4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Function OCSP_request_verify(). This checks the signature on an
7 OCSP request and verifies the signer certificate. The signer
8 certificate is just checked for a generic purpose and OCSP request
12 *) Add OCSP_check_validity() function to check the validity of OCSP
13 responses. OCSP responses are prepared in real time and may only
14 be a few seconds old. Simply checking that the current time lies
15 between thisUpdate and nextUpdate max reject otherwise valid responses
16 caused by either OCSP responder or client clock innacuracy. Instead
17 we allow thisUpdate and nextUpdate to fall within a certain period of
18 the current time. The age of the response can also optionally be
19 checked. Two new options -validity_period and -status_age added to
23 *) If signature or public key algorithm is unrecognized print out its
24 OID rather that just UNKOWN.
27 *) Avoid coredump with unsupported or invalid public keys by checking if
28 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
29 PKCS7_verify() fails with non detached data.
32 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
33 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
34 ID to be generated from the issuer certificate alone which can then be
35 passed to OCSP_id_issuer_cmp().
38 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
39 ASN1 modules to export functions returning ASN1_ITEM pointers
40 instead of the ASN1_ITEM structures themselves. This adds several
41 new macros which allow the underlying ASN1 function/structure to
42 be accessed transparently. As a result code should not use ASN1_ITEM
43 references directly (such as &X509_it) but instead use the relevant
44 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
45 use of the new ASN1 code on platforms where exporting structures
46 is problematical (for example in shared libraries) but exporting
47 functions returning pointers to structures is not.
50 *) Add support for overriding the generation of SSL/TLS session IDs.
51 These callbacks can be registered either in an SSL_CTX or per SSL.
52 The purpose of this is to allow applications to control, if they wish,
53 the arbitrary values chosen for use as session IDs, particularly as it
54 can be useful for session caching in multiple-server environments. A
55 command-line switch for testing this (and any client code that wishes
56 to use such a feature) has been added to "s_server".
57 [Geoff Thorpe, Lutz Jaenicke]
59 *) Modify mkdef.pl to recognise and parse prprocessor conditionals
60 of the form '#if defined(...) || defined(...) || ...' and
61 '#if !defined(...) && !defined(...) && ...'. This also avoids
62 the growing number of special cases it was previously handling.
65 *) Don't use getenv in library functions when run as setuid/setgid.
66 New function OPENSSL_issetugid().
69 *) Make all configuration macros available for application by making
70 sure they are available in opensslconf.h, by giving them names starting
71 with "OPENSSL_" to avoid conflicts with other packages and by making
72 sure e_os2.h will cover all platform-specific cases together with
74 Additionally, it is now possible to define configuration/platform-
75 specific names (called "system identities"). In the C code, these
76 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
77 macro with the name beginning with "OPENSSL_SYS_", which is determined
78 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
82 *) New option -set_serial to 'req' and 'x509' this allows the serial
83 number to use to be specified on the command line. Previously self
84 signed certificates were hard coded with serial number 0 and the
85 CA options of 'x509' had to use a serial number in a file which was
89 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
90 due to incorrect handling of multi-threading:
92 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
94 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
96 3. Count how many times MemCheck_off() has been called so that
97 nested use can be treated correctly. This also avoids
98 inband-signalling in the previous code (which relied on the
99 assumption that thread ID 0 is impossible).
102 *) New options to 'ca' utility to support V2 CRL entry extensions.
103 Currently CRL reason, invalidity date and hold instruction are
104 supported. Add new CRL extensions to V3 code and some new objects.
107 *) Add "-rand" option also to s_client and s_server.
110 *) New function EVP_CIPHER_CTX_set_padding() this is used to
111 disable standard block padding (aka PKCS#5 padding) in the EVP
112 API, which was previously mandatory. This means that the data is
113 not padded in any way and so the total length much be a multiple
114 of the block size, otherwise an error occurs.
117 *) Initial (incomplete) OCSP SSL support.
120 *) Fix CPU detection on Irix 6.x.
121 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
122 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
124 *) New function OCSP_parse_url(). This splits up a URL into its host,
125 port and path components: primarily to parse OCSP URLs. New -url
126 option to ocsp utility.
129 *) New nonce behavior. The return value of OCSP_check_nonce() now
130 reflects the various checks performed. Applications can decide
131 whether to tolerate certain situations such as an absent nonce
132 in a response when one was present in a request: the ocsp application
133 just prints out a warning. New function OCSP_add1_basic_nonce()
134 this is to allow responders to include a nonce in a response even if
135 the request is nonce-less.
138 *) Use the cached encoding of an X509_NAME structure rather than
139 copying it. This is apparently the reason for the libsafe "errors"
140 but the code is actually correct.
143 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
144 skipped when using openssl x509 multiple times on a single input file,
145 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
148 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
149 set string type: to handle setting ASN1_TIME structures. Fix ca
150 utility to correctly initialize revocation date of CRLs.
153 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
154 the clients preferred ciphersuites and rather use its own preferences.
155 Should help to work around M$ SGC (Server Gated Cryptography) bug in
156 Internet Explorer by ensuring unchanged hash method during stepup.
159 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
160 to aes and add a new 'exist' option to print out symbols that don't
164 *) Additional options to ocsp utility to allow flags to be set and
165 additional certificates supplied.
168 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
169 OCSP client a number of certificate to only verify the response
173 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
174 Bleichenbacher's DSA attack.
175 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
176 to be set and top=0 forces the highest bit to be set; top=-1 is new
177 and leaves the highest bit random.
178 [Ulf Moeller, Bodo Moeller]
180 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
181 handle the new API. Currently only ECB, CBC modes supported. Add new
182 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
183 for TLS" draft-ietf-tls-ciphersuite-03.txt.
184 [Ben Laurie, Steve Henson]
186 *) In the NCONF_...-based implementations for CONF_... queries
187 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
188 a temporary CONF structure with the data component set to NULL
189 (which gives segmentation faults in lh_retrieve).
190 Instead, use NULL for the CONF pointer in CONF_get_string and
191 CONF_get_number (which may use environment variables) and directly
192 return NULL from CONF_get_section.
195 *) Fix potential buffer overrun for EBCDIC.
198 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
202 *) Functions for OCSP responders. OCSP_request_onereq_count(),
203 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
204 extract information from a certificate request. OCSP_response_create()
205 creates a response and optionally adds a basic response structure.
206 OCSP_basic_add1_status() adds a complete single response to a basic
207 reponse and returns the OCSP_SINGLERESP structure just added (to allow
208 extensions to be included for example). OCSP_basic_add1_cert() adds a
209 certificate to a basic response and OCSP_basic_sign() signs a basic
210 response with various flags. New helper functions ASN1_TIME_check()
211 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
212 (converts ASN1_TIME to GeneralizedTime).
215 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
216 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
217 structure from a certificate. X509_pubkey_digest() digests tha public_key
218 contents: this is used in various key identifiers.
221 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
222 keyUsage if basicConstraints absent for a CA.
225 *) Make SMIME_write_PKCS7() write mail header values with a format that
226 is more generally accepted (no spaces before the semicolon), since
227 some programs can't parse those values properly otherwise. Also make
228 sure BIO's that break lines after each write do not create invalid
232 *) Make sk_sort() tolerate a NULL argument.
233 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
235 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
236 passed by the function are trusted implicitly. If any of them signed the
237 reponse then it is assumed to be valid and is not verified.
240 *) Zero the premaster secret after deriving the master secret in
244 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
245 to data. This was previously part of the PKCS7 ASN1 code. This
246 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
247 [Steve Henson, reported by Kenneth R. Robinette
248 <support@securenetterm.com>]
250 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
251 routines: without these tracing memory leaks is very painful.
252 Fix leaks in PKCS12 and PKCS7 routines.
255 *) Fix for Irix with NO_ASM.
256 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
258 *) Add some EVP_add_digest_alias registrations (as found in
259 OpenSSL_add_all_digests()), to SSL_library_init()
260 aka OpenSSL_add_ssl_algorithms(). This provides improved
261 compatibility with peers using X.509 certificates
262 with unconventional AlgorithmIdentifier OIDs.
265 *) ./config script fixes.
266 [Ulf Moeller, Richard Levitte]
268 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
269 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
270 effectively meant GeneralizedTime would never be used. Now it
271 is initialised to -1 but X509_time_adj() now has to check the value
272 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
273 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
274 [Steve Henson, reported by Kenneth R. Robinette
275 <support@securenetterm.com>]
277 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
278 result in a zero length in the ASN1_INTEGER structure which was
279 not consistent with the structure when d2i_ASN1_INTEGER() was used
280 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
281 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
282 where it did not print out a minus for negative ASN1_INTEGER.
285 *) Fix 'openssl passwd -1'.
288 *) Add summary printout to ocsp utility. The various functions which
289 convert status values to strings have been renamed to:
290 OCSP_response_status_str(), OCSP_cert_status_str() and
291 OCSP_crl_reason_str() and are no longer static. New options
292 to verify nonce values and to disable verification. OCSP response
293 printout format cleaned up.
296 *) Add additional OCSP certificate checks. These are those specified
297 in RFC2560. This consists of two separate checks: the CA of the
298 certificate being checked must either be the OCSP signer certificate
299 or the issuer of the OCSP signer certificate. In the latter case the
300 OCSP signer certificate must contain the OCSP signing extended key
301 usage. This check is performed by attempting to match the OCSP
302 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
303 in the OCSP_CERTID structures of the response.
306 *) Initial OCSP certificate verification added to OCSP_basic_verify()
307 and related routines. This uses the standard OpenSSL certificate
308 verify routines to perform initial checks (just CA validity) and
309 to obtain the certificate chain. Then additional checks will be
310 performed on the chain. Currently the root CA is checked to see
311 if it is explicitly trusted for OCSP signing. This is used to set
312 a root CA as a global signing root: that is any certificate that
313 chains to that CA is an acceptable OCSP signing certificate.
316 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
317 extensions from a separate configuration file.
318 As when reading extensions from the main configuration file,
319 the '-extensions ...' option may be used for specifying the
321 [Massimiliano Pala <madwolf@comune.modena.it>]
323 *) Change PKCS12_key_gen_asc() so it can cope with non null
324 terminated strings whose length is passed in the passlen
325 parameter, for example from PEM callbacks. This was done
326 by adding an extra length parameter to asc2uni().
327 [Steve Henson, reported by <oddissey@samsung.co.kr>]
329 *) New OCSP utility. Allows OCSP requests to be generated or
330 read. The request can be sent to a responder and the output
331 parsed, outputed or printed in text form. Not complete yet:
332 still needs to check the OCSP response validity.
335 *) New subcommands for 'openssl ca':
336 'openssl ca -status <serial>' prints the status of the cert with
337 the given serial number (according to the index file).
338 'openssl ca -updatedb' updates the expiry status of certificates
340 [Massimiliano Pala <madwolf@comune.modena.it>]
342 *) New '-newreq-nodes' command option to CA.pl. This is like
343 '-newreq', but calls 'openssl req' with the '-nodes' option
344 so that the resulting key is not encrypted.
345 [Damien Miller <djm@mindrot.org>]
347 *) New configuration for the GNU Hurd.
348 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
350 *) Initial code to implement OCSP basic response verify. This
351 is currently incomplete. Currently just finds the signer's
352 certificate and verifies the signature on the response.
355 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
356 value of OPENSSLDIR. This is available via the new '-d' option
357 to 'openssl version', and is also included in 'openssl version -a'.
360 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
361 call failed, free the DSA structure.
364 *) Allowing defining memory allocation callbacks that will be given
365 file name and line number information in additional arguments
366 (a const char* and an int). The basic functionality remains, as
367 well as the original possibility to just replace malloc(),
368 realloc() and free() by functions that do not know about these
369 additional arguments. To register and find out the current
370 settings for extended allocation functions, the following
371 functions are provided:
373 CRYPTO_set_mem_ex_functions
374 CRYPTO_set_locked_mem_ex_functions
375 CRYPTO_get_mem_ex_functions
376 CRYPTO_get_locked_mem_ex_functions
378 These work the same way as CRYPTO_set_mem_functions and friends.
379 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
380 extended allocation function is enabled.
381 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
382 a conventional allocation function is enabled.
383 [Richard Levitte, Bodo Moeller]
385 *) Fix to uni2asc() to cope with zero length Unicode strings.
386 These are present in some PKCS#12 files.
389 *) Finish off removing the remaining LHASH function pointer casts.
390 There should no longer be any prototype-casting required when using
391 the LHASH abstraction, and any casts that remain are "bugs". See
392 the callback types and macros at the head of lhash.h for details
393 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
396 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
397 If an EGD or PRNGD is running and enough entropy is returned, automatic
398 seeding like with /dev/[u]random will be performed.
399 Positions tried are: /etc/entropy, /var/run/egd-pool.
402 *) Change the Unix RAND_poll() variant to be able to poll several
403 random devices, as specified by DEVRANDOM, until a sufficient amount
404 of data has been collected. We spend at most 10 ms on each file
405 (select timeout) and read in non-blocking mode. DEVRANDOM now
406 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
407 (previously it was just the string "/dev/urandom"), so on typical
408 platforms the 10 ms delay will never occur.
409 Also separate out the Unix variant to its own file, rand_unix.c.
410 For VMS, there's a currently-empty rand_vms.c.
413 *) Move OCSP client related routines to ocsp_cl.c. These
414 provide utility functions which an application needing
415 to issue a request to an OCSP responder and analyse the
416 response will typically need: as opposed to those which an
417 OCSP responder itself would need which will be added later.
419 OCSP_request_sign() signs an OCSP request with an API similar
420 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
421 response. OCSP_response_get1_basic() extracts basic response
422 from response. OCSP_resp_find_status(): finds and extracts status
423 information from an OCSP_CERTID structure (which will be created
424 when the request structure is built). These are built from lower
425 level functions which work on OCSP_SINGLERESP structures but
426 wont normally be used unless the application wishes to examine
427 extensions in the OCSP response for example.
429 Replace nonce routines with a pair of functions.
430 OCSP_request_add1_nonce() adds a nonce value and optionally
431 generates a random value. OCSP_check_nonce() checks the
432 validity of the nonce in an OCSP response.
435 *) Change function OCSP_request_add() to OCSP_request_add0_id().
436 This doesn't copy the supplied OCSP_CERTID and avoids the
437 need to free up the newly created id. Change return type
438 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
439 This can then be used to add extensions to the request.
440 Deleted OCSP_request_new(), since most of its functionality
441 is now in OCSP_REQUEST_new() (and the case insensitive name
442 clash) apart from the ability to set the request name which
443 will be added elsewhere.
446 *) Update OCSP API. Remove obsolete extensions argument from
447 various functions. Extensions are now handled using the new
448 OCSP extension code. New simple OCSP HTTP function which
449 can be used to send requests and parse the response.
452 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
453 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
454 uses the special reorder version of SET OF to sort the attributes
455 and reorder them to match the encoded order. This resolves a long
456 standing problem: a verify on a PKCS7 structure just after signing
457 it used to fail because the attribute order did not match the
458 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
459 it uses the received order. This is necessary to tolerate some broken
460 software that does not order SET OF. This is handled by encoding
461 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
462 to produce the required SET OF.
465 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
466 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
467 files to get correct declarations of the ASN.1 item variables.
470 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
471 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
472 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
473 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
474 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
475 ASN1_ITEM and no wrapper functions.
478 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
479 replace the old function pointer based I/O routines. Change most of
480 the *_d2i_bio() and *_d2i_fp() functions to use these.
483 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
484 lines, recognice more "algorithms" that can be deselected, and make
485 it complain about algorithm deselection that isn't recognised.
488 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
489 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
490 to use new functions. Add NO_ASN1_OLD which can be set to remove
491 some old style ASN1 functions: this can be used to determine if old
492 code will still work when these eventually go away.
495 *) New extension functions for OCSP structures, these follow the
496 same conventions as certificates and CRLs.
499 *) New function X509V3_add1_i2d(). This automatically encodes and
500 adds an extension. Its behaviour can be customised with various
501 flags to append, replace or delete. Various wrappers added for
502 certifcates and CRLs.
505 *) Fix to avoid calling the underlying ASN1 print routine when
506 an extension cannot be parsed. Correct a typo in the
507 OCSP_SERVICELOC extension. Tidy up print OCSP format.
510 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
511 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
512 when writing a 32767 byte record.
513 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
515 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
516 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
518 (RSA objects have a reference count access to which is protected
519 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
520 so they are meant to be shared between threads.)
521 [Bodo Moeller, Geoff Thorpe; original patch submitted by
522 "Reddie, Steven" <Steven.Reddie@ca.com>]
524 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
525 entries for variables.
528 *) Fix a deadlock in CRYPTO_mem_leaks().
531 *) Add functionality to apps/openssl.c for detecting locking
532 problems: As the program is single-threaded, all we have
533 to do is register a locking callback using an array for
534 storing which locks are currently held by the program.
537 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
538 SSL_get_ex_data_X509_STORE_idx(), which is used in
539 ssl_verify_cert_chain() and thus can be called at any time
540 during TLS/SSL handshakes so that thread-safety is essential.
541 Unfortunately, the ex_data design is not at all suited
542 for multi-threaded use, so it probably should be abolished.
545 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
546 [Broadcom, tweaked and integrated by Geoff Thorpe]
548 *) Move common extension printing code to new function
549 X509V3_print_extensions(). Reorganise OCSP print routines and
550 implement some needed OCSP ASN1 functions. Add OCSP extensions.
553 *) New function X509_signature_print() to remove duplication in some
557 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
558 set (this was treated exactly the same as SET OF previously). This
559 is used to reorder the STACK representing the structure to match the
560 encoding. This will be used to get round a problem where a PKCS7
561 structure which was signed could not be verified because the STACK
562 order did not reflect the encoded order.
565 *) Reimplement the OCSP ASN1 module using the new code.
568 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
569 for its ASN1 operations. The old style function pointers still exist
570 for now but they will eventually go away.
573 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
574 completely replaces the old ASN1 functionality with a table driven
575 encoder and decoder which interprets an ASN1_ITEM structure describing
576 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
577 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
578 has also been converted to the new form.
581 *) Change BN_mod_exp_recp so that negative moduli are tolerated
582 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
583 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
587 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
588 of not touching the result's sign bit.
591 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
595 *) Changed the LHASH code to use prototypes for callbacks, and created
596 macros to declare and implement thin (optionally static) functions
597 that provide type-safety and avoid function pointer casting for the
598 type-specific callbacks.
601 *) Use better test patterns in bntest.
604 *) Added Kerberos Cipher Suites to be used with TLS, as written in
606 [Veers Staats <staatsvr@asc.hpc.mil>,
607 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
609 *) rand_win.c fix for Borland C.
612 *) BN_rshift bugfix for n == 0.
615 *) Reformat the FAQ so the different questions and answers can be divided
616 in sections depending on the subject.
619 *) Have the zlib compression code load ZLIB.DLL dynamically under
623 *) New function BN_mod_sqrt for computing square roots modulo a prime
624 (using the probabilistic Tonelli-Shanks algorithm unless
625 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
626 be handled deterministically).
627 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
629 *) Store verify_result within SSL_SESSION also for client side to
630 avoid potential security hole. (Re-used sessions on the client side
631 always resulted in verify_result==X509_V_OK, not using the original
632 result of the server certificate verification.)
635 *) Make BN_mod_inverse faster by explicitly handling small quotients
636 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
637 512 bits], about 30% for larger ones [1024 or 2048 bits].)
640 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
641 SSL3_RT_APPLICATION_DATA, return 0.
642 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
646 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
647 releases, have been re-implemented by renaming the previous
648 implementations of ssl2_read and ssl3_read to ssl2_read_internal
649 and ssl3_read_internal, respectively, and adding 'peek' parameters
650 to them. The new ssl[23]_{read,peek} functions are calls to
651 ssl[23]_read_internal with the 'peek' flag set appropriately.
652 A 'peek' parameter has also been added to ssl3_read_bytes, which
653 does the actual work for ssl3_read_internal.
656 *) New function BN_kronecker.
659 *) Fix BN_gcd so that it works on negative inputs; the result is
660 positive unless both parameters are zero.
661 Previously something reasonably close to an infinite loop was
662 possible because numbers could be growing instead of shrinking
663 in the implementation of Euclid's algorithm.
666 *) Fix BN_is_word() and BN_is_one() macros to take into account the
667 sign of the number in question.
669 Fix BN_is_word(a,w) to work correctly for w == 0.
671 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
672 because its test if the absolute value of 'a' equals 'w'.
673 Note that BN_abs_is_word does *not* handle w == 0 reliably;
674 it exists mostly for use in the implementations of BN_is_zero(),
675 BN_is_one(), and BN_is_word().
678 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
679 the method-specific "init()" handler. Also clean up ex_data after
680 calling the method-specific "finish()" handler. Previously, this was
681 happening the other way round.
684 *) New function BN_swap.
687 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
688 the exponentiation functions are more likely to produce reasonable
689 results on negative inputs.
692 *) Change BN_mod_mul so that the result is always non-negative.
693 Previously, it could be negative if one of the factors was negative;
694 I don't think anyone really wanted that behaviour.
697 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
698 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
699 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
700 and add new functions:
713 These functions always generate non-negative results.
715 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
716 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
718 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
719 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
721 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
723 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
724 was actually never needed) and in BN_mul(). The removal in BN_mul()
725 required a small change in bn_mul_part_recursive() and the addition
726 of the functions bn_cmp_part_words(), bn_sub_part_words() and
727 bn_add_part_words(), which do the same thing as bn_cmp_words(),
728 bn_sub_words() and bn_add_words() except they take arrays with
732 *) In 'openssl passwd', verify passwords read from the terminal
733 unless the '-salt' option is used (which usually means that
734 verification would just waste user's time since the resulting
735 hash is going to be compared with some given password hash)
736 or the new '-noverify' option is used.
738 This is an incompatible change, but it does not affect
739 non-interactive use of 'openssl passwd' (passwords on the command
740 line, '-stdin' option, '-in ...' option) and thus should not
744 *) Remove all references to RSAref, since there's no more need for it.
747 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
748 The previous value, 12, was not always sufficient for BN_mod_exp().
751 *) Make DSO load along a path given through an environment variable
752 (SHLIB_PATH) with shl_load().
755 *) Constify the ENGINE code as a result of BIGNUM constification.
756 Also constify the RSA code and most things related to it. In a
757 few places, most notable in the depth of the ASN.1 code, ugly
758 casts back to non-const were required (to be solved at a later
762 *) Make it so the openssl application has all engines loaded by default.
765 *) Constify the BIGNUM routines a little more.
768 *) Make sure that shared libraries get the internal name engine with
769 the full version number and not just 0. This should mark the
770 shared libraries as not backward compatible. Of course, this should
771 be changed again when we can guarantee backward binary compatibility.
774 *) Add the following functions:
780 ENGINE_load_builtin_engines()
782 That way, an application can itself choose if external engines that
783 are built-in in OpenSSL shall ever be used or not. The benefit is
784 that applications won't have to be linked with libdl or other dso
785 libraries unless it's really needed.
787 Changed 'openssl engine' to load all engines on demand.
788 Changed the engine header files to avoid the duplication of some
789 declarations (they differed!).
792 *) 'openssl engine' can now list capabilities.
795 *) Better error reporting in 'openssl engine'.
798 *) Never call load_dh_param(NULL) in s_server.
801 *) Add engine application. It can currently list engines by name and
802 identity, and test if they are actually available.
805 *) Add support for shared libraries under Irix.
806 [Albert Chin-A-Young <china@thewrittenword.com>]
808 *) Improve RPM specification file by forcing symbolic linking and making
809 sure the installed documentation is also owned by root.root.
810 [Damien Miller <djm@mindrot.org>]
812 *) Add configuration option to build on Linux on both big-endian and
814 [Ralf Baechle <ralf@uni-koblenz.de>]
816 *) Give the OpenSSL applications more possibilities to make use of
817 keys (public as well as private) handled by engines.
820 *) Add OCSP code that comes from CertCo.
823 *) Add VMS support for the Rijndael code.
826 *) Added untested support for Nuron crypto accelerator.
829 *) Add support for external cryptographic devices. This code was
830 previously distributed separately as the "engine" branch.
831 [Geoff Thorpe, Richard Levitte]
833 *) Rework the filename-translation in the DSO code. It is now possible to
834 have far greater control over how a "name" is turned into a filename
835 depending on the operating environment and any oddities about the
836 different shared library filenames on each system.
839 *) Support threads on FreeBSD-elf in Configure.
842 *) Add the possibility to create shared libraries on HP-UX
845 *) Fix for SHA1 assembly problem with MASM: it produces
846 warnings about corrupt line number information when assembling
847 with debugging information. This is caused by the overlapping
849 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
852 NCONF_get_number() has no error checking at all. As a replacement,
853 NCONF_get_number_e() is defined (_e for "error checking") and is
854 promoted strongly. The old NCONF_get_number is kept around for
855 binary backward compatibility.
856 Make it possible for methods to load from something other than a BIO,
857 by providing a function pointer that is given a name instead of a BIO.
858 For example, this could be used to load configuration data from an
862 *) Fix typo in get_cert_by_subject() in by_dir.c
863 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
865 *) Rework the system to generate shared libraries:
867 - Make note of the expected extension for the shared libraries and
868 if there is a need for symbolic links from for example libcrypto.so.0
869 to libcrypto.so.0.9.7. There is extended info in Configure for
872 - Make as few rebuilds of the shared libraries as possible.
874 - Still avoid linking the OpenSSL programs with the shared libraries.
876 - When installing, install the shared libraries separately from the
880 *) Fix for non blocking accept BIOs. Added new I/O special reason
881 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
882 with non blocking I/O was not possible because no retry code was
883 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
887 *) Added the beginnings of Rijndael support.
890 *) Fix for bug in DirectoryString mask setting. Add support for
891 X509_NAME_print_ex() in 'req' and X509_print_ex() function
892 to allow certificate printing to more controllable, additional
893 'certopt' option to 'x509' to allow new printing options to be
897 *) Clean old EAY MD5 hack from e_os.h.
900 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
902 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
903 and not in SSL_clear because the latter is also used by the
904 accept/connect functions; previously, the settings made by
905 SSL_set_read_ahead would be lost during the handshake.
906 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
908 *) Correct util/mkdef.pl to be selective about disabled algorithms.
909 Previously, it would create entries for disableed algorithms no
913 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
915 *) In ssl23_get_client_hello, generate an error message when faced
916 with an initial SSL 3.0/TLS record that is too small to contain the
917 first two bytes of the ClientHello message, i.e. client_version.
918 (Note that this is a pathologic case that probably has never happened
919 in real life.) The previous approach was to use the version number
920 from the record header as a substitute; but our protocol choice
921 should not depend on that one because it is not authenticated
922 by the Finished messages.
925 *) More robust randomness gathering functions for Windows.
926 [Jeffrey Altman <jaltman@columbia.edu>]
928 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
929 not set then we don't setup the error code for issuer check errors
930 to avoid possibly overwriting other errors which the callback does
931 handle. If an application does set the flag then we assume it knows
932 what it is doing and can handle the new informational codes
936 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
937 a general "ANY" type, as such it should be able to decode anything
938 including tagged types. However it didn't check the class so it would
939 wrongly interpret tagged types in the same way as their universal
940 counterpart and unknown types were just rejected. Changed so that the
941 tagged and unknown types are handled in the same way as a SEQUENCE:
942 that is the encoding is stored intact. There is also a new type
943 "V_ASN1_OTHER" which is used when the class is not universal, in this
944 case we have no idea what the actual type is so we just lump them all
948 *) On VMS, stdout may very well lead to a file that is written to
949 in a record-oriented fashion. That means that every write() will
950 write a separate record, which will be read separately by the
951 programs trying to read from it. This can be very confusing.
953 The solution is to put a BIO filter in the way that will buffer
954 text until a linefeed is reached, and then write everything a
955 line at a time, so every record written will be an actual line,
956 not chunks of lines and not (usually doesn't happen, but I've
957 seen it once) several lines in one record. BIO_f_linebuffer() is
960 Currently, it's a VMS-only method, because that's where it has
961 been tested well enough.
964 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
965 it can return incorrect results.
966 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
967 but it was in 0.9.6-beta[12].)
970 *) Disable the check for content being present when verifying detached
971 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
972 include zero length content when signing messages.
975 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
976 BIO_ctrl (for BIO pairs).
979 *) Add DSO method for VMS.
982 *) Bug fix: Montgomery multiplication could produce results with the
986 *) Add RPM specification openssl.spec and modify it to build three
987 packages. The default package contains applications, application
988 documentation and run-time libraries. The devel package contains
989 include files, static libraries and function documentation. The
990 doc package contains the contents of the doc directory. The original
991 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
994 *) Add a large number of documentation files for many SSL routines.
995 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
997 *) Add a configuration entry for Sony News 4.
998 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
1000 *) Don't set the two most significant bits to one when generating a
1001 random number < q in the DSA library.
1004 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
1005 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
1006 the underlying transport is blocking) if a handshake took place.
1007 (The default behaviour is needed by applications such as s_client
1008 and s_server that use select() to determine when to use SSL_read;
1009 but for applications that know in advance when to expect data, it
1010 just makes things more complicated.)
1013 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
1017 *) Add a few more EBCDIC conditionals that make `req' and `x509'
1018 work better on such systems.
1019 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1021 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
1022 Update PKCS12_parse() so it copies the friendlyName and the
1023 keyid to the certificates aux info.
1026 *) Fix bug in PKCS7_verify() which caused an infinite loop
1027 if there was more than one signature.
1028 [Sven Uszpelkat <su@celocom.de>]
1030 *) Major change in util/mkdef.pl to include extra information
1031 about each symbol, as well as presentig variables as well
1032 as functions. This change means that there's n more need
1033 to rebuild the .num files when some algorithms are excluded.
1036 *) Allow the verify time to be set by an application,
1037 rather than always using the current time.
1040 *) Phase 2 verify code reorganisation. The certificate
1041 verify code now looks up an issuer certificate by a
1042 number of criteria: subject name, authority key id
1043 and key usage. It also verifies self signed certificates
1044 by the same criteria. The main comparison function is
1045 X509_check_issued() which performs these checks.
1047 Lot of changes were necessary in order to support this
1048 without completely rewriting the lookup code.
1050 Authority and subject key identifier are now cached.
1052 The LHASH 'certs' is X509_STORE has now been replaced
1053 by a STACK_OF(X509_OBJECT). This is mainly because an
1054 LHASH can't store or retrieve multiple objects with
1055 the same hash value.
1057 As a result various functions (which were all internal
1058 use only) have changed to handle the new X509_STORE
1059 structure. This will break anything that messed round
1060 with X509_STORE internally.
1062 The functions X509_STORE_add_cert() now checks for an
1063 exact match, rather than just subject name.
1065 The X509_STORE API doesn't directly support the retrieval
1066 of multiple certificates matching a given criteria, however
1067 this can be worked round by performing a lookup first
1068 (which will fill the cache with candidate certificates)
1069 and then examining the cache for matches. This is probably
1070 the best we can do without throwing out X509_LOOKUP
1071 entirely (maybe later...).
1073 The X509_VERIFY_CTX structure has been enhanced considerably.
1075 All certificate lookup operations now go via a get_issuer()
1076 callback. Although this currently uses an X509_STORE it
1077 can be replaced by custom lookups. This is a simple way
1078 to bypass the X509_STORE hackery necessary to make this
1079 work and makes it possible to use more efficient techniques
1080 in future. A very simple version which uses a simple
1081 STACK for its trusted certificate store is also provided
1082 using X509_STORE_CTX_trusted_stack().
1084 The verify_cb() and verify() callbacks now have equivalents
1085 in the X509_STORE_CTX structure.
1087 X509_STORE_CTX also has a 'flags' field which can be used
1088 to customise the verify behaviour.
1091 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1092 excludes S/MIME capabilities.
1095 *) When a certificate request is read in keep a copy of the
1096 original encoding of the signed data and use it when outputing
1097 again. Signatures then use the original encoding rather than
1098 a decoded, encoded version which may cause problems if the
1099 request is improperly encoded.
1102 *) For consistency with other BIO_puts implementations, call
1103 buffer_write(b, ...) directly in buffer_puts instead of calling
1106 In BIO_puts, increment b->num_write as in BIO_write.
1107 [Peter.Sylvester@EdelWeb.fr]
1109 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1110 BN_zero, we may not return a BIGNUM with an array consisting of
1114 *) Avoid calling abort() from within the library when problems are
1115 detected, except if preprocessor symbols have been defined
1116 (such as REF_CHECK, BN_DEBUG etc.).
1119 *) New openssl application 'rsautl'. This utility can be
1120 used for low level RSA operations. DER public key
1121 BIO/fp routines also added.
1124 *) New Configure entry and patches for compiling on QNX 4.
1125 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1127 *) A demo state-machine implementation was sponsored by
1128 Nuron (http://www.nuron.com/) and is now available in
1129 demos/state_machine.
1132 *) New options added to the 'dgst' utility for signature
1133 generation and verification.
1136 *) Unrecognized PKCS#7 content types are now handled via a
1137 catch all ASN1_TYPE structure. This allows unsupported
1138 types to be stored as a "blob" and an application can
1139 encode and decode it manually.
1142 *) Fix various signed/unsigned issues to make a_strex.c
1144 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1146 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1147 length if passed a buffer. ASN1_INTEGER_to_BN failed
1148 if passed a NULL BN and its argument was negative.
1149 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1151 *) Modification to PKCS#7 encoding routines to output definite
1152 length encoding. Since currently the whole structures are in
1153 memory there's not real point in using indefinite length
1154 constructed encoding. However if OpenSSL is compiled with
1155 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1158 *) Added BIO_vprintf() and BIO_vsnprintf().
1161 *) Added more prefixes to parse for in the the strings written
1162 through a logging bio, to cover all the levels that are available
1163 through syslog. The prefixes are now:
1165 PANIC, EMERG, EMR => LOG_EMERG
1166 ALERT, ALR => LOG_ALERT
1167 CRIT, CRI => LOG_CRIT
1168 ERROR, ERR => LOG_ERR
1169 WARNING, WARN, WAR => LOG_WARNING
1170 NOTICE, NOTE, NOT => LOG_NOTICE
1171 INFO, INF => LOG_INFO
1172 DEBUG, DBG => LOG_DEBUG
1174 and as before, if none of those prefixes are present at the
1175 beginning of the string, LOG_ERR is chosen.
1177 On Win32, the LOG_* levels are mapped according to this:
1179 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1180 LOG_WARNING => EVENTLOG_WARNING_TYPE
1181 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1185 *) Made it possible to reconfigure with just the configuration
1186 argument "reconf" or "reconfigure". The command line arguments
1187 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1188 and are retrieved from there when reconfiguring.
1192 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1194 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1197 *) The obj_dat.pl script was messing up the sorting of object
1198 names. The reason was that it compared the quoted version
1199 of strings as a result "OCSP" > "OCSP Signing" because
1200 " > SPACE. Changed script to store unquoted versions of
1201 names and add quotes on output. It was also omitting some
1202 names from the lookup table if they were given a default
1203 value (that is if SN is missing it is given the same
1204 value as LN and vice versa), these are now added on the
1205 grounds that if an object has a name we should be able to
1206 look it up. Finally added warning output when duplicate
1207 short or long names are found.
1210 *) Changes needed for Tandem NSK.
1211 [Scott Uroff <scott@xypro.com>]
1213 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1214 RSA_padding_check_SSLv23(), special padding was never detected
1215 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1216 version rollback attacks was not effective.
1218 In s23_clnt.c, don't use special rollback-attack detection padding
1219 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1220 client; similarly, in s23_srvr.c, don't do the rollback check if
1221 SSL 2.0 is the only protocol enabled in the server.
1224 *) Make it possible to get hexdumps of unprintable data with 'openssl
1225 asn1parse'. By implication, the functions ASN1_parse_dump() and
1226 BIO_dump_indent() are added.
1229 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1230 these print out strings and name structures based on various
1231 flags including RFC2253 support and proper handling of
1232 multibyte characters. Added options to the 'x509' utility
1233 to allow the various flags to be set.
1236 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1237 Also change the functions X509_cmp_current_time() and
1238 X509_gmtime_adj() work with an ASN1_TIME structure,
1239 this will enable certificates using GeneralizedTime in validity
1240 dates to be checked.
1243 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1244 negative public key encodings) on by default,
1245 NO_NEG_PUBKEY_BUG can be set to disable it.
1248 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1249 content octets. An i2c_ASN1_OBJECT is unnecessary because
1250 the encoding can be trivially obtained from the structure.
1253 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1254 not read locks (CRYPTO_r_[un]lock).
1257 *) A first attempt at creating official support for shared
1258 libraries through configuration. I've kept it so the
1259 default is static libraries only, and the OpenSSL programs
1260 are always statically linked for now, but there are
1261 preparations for dynamic linking in place.
1262 This has been tested on Linux and True64.
1265 *) Randomness polling function for Win9x, as described in:
1266 Peter Gutmann, Software Generation of Practically Strong
1270 *) Fix so PRNG is seeded in req if using an already existing
1274 *) New options to smime application. -inform and -outform
1275 allow alternative formats for the S/MIME message including
1276 PEM and DER. The -content option allows the content to be
1277 specified separately. This should allow things like Netscape
1278 form signing output easier to verify.
1281 *) Fix the ASN1 encoding of tags using the 'long form'.
1284 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1285 STRING types. These convert content octets to and from the
1286 underlying type. The actual tag and length octets are
1287 already assumed to have been read in and checked. These
1288 are needed because all other string types have virtually
1289 identical handling apart from the tag. By having versions
1290 of the ASN1 functions that just operate on content octets
1291 IMPLICIT tagging can be handled properly. It also allows
1292 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1293 and ASN1_INTEGER are identical apart from the tag.
1296 *) Change the handling of OID objects as follows:
1298 - New object identifiers are inserted in objects.txt, following
1299 the syntax given in objects.README.
1300 - objects.pl is used to process obj_mac.num and create a new
1302 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1305 This is currently kind of a hack, and the perl code in objects.pl
1306 isn't very elegant, but it works as I intended. The simplest way
1307 to check that it worked correctly is to look in obj_dat.h and
1308 check the array nid_objs and make sure the objects haven't moved
1309 around (this is important!). Additions are OK, as well as
1310 consistent name changes.
1313 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1316 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1317 The given file adds to whatever has already been seeded into the
1318 random pool through the RANDFILE configuration file option or
1319 environment variable, or the default random state file.
1322 *) mkstack.pl now sorts each macro group into lexical order.
1323 Previously the output order depended on the order the files
1324 appeared in the directory, resulting in needless rewriting
1328 *) Patches to make OpenSSL compile under Win32 again. Mostly
1329 work arounds for the VC++ problem that it treats func() as
1330 func(void). Also stripped out the parts of mkdef.pl that
1331 added extra typesafe functions: these no longer exist.
1334 *) Reorganisation of the stack code. The macros are now all
1335 collected in safestack.h . Each macro is defined in terms of
1336 a "stack macro" of the form SKM_<name>(type, a, b). The
1337 DEBUG_SAFESTACK is now handled in terms of function casts,
1338 this has the advantage of retaining type safety without the
1339 use of additional functions. If DEBUG_SAFESTACK is not defined
1340 then the non typesafe macros are used instead. Also modified the
1341 mkstack.pl script to handle the new form. Needs testing to see
1342 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1343 the default if no major problems. Similar behaviour for ASN1_SET_OF
1344 and PKCS12_STACK_OF.
1347 *) When some versions of IIS use the 'NET' form of private key the
1348 key derivation algorithm is different. Normally MD5(password) is
1349 used as a 128 bit RC4 key. In the modified case
1350 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1351 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1352 as the old Netscape_RSA functions except they have an additional
1353 'sgckey' parameter which uses the modified algorithm. Also added
1354 an -sgckey command line option to the rsa utility. Thanks to
1355 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1356 algorithm to openssl-dev.
1359 *) The evp_local.h macros were using 'c.##kname' which resulted in
1360 invalid expansion on some systems (SCO 5.0.5 for example).
1361 Corrected to 'c.kname'.
1362 [Phillip Porch <root@theporch.com>]
1364 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1365 a STACK of email addresses from a certificate or request, these look
1366 in the subject name and the subject alternative name extensions and
1367 omit any duplicate addresses.
1370 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1371 This makes DSA verification about 2 % faster.
1374 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1375 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1376 plus overhead for 1024 bit moduli).
1377 This makes exponentiations about 0.5 % faster for 1024 bit
1378 exponents (as measured by "openssl speed rsa2048").
1381 *) Rename memory handling macros to avoid conflicts with other
1383 Malloc => OPENSSL_malloc
1384 Malloc_locked => OPENSSL_malloc_locked
1385 Realloc => OPENSSL_realloc
1386 Free => OPENSSL_free
1389 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1390 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1393 *) CygWin32 support.
1394 [John Jarvie <jjarvie@newsguy.com>]
1396 *) The type-safe stack code has been rejigged. It is now only compiled
1397 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1398 by default all type-specific stack functions are "#define"d back to
1399 standard stack functions. This results in more streamlined output
1400 but retains the type-safety checking possibilities of the original
1404 *) The STACK code has been cleaned up, and certain type declarations
1405 that didn't make a lot of sense have been brought in line. This has
1406 also involved a cleanup of sorts in safestack.h to more correctly
1407 map type-safe stack functions onto their plain stack counterparts.
1408 This work has also resulted in a variety of "const"ifications of
1409 lots of the code, especially "_cmp" operations which should normally
1410 be prototyped with "const" parameters anyway.
1413 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1414 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1415 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1416 where all of 'md' is used each time the PRNG is used, but 'state'
1417 is used only indexed by a cyclic counter. As entropy may not be
1418 well distributed from the beginning, 'md' is important as a
1419 chaining variable. However, the output function chains only half
1420 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1421 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1422 in all of 'state' being rewritten, with the new values depending
1423 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1426 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1427 the handshake is continued after ssl_verify_cert_chain();
1428 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1429 can lead to 'unexplainable' connection aborts later.
1430 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1432 *) Major EVP API cipher revision.
1433 Add hooks for extra EVP features. This allows various cipher
1434 parameters to be set in the EVP interface. Support added for variable
1435 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1436 setting of RC2 and RC5 parameters.
1438 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1441 Remove lots of duplicated code from the EVP library. For example *every*
1442 cipher init() function handles the 'iv' in the same way according to the
1443 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1444 for CFB and OFB modes they zero ctx->num.
1446 New functionality allows removal of S/MIME code RC2 hack.
1448 Most of the routines have the same form and so can be declared in terms
1451 By shifting this to the top level EVP_CipherInit() it can be removed from
1452 all individual ciphers. If the cipher wants to handle IVs or keys
1453 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1456 Change lots of functions like EVP_EncryptUpdate() to now return a
1457 value: although software versions of the algorithms cannot fail
1458 any installed hardware versions can.
1461 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1462 this option is set, tolerate broken clients that send the negotiated
1463 protocol version number instead of the requested protocol version
1467 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1468 i.e. non-zero for export ciphersuites, zero otherwise.
1469 Previous versions had this flag inverted, inconsistent with
1470 rsa_tmp_cb (..._TMP_RSA_CB).
1471 [Bodo Moeller; problem reported by Amit Chopra]
1473 *) Add missing DSA library text string. Work around for some IIS
1474 key files with invalid SEQUENCE encoding.
1477 *) Add a document (doc/standards.txt) that list all kinds of standards
1478 and so on that are implemented in OpenSSL.
1481 *) Enhance c_rehash script. Old version would mishandle certificates
1482 with the same subject name hash and wouldn't handle CRLs at all.
1483 Added -fingerprint option to crl utility, to support new c_rehash
1487 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1490 *) Fix for SSL server purpose checking. Server checking was
1491 rejecting certificates which had extended key usage present
1492 but no ssl client purpose.
1493 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1495 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1496 is a little unclear about how a blank password is handled.
1497 Since the password in encoded as a BMPString with terminating
1498 double NULL a zero length password would end up as just the
1499 double NULL. However no password at all is different and is
1500 handled differently in the PKCS#12 key generation code. NS
1501 treats a blank password as zero length. MSIE treats it as no
1502 password on export: but it will try both on import. We now do
1503 the same: PKCS12_parse() tries zero length and no password if
1504 the password is set to "" or NULL (NULL is now a valid password:
1505 it wasn't before) as does the pkcs12 application.
1508 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1509 perror when PEM_read_bio_X509_REQ fails, the error message must
1510 be obtained from the error queue.
1513 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1514 it in ERR_remove_state if appropriate, and change ERR_get_state
1515 accordingly to avoid race conditions (this is necessary because
1516 thread_hash is no longer constant once set).
1519 *) Bugfix for linux-elf makefile.one.
1522 *) RSA_get_default_method() will now cause a default
1523 RSA_METHOD to be chosen if one doesn't exist already.
1524 Previously this was only set during a call to RSA_new()
1525 or RSA_new_method(NULL) meaning it was possible for
1526 RSA_get_default_method() to return NULL.
1529 *) Added native name translation to the existing DSO code
1530 that will convert (if the flag to do so is set) filenames
1531 that are sufficiently small and have no path information
1532 into a canonical native form. Eg. "blah" converted to
1533 "libblah.so" or "blah.dll" etc.
1536 *) New function ERR_error_string_n(e, buf, len) which is like
1537 ERR_error_string(e, buf), but writes at most 'len' bytes
1538 including the 0 terminator. For ERR_error_string_n, 'buf'
1540 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1542 *) CONF library reworked to become more general. A new CONF
1543 configuration file reader "class" is implemented as well as a
1544 new functions (NCONF_*, for "New CONF") to handle it. The now
1545 old CONF_* functions are still there, but are reimplemented to
1546 work in terms of the new functions. Also, a set of functions
1547 to handle the internal storage of the configuration data is
1548 provided to make it easier to write new configuration file
1549 reader "classes" (I can definitely see something reading a
1550 configuration file in XML format, for example), called _CONF_*,
1551 or "the configuration storage API"...
1553 The new configuration file reading functions are:
1555 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1556 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1558 NCONF_default, NCONF_WIN32
1560 NCONF_dump_fp, NCONF_dump_bio
1562 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1563 NCONF_new creates a new CONF object. This works in the same way
1564 as other interfaces in OpenSSL, like the BIO interface.
1565 NCONF_dump_* dump the internal storage of the configuration file,
1566 which is useful for debugging. All other functions take the same
1567 arguments as the old CONF_* functions wth the exception of the
1568 first that must be a `CONF *' instead of a `LHASH *'.
1570 To make it easer to use the new classes with the old CONF_* functions,
1571 the function CONF_set_default_method is provided.
1574 *) Add '-tls1' option to 'openssl ciphers', which was already
1575 mentioned in the documentation but had not been implemented.
1576 (This option is not yet really useful because even the additional
1577 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1580 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1581 OpenSSL-based applications) load shared libraries and bind to
1582 them in a portable way.
1583 [Geoff Thorpe, with contributions from Richard Levitte]
1585 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1587 *) Make sure _lrotl and _lrotr are only used with MSVC.
1589 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1590 (the default implementation of RAND_status).
1592 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1593 to '-clrext' (= clear extensions), as intended and documented.
1594 [Bodo Moeller; inconsistency pointed out by Michael Attili
1597 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1598 was larger than the MD block size.
1599 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1601 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1602 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1603 using the passed key: if the passed key was a private key the result
1604 of X509_print(), for example, would be to print out all the private key
1608 *) des_quad_cksum() byte order bug fix.
1609 [Ulf Möller, using the problem description in krb4-0.9.7, where
1610 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1612 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1614 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1616 *) For easily testing in shell scripts whether some command
1617 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1618 returns with exit code 0 iff no command of the given name is available.
1619 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1620 the output goes to stdout and nothing is printed to stderr.
1621 Additional arguments are always ignored.
1623 Since for each cipher there is a command of the same name,
1624 the 'no-cipher' compilation switches can be tested this way.
1626 ('openssl no-XXX' is not able to detect pseudo-commands such
1627 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1630 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1633 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1634 is set; it will be thrown away anyway because each handshake creates
1636 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1637 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1638 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1639 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1642 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1643 'Q' and 'R' lose their special meanings (quit/renegotiate).
1644 This is part of what -quiet does; unlike -quiet, -ign_eof
1645 does not suppress any output.
1648 *) Add compatibility options to the purpose and trust code. The
1649 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1650 accepts a certificate or CA, this was the previous behaviour,
1651 with all the associated security issues.
1653 X509_TRUST_COMPAT is the old trust behaviour: only and
1654 automatically trust self signed roots in certificate store. A
1655 new trust setting X509_TRUST_DEFAULT is used to specify that
1656 a purpose has no associated trust setting and it should instead
1657 use the value in the default purpose.
1660 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1661 and fix a memory leak.
1664 *) In util/mkerr.pl (which implements 'make errors'), preserve
1665 reason strings from the previous version of the .c file, as
1666 the default to have only downcase letters (and digits) in
1667 automatically generated reasons codes is not always appropriate.
1670 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1671 using strerror. Previously, ERR_reason_error_string() returned
1672 library names as reason strings for SYSerr; but SYSerr is a special
1673 case where small numbers are errno values, not library numbers.
1676 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1677 converts DSA parameters into DH parameters. (When creating parameters,
1678 DSA_generate_parameters is used.)
1681 *) Include 'length' (recommended exponent length) in C code generated
1682 by 'openssl dhparam -C'.
1685 *) The second argument to set_label in perlasm was already being used
1686 so couldn't be used as a "file scope" flag. Moved to third argument
1690 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1691 instead of RAND_bytes for encryption IVs and salts.
1694 *) Include RAND_status() into RAND_METHOD instead of implementing
1695 it only for md_rand.c Otherwise replacing the PRNG by calling
1696 RAND_set_rand_method would be impossible.
1699 *) Don't let DSA_generate_key() enter an infinite loop if the random
1700 number generation fails.
1703 *) New 'rand' application for creating pseudo-random output.
1706 *) Added configuration support for Linux/IA64
1707 [Rolf Haberrecker <rolf@suse.de>]
1709 *) Assembler module support for Mingw32.
1712 *) Shared library support for HPUX (in shlib/).
1713 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1715 *) Shared library support for Solaris gcc.
1716 [Lutz Behnke <behnke@trustcenter.de>]
1718 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1720 *) PKCS7_encrypt() was adding text MIME headers twice because they
1721 were added manually and by SMIME_crlf_copy().
1724 *) In bntest.c don't call BN_rand with zero bits argument.
1725 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1727 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1728 case was implemented. This caused BN_div_recp() to fail occasionally.
1731 *) Add an optional second argument to the set_label() in the perl
1732 assembly language builder. If this argument exists and is set
1733 to 1 it signals that the assembler should use a symbol whose
1734 scope is the entire file, not just the current function. This
1735 is needed with MASM which uses the format label:: for this scope.
1736 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1738 *) Change the ASN1 types so they are typedefs by default. Before
1739 almost all types were #define'd to ASN1_STRING which was causing
1740 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1744 *) Change names of new functions to the new get1/get0 naming
1745 convention: After 'get1', the caller owns a reference count
1746 and has to call ..._free; 'get0' returns a pointer to some
1747 data structure without incrementing reference counters.
1748 (Some of the existing 'get' functions increment a reference
1749 counter, some don't.)
1750 Similarly, 'set1' and 'add1' functions increase reference
1751 counters or duplicate objects.
1754 *) Allow for the possibility of temp RSA key generation failure:
1755 the code used to assume it always worked and crashed on failure.
1758 *) Fix potential buffer overrun problem in BIO_printf().
1759 [Ulf Möller, using public domain code by Patrick Powell; problem
1760 pointed out by David Sacerdote <das33@cornell.edu>]
1762 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1763 RAND_egd() and RAND_status(). In the command line application,
1764 the EGD socket can be specified like a seed file using RANDFILE
1768 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1769 Some CAs (e.g. Verisign) distribute certificates in this form.
1772 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1773 list to exclude them. This means that no special compilation option
1774 is needed to use anonymous DH: it just needs to be included in the
1778 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1779 EVP_MD_type. The old functionality is available in a new macro called
1780 EVP_MD_md(). Change code that uses it and update docs.
1783 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1784 where the 'void *' argument is replaced by a function pointer argument.
1785 Previously 'void *' was abused to point to functions, which works on
1786 many platforms, but is not correct. As these functions are usually
1787 called by macros defined in OpenSSL header files, most source code
1788 should work without changes.
1791 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1792 sections with information on -D... compiler switches used for
1793 compiling the library so that applications can see them. To enable
1794 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1795 must be defined. E.g.,
1796 #define OPENSSL_ALGORITHM_DEFINES
1797 #include <openssl/opensslconf.h>
1798 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1799 [Richard Levitte, Ulf and Bodo Möller]
1801 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1805 *) Change the 'other' type in certificate aux info to a STACK_OF
1806 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1807 the required ASN1 format: arbitrary types determined by an OID.
1810 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1811 argument to 'req'. This is not because the function is newer or
1812 better than others it just uses the work 'NEW' in the certificate
1813 request header lines. Some software needs this.
1816 *) Reorganise password command line arguments: now passwords can be
1817 obtained from various sources. Delete the PEM_cb function and make
1818 it the default behaviour: i.e. if the callback is NULL and the
1819 usrdata argument is not NULL interpret it as a null terminated pass
1820 phrase. If usrdata and the callback are NULL then the pass phrase
1821 is prompted for as usual.
1824 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1825 the support is automatically enabled. The resulting binaries will
1826 autodetect the card and use it if present.
1827 [Ben Laurie and Compaq Inc.]
1829 *) Work around for Netscape hang bug. This sends certificate request
1830 and server done in one record. Since this is perfectly legal in the
1831 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1832 the bugs/SSLv3 entry for more info.
1835 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1838 *) Add -rand argument to smime and pkcs12 applications and read/write
1842 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1845 *) Add command line password options to the remaining applications.
1848 *) Bug fix for BN_div_recp() for numerators with an even number of
1852 *) More tests in bntest.c, and changed test_bn output.
1855 *) ./config recognizes MacOS X now.
1858 *) Bug fix for BN_div() when the first words of num and divsor are
1859 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1862 *) Add support for various broken PKCS#8 formats, and command line
1863 options to produce them.
1866 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1867 get temporary BIGNUMs from a BN_CTX.
1870 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1874 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1875 include a #define from the old name to the new. The original intent
1876 was that statically linked binaries could for example just call
1877 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1878 link with digests. This never worked becayse SSLeay_add_all_digests()
1879 and SSLeay_add_all_ciphers() were in the same source file so calling
1880 one would link with the other. They are now in separate source files.
1883 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1886 *) Use a less unusual form of the Miller-Rabin primality test (it used
1887 a binary algorithm for exponentiation integrated into the Miller-Rabin
1888 loop, our standard modexp algorithms are faster).
1891 *) Support for the EBCDIC character set completed.
1892 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1894 *) Source code cleanups: use const where appropriate, eliminate casts,
1895 use void * instead of char * in lhash.
1898 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1899 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1900 this the server could overwrite ephemeral keys that the client
1904 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1905 using 50 iterations of the Rabin-Miller test.
1907 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1908 iterations of the Rabin-Miller test as required by the appendix
1909 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1910 As BN_is_prime_fasttest includes trial division, DSA parameter
1911 generation becomes much faster.
1913 This implies a change for the callback functions in DSA_is_prime
1914 and DSA_generate_parameters: The callback function is called once
1915 for each positive witness in the Rabin-Miller test, not just
1916 occasionally in the inner loop; and the parameters to the
1917 callback function now provide an iteration count for the outer
1918 loop rather than for the current invocation of the inner loop.
1919 DSA_generate_parameters additionally can call the callback
1920 function with an 'iteration count' of -1, meaning that a
1921 candidate has passed the trial division test (when q is generated
1922 from an application-provided seed, trial division is skipped).
1925 *) New function BN_is_prime_fasttest that optionally does trial
1926 division before starting the Rabin-Miller test and has
1927 an additional BN_CTX * argument (whereas BN_is_prime always
1928 has to allocate at least one BN_CTX).
1929 'callback(1, -1, cb_arg)' is called when a number has passed the
1930 trial division stage.
1933 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1937 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1940 *) New function BN_pseudo_rand().
1943 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1944 bignum version of BN_from_montgomery() with the working code from
1945 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1949 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1950 made it impossible to use the same SSL_SESSION data structure in
1951 SSL2 clients in multiple threads.
1954 *) The return value of RAND_load_file() no longer counts bytes obtained
1955 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1956 to seed the PRNG (previously an explicit byte count was required).
1957 [Ulf Möller, Bodo Möller]
1959 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1960 used (char *) instead of (void *) and had casts all over the place.
1963 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1966 *) Retain source code compatibility for BN_prime_checks macro:
1967 BN_is_prime(..., BN_prime_checks, ...) now uses
1968 BN_prime_checks_for_size to determine the appropriate number of
1969 Rabin-Miller iterations.
1972 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1973 DH_CHECK_P_NOT_SAFE_PRIME.
1974 (Check if this is true? OpenPGP calls them "strong".)
1977 *) Merge the functionality of "dh" and "gendh" programs into a new program
1978 "dhparam". The old programs are retained for now but will handle DH keys
1979 (instead of parameters) in future.
1982 *) Make the ciphers, s_server and s_client programs check the return values
1983 when a new cipher list is set.
1986 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1987 ciphers. Before when the 56bit ciphers were enabled the sorting was
1990 The syntax for the cipher sorting has been extended to support sorting by
1991 cipher-strength (using the strength_bits hard coded in the tables).
1992 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1994 Fix a bug in the cipher-command parser: when supplying a cipher command
1995 string with an "undefined" symbol (neither command nor alphanumeric
1996 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1997 an error is flagged.
1999 Due to the strength-sorting extension, the code of the
2000 ssl_create_cipher_list() function was completely rearranged. I hope that
2001 the readability was also increased :-)
2002 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2004 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
2005 for the first serial number and places 2 in the serial number file. This
2006 avoids problems when the root CA is created with serial number zero and
2007 the first user certificate has the same issuer name and serial number
2011 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
2012 the new code. Add documentation for this stuff.
2015 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
2016 X509_*() to X509at_*() on the grounds that they don't handle X509
2017 structures and behave in an analagous way to the X509v3 functions:
2018 they shouldn't be called directly but wrapper functions should be used
2021 So we also now have some wrapper functions that call the X509at functions
2022 when passed certificate requests. (TO DO: similar things can be done with
2023 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
2024 things. Some of these need some d2i or i2d and print functionality
2025 because they handle more complex structures.)
2028 *) Add missing #ifndefs that caused missing symbols when building libssl
2029 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
2030 NO_RSA in ssl/s2*.c.
2031 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
2033 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
2034 has a return value which indicates the quality of the random data
2035 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
2036 error queue. New function RAND_pseudo_bytes() generates output that is
2037 guaranteed to be unique but not unpredictable. RAND_add is like
2038 RAND_seed, but takes an extra argument for an entropy estimate
2039 (RAND_seed always assumes full entropy).
2042 *) Do more iterations of Rabin-Miller probable prime test (specifically,
2043 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
2044 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
2045 in crypto/bn/bn_prime.c for the complete table). This guarantees a
2046 false-positive rate of at most 2^-80 for random input.
2049 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
2052 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
2053 in the 0.9.5 release), this returns the chain
2054 from an X509_CTX structure with a dup of the stack and all
2055 the X509 reference counts upped: so the stack will exist
2056 after X509_CTX_cleanup() has been called. Modify pkcs12.c
2059 Also make SSL_SESSION_print() print out the verify return
2063 *) Add manpage for the pkcs12 command. Also change the default
2064 behaviour so MAC iteration counts are used unless the new
2065 -nomaciter option is used. This improves file security and
2066 only older versions of MSIE (4.0 for example) need it.
2069 *) Honor the no-xxx Configure options when creating .DEF files.
2072 *) Add PKCS#10 attributes to field table: challengePassword,
2073 unstructuredName and unstructuredAddress. These are taken from
2074 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
2075 international characters are used.
2077 More changes to X509_ATTRIBUTE code: allow the setting of types
2078 based on strings. Remove the 'loc' parameter when adding
2079 attributes because these will be a SET OF encoding which is sorted
2083 *) Initial changes to the 'req' utility to allow request generation
2084 automation. This will allow an application to just generate a template
2085 file containing all the field values and have req construct the
2088 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2089 used all over the place including certificate requests and PKCS#7
2090 structures. They are currently handled manually where necessary with
2091 some primitive wrappers for PKCS#7. The new functions behave in a
2092 manner analogous to the X509 extension functions: they allow
2093 attributes to be looked up by NID and added.
2095 Later something similar to the X509V3 code would be desirable to
2096 automatically handle the encoding, decoding and printing of the
2097 more complex types. The string types like challengePassword can
2098 be handled by the string table functions.
2100 Also modified the multi byte string table handling. Now there is
2101 a 'global mask' which masks out certain types. The table itself
2102 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2103 is useful when for example there is only one permissible type
2104 (as in countryName) and using the mask might result in no valid
2108 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2109 SSL_get_peer_finished to allow applications to obtain the latest
2110 Finished messages sent to the peer or expected from the peer,
2111 respectively. (SSL_get_peer_finished is usually the Finished message
2112 actually received from the peer, otherwise the protocol will be aborted.)
2114 As the Finished message are message digests of the complete handshake
2115 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2116 be used for external authentication procedures when the authentication
2117 provided by SSL/TLS is not desired or is not enough.
2120 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2121 the host supports BWX extension and if Compaq C is present on the
2122 $PATH. Just exploiting of the BWX extension results in 20-30%
2123 performance kick for some algorithms, e.g. DES and RC4 to mention
2124 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2128 *) Add support for MS "fast SGC". This is arguably a violation of the
2129 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2130 weak crypto and after checking the certificate is SGC a second one
2131 with strong crypto. MS SGC stops the first handshake after receiving
2132 the server certificate message and sends a second client hello. Since
2133 a server will typically do all the time consuming operations before
2134 expecting any further messages from the client (server key exchange
2135 is the most expensive) there is little difference between the two.
2137 To get OpenSSL to support MS SGC we have to permit a second client
2138 hello message after we have sent server done. In addition we have to
2139 reset the MAC if we do get this second client hello.
2142 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2143 if a DER encoded private key is RSA or DSA traditional format. Changed
2144 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2145 format DER encoded private key. Newer code should use PKCS#8 format which
2146 has the key type encoded in the ASN1 structure. Added DER private key
2147 support to pkcs8 application.
2150 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2151 ciphersuites has been selected (as required by the SSL 3/TLS 1
2152 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2153 is set, we interpret this as a request to violate the specification
2154 (the worst that can happen is a handshake failure, and 'correct'
2155 behaviour would result in a handshake failure anyway).
2158 *) In SSL_CTX_add_session, take into account that there might be multiple
2159 SSL_SESSION structures with the same session ID (e.g. when two threads
2160 concurrently obtain them from an external cache).
2161 The internal cache can handle only one SSL_SESSION with a given ID,
2162 so if there's a conflict, we now throw out the old one to achieve
2166 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2167 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2168 some routines that use cipher OIDs: some ciphers do not have OIDs
2169 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2173 *) Simplify the trust setting structure and code. Now we just have
2174 two sequences of OIDs for trusted and rejected settings. These will
2175 typically have values the same as the extended key usage extension
2176 and any application specific purposes.
2178 The trust checking code now has a default behaviour: it will just
2179 check for an object with the same NID as the passed id. Functions can
2180 be provided to override either the default behaviour or the behaviour
2181 for a given id. SSL client, server and email already have functions
2182 in place for compatibility: they check the NID and also return "trusted"
2183 if the certificate is self signed.
2186 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2187 traditional format into an EVP_PKEY structure.
2190 *) Add a password callback function PEM_cb() which either prompts for
2191 a password if usr_data is NULL or otherwise assumes it is a null
2192 terminated password. Allow passwords to be passed on command line
2193 environment or config files in a few more utilities.
2196 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2197 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2198 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2199 Update documentation.
2202 *) Support for ASN1 "NULL" type. This could be handled before by using
2203 ASN1_TYPE but there wasn't any function that would try to read a NULL
2204 and produce an error if it couldn't. For compatibility we also have
2205 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2206 don't allocate anything because they don't need to.
2209 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2211 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2213 *) Rebuild of the memory allocation routines used by OpenSSL code and
2214 possibly others as well. The purpose is to make an interface that
2215 provide hooks so anyone can build a separate set of allocation and
2216 deallocation routines to be used by OpenSSL, for example memory
2217 pool implementations, or something else, which was previously hard
2218 since Malloc(), Realloc() and Free() were defined as macros having
2219 the values malloc, realloc and free, respectively (except for Win32
2220 compilations). The same is provided for memory debugging code.
2221 OpenSSL already comes with functionality to find memory leaks, but
2222 this gives people a chance to debug other memory problems.
2224 With these changes, a new set of functions and macros have appeared:
2226 CRYPTO_set_mem_debug_functions() [F]
2227 CRYPTO_get_mem_debug_functions() [F]
2228 CRYPTO_dbg_set_options() [F]
2229 CRYPTO_dbg_get_options() [F]
2230 CRYPTO_malloc_debug_init() [M]
2232 The memory debug functions are NULL by default, unless the library
2233 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2234 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2235 gives the standard debugging functions that come with OpenSSL) or
2236 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2237 provided by the library user) must be used. When the standard
2238 debugging functions are used, CRYPTO_dbg_set_options can be used to
2239 request additional information:
2240 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2241 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2243 Also, things like CRYPTO_set_mem_functions will always give the
2244 expected result (the new set of functions is used for allocation
2245 and deallocation) at all times, regardless of platform and compiler
2248 To finish it up, some functions that were never use in any other
2249 way than through macros have a new API and new semantic:
2252 CRYPTO_dbg_realloc()
2255 All macros of value have retained their old syntax.
2256 [Richard Levitte and Bodo Moeller]
2258 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2259 ordering of SMIMECapabilities wasn't in "strength order" and there
2260 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2264 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2265 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2266 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2268 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2269 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2270 functionality to handle multipart/signed properly) and a utility
2271 called 'smime' to call all this stuff. This is based on code I
2272 originally wrote for Celo who have kindly allowed it to be
2273 included in OpenSSL.
2276 *) Add variants des_set_key_checked and des_set_key_unchecked of
2277 des_set_key (aka des_key_sched). Global variable des_check_key
2278 decides which of these is called by des_set_key; this way
2279 des_check_key behaves as it always did, but applications and
2280 the library itself, which was buggy for des_check_key == 1,
2281 have a cleaner way to pick the version they need.
2284 *) New function PKCS12_newpass() which changes the password of a
2288 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2289 dynamic mix. In both cases the ids can be used as an index into the
2290 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2291 functions so they accept a list of the field values and the
2292 application doesn't need to directly manipulate the X509_TRUST
2296 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2300 *) Modify the way the V3 extension code looks up extensions. This now
2301 works in a similar way to the object code: we have some "standard"
2302 extensions in a static table which is searched with OBJ_bsearch()
2303 and the application can add dynamic ones if needed. The file
2304 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2305 updated whenever a new extension is added to the core code and kept
2306 in ext_nid order. There is a simple program 'tabtest.c' which checks
2307 this. New extensions are not added too often so this file can readily
2308 be maintained manually.
2310 There are two big advantages in doing things this way. The extensions
2311 can be looked up immediately and no longer need to be "added" using
2312 X509V3_add_standard_extensions(): this function now does nothing.
2313 [Side note: I get *lots* of email saying the extension code doesn't
2314 work because people forget to call this function]
2315 Also no dynamic allocation is done unless new extensions are added:
2316 so if we don't add custom extensions there is no need to call
2317 X509V3_EXT_cleanup().
2320 *) Modify enc utility's salting as follows: make salting the default. Add a
2321 magic header, so unsalted files fail gracefully instead of just decrypting
2322 to garbage. This is because not salting is a big security hole, so people
2323 should be discouraged from doing it.
2326 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2327 digest to be passed on the command line but it only used this
2328 parameter when signing a certificate. Modified so all relevant
2329 operations are affected by the digest parameter including the
2330 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2331 DSA key was used because it didn't fix the digest.
2334 *) Initial certificate chain verify code. Currently tests the untrusted
2335 certificates for consistency with the verify purpose (which is set
2336 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2338 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2339 this is because it will reject chains with invalid extensions whereas
2340 every previous version of OpenSSL and SSLeay made no checks at all.
2342 Trust code: checks the root CA for the relevant trust settings. Trust
2343 settings have an initial value consistent with the verify purpose: e.g.
2344 if the verify purpose is for SSL client use it expects the CA to be
2345 trusted for SSL client use. However the default value can be changed to
2346 permit custom trust settings: one example of this would be to only trust
2347 certificates from a specific "secure" set of CAs.
2349 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2350 which should be used for version portability: especially since the
2351 verify structure is likely to change more often now.
2353 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2354 to set them. If not set then assume SSL clients will verify SSL servers
2357 Two new options to the verify program: -untrusted allows a set of
2358 untrusted certificates to be passed in and -purpose which sets the
2359 intended purpose of the certificate. If a purpose is set then the
2360 new chain verify code is used to check extension consistency.
2363 *) Support for the authority information access extension.
2366 *) Modify RSA and DSA PEM read routines to transparently handle
2367 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2368 public keys in a format compatible with certificate
2369 SubjectPublicKeyInfo structures. Unfortunately there were already
2370 functions called *_PublicKey_* which used various odd formats so
2371 these are retained for compatibility: however the DSA variants were
2372 never in a public release so they have been deleted. Changed dsa/rsa
2373 utilities to handle the new format: note no releases ever handled public
2374 keys so we should be OK.
2376 The primary motivation for this change is to avoid the same fiasco
2377 that dogs private keys: there are several incompatible private key
2378 formats some of which are standard and some OpenSSL specific and
2379 require various evil hacks to allow partial transparent handling and
2380 even then it doesn't work with DER formats. Given the option anything
2381 other than PKCS#8 should be dumped: but the other formats have to
2382 stay in the name of compatibility.
2384 With public keys and the benefit of hindsight one standard format
2385 is used which works with EVP_PKEY, RSA or DSA structures: though
2386 it clearly returns an error if you try to read the wrong kind of key.
2388 Added a -pubkey option to the 'x509' utility to output the public key.
2389 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2390 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2391 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2392 that do the same as the EVP_PKEY_assign_*() except they up the
2393 reference count of the added key (they don't "swallow" the
2397 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2398 CRLs would fail if the file contained no certificates or no CRLs:
2399 added a new function to read in both types and return the number
2400 read: this means that if none are read it will be an error. The
2401 DER versions of the certificate and CRL reader would always fail
2402 because it isn't possible to mix certificates and CRLs in DER format
2403 without choking one or the other routine. Changed this to just read
2404 a certificate: this is the best we can do. Also modified the code
2405 in apps/verify.c to take notice of return codes: it was previously
2406 attempting to read in certificates from NULL pointers and ignoring
2407 any errors: this is one reason why the cert and CRL reader seemed
2408 to work. It doesn't check return codes from the default certificate
2409 routines: these may well fail if the certificates aren't installed.
2412 *) Code to support otherName option in GeneralName.
2415 *) First update to verify code. Change the verify utility
2416 so it warns if it is passed a self signed certificate:
2417 for consistency with the normal behaviour. X509_verify
2418 has been modified to it will now verify a self signed
2419 certificate if *exactly* the same certificate appears
2420 in the store: it was previously impossible to trust a
2421 single self signed certificate. This means that:
2422 openssl verify ss.pem
2423 now gives a warning about a self signed certificate but
2424 openssl verify -CAfile ss.pem ss.pem
2428 *) For servers, store verify_result in SSL_SESSION data structure
2429 (and add it to external session representation).
2430 This is needed when client certificate verifications fails,
2431 but an application-provided verification callback (set by
2432 SSL_CTX_set_cert_verify_callback) allows accepting the session
2433 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2434 but returns 1): When the session is reused, we have to set
2435 ssl->verify_result to the appropriate error code to avoid
2437 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2439 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2440 case in PKCS7_dataInit() where the signed PKCS7 structure
2441 didn't contain any existing data because it was being created.
2442 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2444 *) Add a salt to the key derivation routines in enc.c. This
2445 forms the first 8 bytes of the encrypted file. Also add a
2446 -S option to allow a salt to be input on the command line.
2449 *) New function X509_cmp(). Oddly enough there wasn't a function
2450 to compare two certificates. We do this by working out the SHA1
2451 hash and comparing that. X509_cmp() will be needed by the trust
2455 *) SSL_get1_session() is like SSL_get_session(), but increments
2456 the reference count in the SSL_SESSION returned.
2457 [Geoff Thorpe <geoff@eu.c2.net>]
2459 *) Fix for 'req': it was adding a null to request attributes.
2460 Also change the X509_LOOKUP and X509_INFO code to handle
2461 certificate auxiliary information.
2464 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2468 *) Add the possibility to add extra information to the memory leak
2469 detecting output, to form tracebacks, showing from where each
2470 allocation was originated: CRYPTO_push_info("constant string") adds
2471 the string plus current file name and line number to a per-thread
2472 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2473 is like calling CYRPTO_pop_info() until the stack is empty.
2474 Also updated memory leak detection code to be multi-thread-safe.
2477 *) Add options -text and -noout to pkcs7 utility and delete the
2478 encryption options which never did anything. Update docs.
2481 *) Add options to some of the utilities to allow the pass phrase
2482 to be included on either the command line (not recommended on
2483 OSes like Unix) or read from the environment. Update the
2484 manpages and fix a few bugs.
2487 *) Add a few manpages for some of the openssl commands.
2490 *) Fix the -revoke option in ca. It was freeing up memory twice,
2491 leaking and not finding already revoked certificates.
2494 *) Extensive changes to support certificate auxiliary information.
2495 This involves the use of X509_CERT_AUX structure and X509_AUX
2496 functions. An X509_AUX function such as PEM_read_X509_AUX()
2497 can still read in a certificate file in the usual way but it
2498 will also read in any additional "auxiliary information". By
2499 doing things this way a fair degree of compatibility can be
2500 retained: existing certificates can have this information added
2501 using the new 'x509' options.
2503 Current auxiliary information includes an "alias" and some trust
2504 settings. The trust settings will ultimately be used in enhanced
2505 certificate chain verification routines: currently a certificate
2506 can only be trusted if it is self signed and then it is trusted
2510 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2511 The problem was that one of the replacement routines had not been working
2512 since SSLeay releases. For now the offending routine has been replaced
2513 with non-optimised assembler. Even so, this now gives around 95%
2514 performance improvement for 1024 bit RSA signs.
2517 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2518 handling. Most clients have the effective key size in bits equal to
2519 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2520 A few however don't do this and instead use the size of the decrypted key
2521 to determine the RC2 key length and the AlgorithmIdentifier to determine
2522 the effective key length. In this case the effective key length can still
2523 be 40 bits but the key length can be 168 bits for example. This is fixed
2524 by manually forcing an RC2 key into the EVP_PKEY structure because the
2525 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2526 the key length and effective key length are equal.
2529 *) Add a bunch of functions that should simplify the creation of
2530 X509_NAME structures. Now you should be able to do:
2531 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2532 and have it automatically work out the correct field type and fill in
2533 the structures. The more adventurous can try:
2534 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2535 and it will (hopefully) work out the correct multibyte encoding.
2538 *) Change the 'req' utility to use the new field handling and multibyte
2539 copy routines. Before the DN field creation was handled in an ad hoc
2540 way in req, ca, and x509 which was rather broken and didn't support
2541 BMPStrings or UTF8Strings. Since some software doesn't implement
2542 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2543 using the dirstring_type option. See the new comment in the default
2544 openssl.cnf for more info.
2547 *) Make crypto/rand/md_rand.c more robust:
2548 - Assure unique random numbers after fork().
2549 - Make sure that concurrent threads access the global counter and
2550 md serializably so that we never lose entropy in them
2551 or use exactly the same state in multiple threads.
2552 Access to the large state is not always serializable because
2553 the additional locking could be a performance killer, and
2554 md should be large enough anyway.
2557 *) New file apps/app_rand.c with commonly needed functionality
2558 for handling the random seed file.
2560 Use the random seed file in some applications that previously did not:
2562 dsaparam -genkey (which also ignored its '-rand' option),
2565 x509 (when signing).
2566 Except on systems with /dev/urandom, it is crucial to have a random
2567 seed file at least for key creation, DSA signing, and for DH exchanges;
2568 for RSA signatures we could do without one.
2570 gendh and gendsa (unlike genrsa) used to read only the first byte
2571 of each file listed in the '-rand' option. The function as previously
2572 found in genrsa is now in app_rand.c and is used by all programs
2573 that support '-rand'.
2576 *) In RAND_write_file, use mode 0600 for creating files;
2577 don't just chmod when it may be too late.
2580 *) Report an error from X509_STORE_load_locations
2581 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2584 *) New function ASN1_mbstring_copy() this copies a string in either
2585 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2586 into an ASN1_STRING type. A mask of permissible types is passed
2587 and it chooses the "minimal" type to use or an error if not type
2591 *) Add function equivalents to the various macros in asn1.h. The old
2592 macros are retained with an M_ prefix. Code inside the library can
2593 use the M_ macros. External code (including the openssl utility)
2594 should *NOT* in order to be "shared library friendly".
2597 *) Add various functions that can check a certificate's extensions
2598 to see if it usable for various purposes such as SSL client,
2599 server or S/MIME and CAs of these types. This is currently
2600 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2601 verification. Also added a -purpose flag to x509 utility to
2602 print out all the purposes.
2605 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2609 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2610 for, obtain and decode and extension and obtain its critical flag.
2611 This allows all the necessary extension code to be handled in a
2612 single function call.
2615 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2616 platforms. See crypto/rc4/rc4_enc.c for further details.
2619 *) New -noout option to asn1parse. This causes no output to be produced
2620 its main use is when combined with -strparse and -out to extract data
2621 from a file (which may not be in ASN.1 format).
2624 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2625 when producing the local key id.
2626 [Richard Levitte <levitte@stacken.kth.se>]
2628 *) New option -dhparam in s_server. This allows a DH parameter file to be
2629 stated explicitly. If it is not stated then it tries the first server
2630 certificate file. The previous behaviour hard coded the filename
2634 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2635 a public key to be input or output. For example:
2636 openssl rsa -in key.pem -pubout -out pubkey.pem
2637 Also added necessary DSA public key functions to handle this.
2640 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2641 in the message. This was handled by allowing
2642 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2643 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2645 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2646 to the end of the strings whereas this didn't. This would cause problems
2647 if strings read with d2i_ASN1_bytes() were later modified.
2648 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2650 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2651 data and it contains EOF it will end up returning an error. This is
2652 caused by input 46 bytes long. The cause is due to the way base64
2653 BIOs find the start of base64 encoded data. They do this by trying a
2654 trial decode on each line until they find one that works. When they
2655 do a flag is set and it starts again knowing it can pass all the
2656 data directly through the decoder. Unfortunately it doesn't reset
2657 the context it uses. This means that if EOF is reached an attempt
2658 is made to pass two EOFs through the context and this causes the
2659 resulting error. This can also cause other problems as well. As is
2660 usual with these problems it takes *ages* to find and the fix is
2661 trivial: move one line.
2662 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2664 *) Ugly workaround to get s_client and s_server working under Windows. The
2665 old code wouldn't work because it needed to select() on sockets and the
2666 tty (for keypresses and to see if data could be written). Win32 only
2667 supports select() on sockets so we select() with a 1s timeout on the
2668 sockets and then see if any characters are waiting to be read, if none
2669 are present then we retry, we also assume we can always write data to
2670 the tty. This isn't nice because the code then blocks until we've
2671 received a complete line of data and it is effectively polling the
2672 keyboard at 1s intervals: however it's quite a bit better than not
2673 working at all :-) A dedicated Windows application might handle this
2674 with an event loop for example.
2677 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2678 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2679 will be called when RSA_sign() and RSA_verify() are used. This is useful
2680 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2681 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2682 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2683 This necessitated the support of an extra signature type NID_md5_sha1
2684 for SSL signatures and modifications to the SSL library to use it instead
2685 of calling RSA_public_decrypt() and RSA_private_encrypt().
2688 *) Add new -verify -CAfile and -CApath options to the crl program, these
2689 will lookup a CRL issuers certificate and verify the signature in a
2690 similar way to the verify program. Tidy up the crl program so it
2691 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2692 less strict. It will now permit CRL extensions even if it is not
2693 a V2 CRL: this will allow it to tolerate some broken CRLs.
2696 *) Initialize all non-automatic variables each time one of the openssl
2697 sub-programs is started (this is necessary as they may be started
2698 multiple times from the "OpenSSL>" prompt).
2699 [Lennart Bang, Bodo Moeller]
2701 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2702 removing all other RSA functionality (this is what NO_RSA does). This
2703 is so (for example) those in the US can disable those operations covered
2704 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2708 *) Non-copying interface to BIO pairs.
2709 (still largely untested)
2712 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2713 ASCII string. This was handled independently in various places before.
2716 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2717 UTF8 strings a character at a time.
2720 *) Use client_version from client hello to select the protocol
2721 (s23_srvr.c) and for RSA client key exchange verification
2722 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2725 *) Add various utility functions to handle SPKACs, these were previously
2726 handled by poking round in the structure internals. Added new function
2727 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2728 print, verify and generate SPKACs. Based on an original idea from
2729 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2732 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2735 *) Allow the config file extension section to be overwritten on the
2736 command line. Based on an original idea from Massimiliano Pala
2737 <madwolf@comune.modena.it>. The new option is called -extensions
2738 and can be applied to ca, req and x509. Also -reqexts to override
2739 the request extensions in req and -crlexts to override the crl extensions
2743 *) Add new feature to the SPKAC handling in ca. Now you can include
2744 the same field multiple times by preceding it by "XXXX." for example:
2747 this is the same syntax as used in the req config file.
2750 *) Allow certificate extensions to be added to certificate requests. These
2751 are specified in a 'req_extensions' option of the req section of the
2752 config file. They can be printed out with the -text option to req but
2753 are otherwise ignored at present.
2756 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2757 data read consists of only the final block it would not decrypted because
2758 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2759 A misplaced 'break' also meant the decrypted final block might not be
2760 copied until the next read.
2763 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2764 a few extra parameters to the DH structure: these will be useful if
2765 for example we want the value of 'q' or implement X9.42 DH.
2768 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2769 provides hooks that allow the default DSA functions or functions on a
2770 "per key" basis to be replaced. This allows hardware acceleration and
2771 hardware key storage to be handled without major modification to the
2772 library. Also added low level modexp hooks and CRYPTO_EX structure and
2773 associated functions.
2776 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2777 as "read only": it can't be written to and the buffer it points to will
2778 not be freed. Reading from a read only BIO is much more efficient than
2779 a normal memory BIO. This was added because there are several times when
2780 an area of memory needs to be read from a BIO. The previous method was
2781 to create a memory BIO and write the data to it, this results in two
2782 copies of the data and an O(n^2) reading algorithm. There is a new
2783 function BIO_new_mem_buf() which creates a read only memory BIO from
2784 an area of memory. Also modified the PKCS#7 routines to use read only
2788 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2789 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2790 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2791 but a retry condition occured while trying to read the rest.
2794 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2795 NID_pkcs7_encrypted by default: this was wrong since this should almost
2796 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2797 the encrypted data type: this is a more sensible place to put it and it
2798 allows the PKCS#12 code to be tidied up that duplicated this
2802 *) Changed obj_dat.pl script so it takes its input and output files on
2803 the command line. This should avoid shell escape redirection problems
2807 *) Initial support for certificate extension requests, these are included
2808 in things like Xenroll certificate requests. Included functions to allow
2809 extensions to be obtained and added.
2812 *) -crlf option to s_client and s_server for sending newlines as
2813 CRLF (as required by many protocols).
2816 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2818 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2819 [Ralf S. Engelschall]
2821 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2822 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2824 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2828 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2829 DH parameters/keys (q is lost during that conversion, but the resulting
2830 DH parameters contain its length).
2832 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2833 much faster than DH_generate_parameters (which creates parameters
2834 where p = 2*q + 1), and also the smaller q makes DH computations
2835 much more efficient (160-bit exponentiation instead of 1024-bit
2836 exponentiation); so this provides a convenient way to support DHE
2837 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2838 utter importance to use
2839 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2841 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2842 when such DH parameters are used, because otherwise small subgroup
2843 attacks may become possible!
2846 *) Avoid memory leak in i2d_DHparams.
2849 *) Allow the -k option to be used more than once in the enc program:
2850 this allows the same encrypted message to be read by multiple recipients.
2853 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2854 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2855 it will always use the numerical form of the OID, even if it has a short
2859 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2860 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2861 otherwise bn_mod_exp was called. In the case of hardware keys for example
2862 no private key components need be present and it might store extra data
2863 in the RSA structure, which cannot be accessed from bn_mod_exp.
2864 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2865 private key operations.
2868 *) Added support for SPARC Linux.
2871 *) pem_password_cb function type incompatibly changed from
2872 typedef int pem_password_cb(char *buf, int size, int rwflag);
2874 ....(char *buf, int size, int rwflag, void *userdata);
2875 so that applications can pass data to their callbacks:
2876 The PEM[_ASN1]_{read,write}... functions and macros now take an
2877 additional void * argument, which is just handed through whenever
2878 the password callback is called.
2879 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2881 New function SSL_CTX_set_default_passwd_cb_userdata.
2883 Compatibility note: As many C implementations push function arguments
2884 onto the stack in reverse order, the new library version is likely to
2885 interoperate with programs that have been compiled with the old
2886 pem_password_cb definition (PEM_whatever takes some data that
2887 happens to be on the stack as its last argument, and the callback
2888 just ignores this garbage); but there is no guarantee whatsoever that
2891 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2892 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2893 problems not only on Windows, but also on some Unix platforms.
2894 To avoid problematic command lines, these definitions are now in an
2895 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2896 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2899 *) MIPS III/IV assembler module is reimplemented.
2902 *) More DES library cleanups: remove references to srand/rand and
2903 delete an unused file.
2906 *) Add support for the the free Netwide assembler (NASM) under Win32,
2907 since not many people have MASM (ml) and it can be hard to obtain.
2908 This is currently experimental but it seems to work OK and pass all
2909 the tests. Check out INSTALL.W32 for info.
2912 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2913 without temporary keys kept an extra copy of the server key,
2914 and connections with temporary keys did not free everything in case
2918 *) New function RSA_check_key and new openssl rsa option -check
2919 for verifying the consistency of RSA keys.
2920 [Ulf Moeller, Bodo Moeller]
2922 *) Various changes to make Win32 compile work:
2923 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2924 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2925 comparison" warnings.
2926 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2929 *) Add a debugging option to PKCS#5 v2 key generation function: when
2930 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2931 derived keys are printed to stderr.
2934 *) Copy the flags in ASN1_STRING_dup().
2935 [Roman E. Pavlov <pre@mo.msk.ru>]
2937 *) The x509 application mishandled signing requests containing DSA
2938 keys when the signing key was also DSA and the parameters didn't match.
2940 It was supposed to omit the parameters when they matched the signing key:
2941 the verifying software was then supposed to automatically use the CA's
2942 parameters if they were absent from the end user certificate.
2944 Omitting parameters is no longer recommended. The test was also
2945 the wrong way round! This was probably due to unusual behaviour in
2946 EVP_cmp_parameters() which returns 1 if the parameters match.
2947 This meant that parameters were omitted when they *didn't* match and
2948 the certificate was useless. Certificates signed with 'ca' didn't have
2950 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2952 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2953 The interface is as follows:
2954 Applications can use
2955 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2956 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2957 "off" is now the default.
2958 The library internally uses
2959 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2960 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2961 to disable memory-checking temporarily.
2963 Some inconsistent states that previously were possible (and were
2964 even the default) are now avoided.
2966 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2967 with each memory chunk allocated; this is occasionally more helpful
2968 than just having a counter.
2970 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2972 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2976 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2977 which largely parallels "options", but is for changing API behaviour,
2978 whereas "options" are about protocol behaviour.
2979 Initial "mode" flags are:
2981 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2982 a single record has been written.
2983 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2984 retries use the same buffer location.
2985 (But all of the contents must be
2989 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2992 *) Fix problems with no-hmac etc.
2993 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2995 *) New functions RSA_get_default_method(), RSA_set_method() and
2996 RSA_get_method(). These allows replacement of RSA_METHODs without having
2997 to mess around with the internals of an RSA structure.
3000 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
3001 Also really enable memory leak checks in openssl.c and in some
3003 [Chad C. Mulligan, Bodo Moeller]
3005 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
3006 up the length of negative integers. This has now been simplified to just
3007 store the length when it is first determined and use it later, rather
3008 than trying to keep track of where data is copied and updating it to
3010 [Steve Henson, reported by Brien Wheeler
3011 <bwheeler@authentica-security.com>]
3013 *) Add a new function PKCS7_signatureVerify. This allows the verification
3014 of a PKCS#7 signature but with the signing certificate passed to the
3015 function itself. This contrasts with PKCS7_dataVerify which assumes the
3016 certificate is present in the PKCS#7 structure. This isn't always the
3017 case: certificates can be omitted from a PKCS#7 structure and be
3018 distributed by "out of band" means (such as a certificate database).
3021 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
3022 function prototypes in pem.h, also change util/mkdef.pl to add the
3023 necessary function names.
3026 *) mk1mf.pl (used by Windows builds) did not properly read the
3027 options set by Configure in the top level Makefile, and Configure
3028 was not even able to write more than one option correctly.
3029 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
3032 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
3033 file to be loaded from a BIO or FILE pointer. The BIO version will
3034 for example allow memory BIOs to contain config info.
3037 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
3038 Whoever hopes to achieve shared-library compatibility across versions
3039 must use this, not the compile-time macro.
3040 (Exercise 0.9.4: Which is the minimum library version required by
3042 Note: All this applies only to multi-threaded programs, others don't
3046 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
3047 through a BIO pair triggered the default case, i.e.
3048 SSLerr(...,SSL_R_UNKNOWN_STATE).
3051 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
3052 can use the SSL library even if none of the specific BIOs is
3056 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
3057 for the encoded length.
3058 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
3060 *) Add initial documentation of the X509V3 functions.
3063 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
3064 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
3065 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
3066 secure PKCS#8 private key format with a high iteration count.
3069 *) Fix determination of Perl interpreter: A perl or perl5
3070 _directory_ in $PATH was also accepted as the interpreter.
3071 [Ralf S. Engelschall]
3073 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
3074 wrong with it but it was very old and did things like calling
3075 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
3079 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
3080 to use the new extension code.
3083 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
3084 with macros. This should make it easier to change their form, add extra
3085 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3089 *) Add to configuration table a new entry that can specify an alternative
3090 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3091 according to Mark Crispin <MRC@Panda.COM>.
3095 *) DES CBC did not update the IV. Weird.
3098 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3099 Changing the behaviour of the former might break existing programs --
3100 where IV updating is needed, des_ncbc_encrypt can be used.
3103 *) When bntest is run from "make test" it drives bc to check its
3104 calculations, as well as internally checking them. If an internal check
3105 fails, it needs to cause bc to give a non-zero result or make test carries
3106 on without noticing the failure. Fixed.
3109 *) DES library cleanups.
3112 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3113 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3114 ciphers. NOTE: although the key derivation function has been verified
3115 against some published test vectors it has not been extensively tested
3116 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3120 *) Instead of "mkdir -p", which is not fully portable, use new
3121 Perl script "util/mkdir-p.pl".
3124 *) Rewrite the way password based encryption (PBE) is handled. It used to
3125 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3126 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3127 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3128 the 'parameter' field of the AlgorithmIdentifier is passed to the
3129 underlying key generation function so it must do its own ASN1 parsing.
3130 This has also changed the EVP_PBE_CipherInit() function which now has a
3131 'parameter' argument instead of literal salt and iteration count values
3132 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3135 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3136 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3137 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3138 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3139 value was just used as a "magic string" and not used directly its
3140 value doesn't matter.
3143 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3147 *) "linux-sparc64" configuration (ultrapenguin).
3148 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3149 "linux-sparc" configuration.
3150 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3152 *) config now generates no-xxx options for missing ciphers.
3155 *) Support the EBCDIC character set (work in progress).
3156 File ebcdic.c not yet included because it has a different license.
3157 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3159 *) Support BS2000/OSD-POSIX.
3160 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3162 *) Make callbacks for key generation use void * instead of char *.
3165 *) Make S/MIME samples compile (not yet tested).
3168 *) Additional typesafe stacks.
3171 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3175 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3177 *) New configuration variant "sco5-gcc".
3179 *) Updated some demos.
3180 [Sean O Riordain, Wade Scholine]
3182 *) Add missing BIO_free at exit of pkcs12 application.
3185 *) Fix memory leak in conf.c.
3188 *) Updates for Win32 to assembler version of MD5.
3191 *) Set #! path to perl in apps/der_chop to where we found it
3192 instead of using a fixed path.
3195 *) SHA library changes for irix64-mips4-cc.
3198 *) Improvements for VMS support.
3202 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3204 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3205 This also avoids the problems with SC4.2 and unpatched SC5.
3206 [Andy Polyakov <appro@fy.chalmers.se>]
3208 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3209 These are required because of the typesafe stack would otherwise break
3210 existing code. If old code used a structure member which used to be STACK
3211 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3212 sk_num or sk_value it would produce an error because the num, data members
3213 are not present in STACK_OF. Now it just produces a warning. sk_set
3214 replaces the old method of assigning a value to sk_value
3215 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3216 that does this will no longer work (and should use sk_set instead) but
3217 this could be regarded as a "questionable" behaviour anyway.
3220 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3221 correctly handle encrypted S/MIME data.
3224 *) Change type of various DES function arguments from des_cblock
3225 (which means, in function argument declarations, pointer to char)
3226 to des_cblock * (meaning pointer to array with 8 char elements),
3227 which allows the compiler to do more typechecking; it was like
3228 that back in SSLeay, but with lots of ugly casts.
3230 Introduce new type const_des_cblock.
3233 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3234 problems: find RecipientInfo structure that matches recipient certificate
3235 and initialise the ASN1 structures properly based on passed cipher.
3238 *) Belatedly make the BN tests actually check the results.
3241 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3242 to and from BNs: it was completely broken. New compilation option
3243 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3244 key elements as negative integers.
3247 *) Reorganize and speed up MD5.
3248 [Andy Polyakov <appro@fy.chalmers.se>]
3251 [Richard Levitte <richard@levitte.org>]
3253 *) New option -out to asn1parse to allow the parsed structure to be
3254 output to a file. This is most useful when combined with the -strparse
3255 option to examine the output of things like OCTET STRINGS.
3258 *) Make SSL library a little more fool-proof by not requiring any longer
3259 that SSL_set_{accept,connect}_state be called before
3260 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3261 in many applications because usually everything *appeared* to work as
3262 intended anyway -- now it really works as intended).
3265 *) Move openssl.cnf out of lib/.
3268 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3269 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3270 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3271 [Ralf S. Engelschall]
3273 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3274 handle PKCS#7 enveloped data properly.
3275 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3277 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3278 copying pointers. The cert_st handling is changed by this in
3279 various ways (and thus what used to be known as ctx->default_cert
3280 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3281 any longer when s->cert does not give us what we need).
3282 ssl_cert_instantiate becomes obsolete by this change.
3283 As soon as we've got the new code right (possibly it already is?),
3284 we have solved a couple of bugs of the earlier code where s->cert
3285 was used as if it could not have been shared with other SSL structures.
3287 Note that using the SSL API in certain dirty ways now will result
3288 in different behaviour than observed with earlier library versions:
3289 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3290 does not influence s as it used to.
3292 In order to clean up things more thoroughly, inside SSL_SESSION
3293 we don't use CERT any longer, but a new structure SESS_CERT
3294 that holds per-session data (if available); currently, this is
3295 the peer's certificate chain and, for clients, the server's certificate
3296 and temporary key. CERT holds only those values that can have
3297 meaningful defaults in an SSL_CTX.
3300 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3301 from the internal representation. Various PKCS#7 fixes: remove some
3302 evil casts and set the enc_dig_alg field properly based on the signing
3306 *) Allow PKCS#12 password to be set from the command line or the
3307 environment. Let 'ca' get its config file name from the environment
3308 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3312 *) Allow certificate policies extension to use an IA5STRING for the
3313 organization field. This is contrary to the PKIX definition but
3314 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3318 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3319 without disallowing inline assembler and the like for non-pedantic builds.
3322 *) Support Borland C++ builder.
3323 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3328 *) SHA-1 cleanups and performance enhancements.
3329 [Andy Polyakov <appro@fy.chalmers.se>]
3331 *) Sparc v8plus assembler for the bignum library.
3332 [Andy Polyakov <appro@fy.chalmers.se>]
3334 *) Accept any -xxx and +xxx compiler options in Configure.
3337 *) Update HPUX configuration.
3340 *) Add missing sk_<type>_unshift() function to safestack.h
3341 [Ralf S. Engelschall]
3343 *) New function SSL_CTX_use_certificate_chain_file that sets the
3344 "extra_cert"s in addition to the certificate. (This makes sense
3345 only for "PEM" format files, as chains as a whole are not
3349 *) Support verify_depth from the SSL API.
3350 x509_vfy.c had what can be considered an off-by-one-error:
3351 Its depth (which was not part of the external interface)
3352 was actually counting the number of certificates in a chain;
3353 now it really counts the depth.
3356 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3357 instead of X509err, which often resulted in confusing error
3358 messages since the error codes are not globally unique
3359 (e.g. an alleged error in ssl3_accept when a certificate
3360 didn't match the private key).
3362 *) New function SSL_CTX_set_session_id_context that allows to set a default
3363 value (so that you don't need SSL_set_session_id_context for each
3364 connection using the SSL_CTX).
3367 *) OAEP decoding bug fix.
3370 *) Support INSTALL_PREFIX for package builders, as proposed by
3374 *) New Configure options "threads" and "no-threads". For systems
3375 where the proper compiler options are known (currently Solaris
3376 and Linux), "threads" is the default.
3379 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3382 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3383 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3384 such as /usr/local/bin.
3387 *) "make linux-shared" to build shared libraries.
3388 [Niels Poppe <niels@netbox.org>]
3390 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3393 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3394 extension adding in x509 utility.
3397 *) Remove NOPROTO sections and error code comments.
3400 *) Partial rewrite of the DEF file generator to now parse the ANSI
3404 *) New Configure options --prefix=DIR and --openssldir=DIR.
3407 *) Complete rewrite of the error code script(s). It is all now handled
3408 by one script at the top level which handles error code gathering,
3409 header rewriting and C source file generation. It should be much better
3410 than the old method: it now uses a modified version of Ulf's parser to
3411 read the ANSI prototypes in all header files (thus the old K&R definitions
3412 aren't needed for error creation any more) and do a better job of
3413 translating function codes into names. The old 'ASN1 error code imbedded
3414 in a comment' is no longer necessary and it doesn't use .err files which
3415 have now been deleted. Also the error code call doesn't have to appear all
3416 on one line (which resulted in some large lines...).
3419 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3422 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3423 0 (which usually indicates a closed connection), but continue reading.
3426 *) Fix some race conditions.
3429 *) Add support for CRL distribution points extension. Add Certificate
3430 Policies and CRL distribution points documentation.
3433 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3436 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3437 8 of keying material. Merlin has also confirmed interop with this fix
3438 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3439 [Merlin Hughes <merlin@baltimore.ie>]
3441 *) Fix lots of warnings.
3442 [Richard Levitte <levitte@stacken.kth.se>]
3444 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3445 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3446 [Richard Levitte <levitte@stacken.kth.se>]
3448 *) Fix problems with sizeof(long) == 8.
3449 [Andy Polyakov <appro@fy.chalmers.se>]
3451 *) Change functions to ANSI C.
3454 *) Fix typos in error codes.
3455 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3457 *) Remove defunct assembler files from Configure.
3460 *) SPARC v8 assembler BIGNUM implementation.
3461 [Andy Polyakov <appro@fy.chalmers.se>]
3463 *) Support for Certificate Policies extension: both print and set.
3464 Various additions to support the r2i method this uses.
3467 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3468 return a const string when you are expecting an allocated buffer.
3471 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3472 types DirectoryString and DisplayText.
3475 *) Add code to allow r2i extensions to access the configuration database,
3476 add an LHASH database driver and add several ctx helper functions.
3479 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3480 fail when they extended the size of a BIGNUM.
3483 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3484 support typesafe stack.
3487 *) Fix typo in SSL_[gs]et_options().
3488 [Nils Frostberg <nils@medcom.se>]
3490 *) Delete various functions and files that belonged to the (now obsolete)
3491 old X509V3 handling code.
3494 *) New Configure option "rsaref".
3497 *) Don't auto-generate pem.h.
3500 *) Introduce type-safe ASN.1 SETs.
3503 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3504 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3506 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3507 that links with OpenSSL (well at least cause lots of warnings), but fear
3508 not: the conversion is trivial, and it eliminates loads of evil casts. A
3509 few STACKed things have been converted already. Feel free to convert more.
3510 In the fullness of time, I'll do away with the STACK type altogether.
3513 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3514 specified in <certfile> by updating the entry in the index.txt file.
3515 This way one no longer has to edit the index.txt file manually for
3516 revoking a certificate. The -revoke option does the gory details now.
3517 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3519 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3520 `-text' option at all and this way the `-noout -text' combination was
3521 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3522 [Ralf S. Engelschall]
3524 *) Make sure a corresponding plain text error message exists for the
3525 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3526 verify callback function determined that a certificate was revoked.
3527 [Ralf S. Engelschall]
3529 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3530 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3531 all available cipers including rc5, which was forgotten until now.
3532 In order to let the testing shell script know which algorithms
3533 are available, a new (up to now undocumented) command
3534 "openssl list-cipher-commands" is used.
3537 *) Bugfix: s_client occasionally would sleep in select() when
3538 it should have checked SSL_pending() first.
3541 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3542 the raw DSA values prior to ASN.1 encoding.
3545 *) Tweaks to Configure
3546 [Niels Poppe <niels@netbox.org>]
3548 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3552 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3555 *) New config option to avoid instructions that are illegal on the 80386.
3556 The default code is faster, but requires at least a 486.
3559 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3560 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3561 same as SSL2_VERSION anyway.
3564 *) New "-showcerts" option for s_client.
3567 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3568 application. Various cleanups and fixes.
3571 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3572 modify error routines to work internally. Add error codes and PBE init
3573 to library startup routines.
3576 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3577 packing functions to asn1 and evp. Changed function names and error
3578 codes along the way.
3581 *) PKCS12 integration: and so it begins... First of several patches to
projects / openssl.git / blob