problems with sizeof(long) == 8.
3455 [Andy Polyakov <appro@fy.chalmers.se>]
3457 *) Change functions to ANSI C.
3460 *) Fix typos in error codes.
3461 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3463 *) Remove defunct assembler files from Configure.
3466 *) SPARC v8 assembler BIGNUM implementation.
3467 [Andy Polyakov <appro@fy.chalmers.se>]
3469 *) Support for Certificate Policies extension: both print and set.
3470 Various additions to support the r2i method this uses.
3473 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3474 return a const string when you are expecting an allocated buffer.
3477 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3478 types DirectoryString and DisplayText.
3481 *) Add code to allow r2i extensions to access the configuration database,
3482 add an LHASH database driver and add several ctx helper functions.
3485 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3486 fail when they extended the size of a BIGNUM.
3489 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3490 support typesafe stack.
3493 *) Fix typo in SSL_[gs]et_options().
3494 [Nils Frostberg <nils@medcom.se>]
3496 *) Delete various functions and files that belonged to the (now obsolete)
3497 old X509V3 handling code.
3500 *) New Configure option "rsaref".
3503 *) Don't auto-generate pem.h.
3506 *) Introduce type-safe ASN.1 SETs.
3509 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3510 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3512 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3513 that links with OpenSSL (well at least cause lots of warnings), but fear
3514 not: the conversion is trivial, and it eliminates loads of evil casts. A
3515 few STACKed things have been converted already. Feel free to convert more.
3516 In the fullness of time, I'll do away with the STACK type altogether.
3519 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3520 specified in <certfile> by updating the entry in the index.txt file.
3521 This way one no longer has to edit the index.txt file manually for
3522 revoking a certificate. The -revoke option does the gory details now.
3523 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3525 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3526 `-text' option at all and this way the `-noout -text' combination was
3527 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3528 [Ralf S. Engelschall]
3530 *) Make sure a corresponding plain text error message exists for the
3531 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3532 verify callback function determined that a certificate was revoked.
3533 [Ralf S. Engelschall]
3535 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3536 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3537 all available cipers including rc5, which was forgotten until now.
3538 In order to let the testing shell script know which algorithms
3539 are available, a new (up to now undocumented) command
3540 "openssl list-cipher-commands" is used.
3543 *) Bugfix: s_client occasionally would sleep in select() when
3544 it should have checked SSL_pending() first.
3547 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3548 the raw DSA values prior to ASN.1 encoding.
3551 *) Tweaks to Configure
3552 [Niels Poppe <niels@netbox.org>]
3554 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3558 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3561 *) New config option to avoid instructions that are illegal on the 80386.
3562 The default code is faster, but requires at least a 486.
3565 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3566 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3567 same as SSL2_VERSION anyway.
3570 *) New "-showcerts" option for s_client.
3573 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3574 application. Various cleanups and fixes.
3577 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3578 modify error routines to work internally. Add error codes and PBE init
3579 to library startup routines.
projects / openssl.git / blob