4 Changes between 0.9.6 and 0.9.7 [xx XXX 2000]
6 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7 due to incorrect handling of multi-threading:
9 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
11 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13 3. Count how many times MemCheck_off() has been called so that
14 nested use can be treated correctly. This also avoids
15 inband-signalling in the previous code (which relied on the
16 assumption that thread ID 0 is impossible).
19 *) New options to 'ca' utility to support V2 CRL entry extensions.
20 Currently CRL reason, invalidity date and hold instruction are
21 supported. Add new CRL extensions to V3 code and some new objects.
24 *) Add "-rand" option also to s_client and s_server.
27 *) New function EVP_CIPHER_CTX_set_padding() this is used to
28 disable standard block padding (aka PKCS#5 padding) in the EVP
29 API, which was previously mandatory. This means that the data is
30 not padded in any way and so the total length much be a multiple
31 of the block size, otherwise an error occurs.
34 *) Initial (incomplete) OCSP SSL support.
37 *) Fix CPU detection on Irix 6.x.
38 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
39 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
41 *) New function OCSP_parse_url(). This splits up a URL into its host,
42 port and path components: primarily to parse OCSP URLs. New -url
43 option to ocsp utility.
46 *) New nonce behavior. The return value of OCSP_check_nonce() now
47 reflects the various checks performed. Applications can decide
48 whether to tolerate certain situations such as an absent nonce
49 in a response when one was present in a request: the ocsp application
50 just prints out a warning. New function OCSP_add1_basic_nonce()
51 this is to allow responders to include a nonce in a response even if
52 the request is nonce-less.
55 *) Use the cached encoding of an X509_NAME structure rather than
56 copying it. This is apparently the reason for the libsafe "errors"
57 but the code is actually correct.
60 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
61 skipped when using openssl x509 multiple times on a single input file,
62 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
65 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
66 set string type: to handle setting ASN1_TIME structures. Fix ca
67 utility to correctly initialize revocation date of CRLs.
70 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
71 the clients preferred ciphersuites and rather use its own preferences.
72 Should help to work around M$ SGC (Server Gated Cryptography) bug in
73 Internet Explorer by ensuring unchanged hash method during stepup.
76 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
77 to aes and add a new 'exist' option to print out symbols that don't
81 *) Additional options to ocsp utility to allow flags to be set and
82 additional certificates supplied.
85 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
86 OCSP client a number of certificate to only verify the response
90 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
91 Bleichenbacher's DSA attack.
92 [Ulf Moeller, Bodo Moeller]
94 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
95 handle the new API. Currently only ECB, CBC modes supported. Add new
96 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
97 for TLS" draft-ietf-tls-ciphersuite-03.txt.
98 [Ben Laurie, Steve Henson]
100 *) In the NCONF_...-based implementations for CONF_... queries
101 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
102 a temporary CONF structure with the data component set to NULL
103 (which gives segmentation faults in lh_retrieve).
104 Instead, use NULL for the CONF pointer in CONF_get_string and
105 CONF_get_number (which may use environment variables) and directly
106 return NULL from CONF_get_section.
109 *) Fix potential buffer overrun for EBCDIC.
112 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
116 *) Functions for OCSP responders. OCSP_request_onereq_count(),
117 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
118 extract information from a certificate request. OCSP_response_create()
119 creates a response and optionally adds a basic response structure.
120 OCSP_basic_add1_status() adds a complete single response to a basic
121 reponse and returns the OCSP_SINGLERESP structure just added (to allow
122 extensions to be included for example). OCSP_basic_add1_cert() adds a
123 certificate to a basic response and OCSP_basic_sign() signs a basic
124 response with various flags. New helper functions ASN1_TIME_check()
125 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
126 (converts ASN1_TIME to GeneralizedTime).
129 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
130 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
131 structure from a certificate. X509_pubkey_digest() digests tha public_key
132 contents: this is used in various key identifiers.
135 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
136 keyUsage if basicConstraints absent for a CA.
139 *) Make SMIME_write_PKCS7() write mail header values with a format that
140 is more generally accepted (no spaces before the semicolon), since
141 some programs can't parse those values properly otherwise. Also make
142 sure BIO's that break lines after each write do not create invalid
146 *) Make sk_sort() tolerate a NULL argument.
147 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
149 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
150 passed by the function are trusted implicitly. If any of them signed the
151 reponse then it is assumed to be valid and is not verified.
154 *) Zero the premaster secret after deriving the master secret in
158 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
159 to data. This was previously part of the PKCS7 ASN1 code. This
160 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
161 [Steve Henson, reported by Kenneth R. Robinette
162 <support@securenetterm.com>]
164 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
165 routines: without these tracing memory leaks is very painful.
166 Fix leaks in PKCS12 and PKCS7 routines.
169 *) Fix for Irix with NO_ASM.
170 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
172 *) Add some EVP_add_digest_alias registrations (as found in
173 OpenSSL_add_all_digests()), to SSL_library_init()
174 aka OpenSSL_add_ssl_algorithms(). This provides improved
175 compatibility with peers using X.509 certificates
176 with unconventional AlgorithmIdentifier OIDs.
179 *) ./config script fixes.
180 [Ulf Moeller, Richard Levitte]
182 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
183 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
184 effectively meant GeneralizedTime would never be used. Now it
185 is initialised to -1 but X509_time_adj() now has to check the value
186 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
187 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
188 [Steve Henson, reported by Kenneth R. Robinette
189 <support@securenetterm.com>]
191 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
192 result in a zero length in the ASN1_INTEGER structure which was
193 not consistent with the structure when d2i_ASN1_INTEGER() was used
194 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
195 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
196 where it did not print out a minus for negative ASN1_INTEGER.
199 *) Fix 'openssl passwd -1'.
202 *) Add summary printout to ocsp utility. The various functions which
203 convert status values to strings have been renamed to:
204 OCSP_response_status_str(), OCSP_cert_status_str() and
205 OCSP_crl_reason_str() and are no longer static. New options
206 to verify nonce values and to disable verification. OCSP response
207 printout format cleaned up.
210 *) Add additional OCSP certificate checks. These are those specified
211 in RFC2560. This consists of two separate checks: the CA of the
212 certificate being checked must either be the OCSP signer certificate
213 or the issuer of the OCSP signer certificate. In the latter case the
214 OCSP signer certificate must contain the OCSP signing extended key
215 usage. This check is performed by attempting to match the OCSP
216 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
217 in the OCSP_CERTID structures of the response.
220 *) Initial OCSP certificate verification added to OCSP_basic_verify()
221 and related routines. This uses the standard OpenSSL certificate
222 verify routines to perform initial checks (just CA validity) and
223 to obtain the certificate chain. Then additional checks will be
224 performed on the chain. Currently the root CA is checked to see
225 if it is explicitly trusted for OCSP signing. This is used to set
226 a root CA as a global signing root: that is any certificate that
227 chains to that CA is an acceptable OCSP signing certificate.
230 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
231 extensions from a separate configuration file.
232 As when reading extensions from the main configuration file,
233 the '-extensions ...' option may be used for specifying the
235 [Massimiliano Pala <madwolf@comune.modena.it>]
237 *) Change PKCS12_key_gen_asc() so it can cope with non null
238 terminated strings whose length is passed in the passlen
239 parameter, for example from PEM callbacks. This was done
240 by adding an extra length parameter to asc2uni().
241 [Steve Henson, reported by <oddissey@samsung.co.kr>]
243 *) New OCSP utility. Allows OCSP requests to be generated or
244 read. The request can be sent to a responder and the output
245 parsed, outputed or printed in text form. Not complete yet:
246 still needs to check the OCSP response validity.
249 *) New subcommands for 'openssl ca':
250 'openssl ca -status <serial>' prints the status of the cert with
251 the given serial number (according to the index file).
252 'openssl ca -updatedb' updates the expiry status of certificates
254 [Massimiliano Pala <madwolf@comune.modena.it>]
256 *) New '-newreq-nodes' command option to CA.pl. This is like
257 '-newreq', but calls 'openssl req' with the '-nodes' option
258 so that the resulting key is not encrypted.
259 [Damien Miller <djm@mindrot.org>]
261 *) New configuration for the GNU Hurd.
262 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
264 *) Initial code to implement OCSP basic response verify. This
265 is currently incomplete. Currently just finds the signer's
266 certificate and verifies the signature on the response.
269 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
270 value of OPENSSLDIR. This is available via the new '-d' option
271 to 'openssl version', and is also included in 'openssl version -a'.
274 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
275 call failed, free the DSA structure.
278 *) Allowing defining memory allocation callbacks that will be given
279 file name and line number information in additional arguments
280 (a const char* and an int). The basic functionality remains, as
281 well as the original possibility to just replace malloc(),
282 realloc() and free() by functions that do not know about these
283 additional arguments. To register and find out the current
284 settings for extended allocation functions, the following
285 functions are provided:
287 CRYPTO_set_mem_ex_functions
288 CRYPTO_set_locked_mem_ex_functions
289 CRYPTO_get_mem_ex_functions
290 CRYPTO_get_locked_mem_ex_functions
292 These work the same way as CRYPTO_set_mem_functions and friends.
293 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
294 extended allocation function is enabled.
295 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
296 a conventional allocation function is enabled.
297 [Richard Levitte, Bodo Moeller]
299 *) Fix to uni2asc() to cope with zero length Unicode strings.
300 These are present in some PKCS#12 files.
303 *) Finish off removing the remaining LHASH function pointer casts.
304 There should no longer be any prototype-casting required when using
305 the LHASH abstraction, and any casts that remain are "bugs". See
306 the callback types and macros at the head of lhash.h for details
307 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
310 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
311 If an EGD or PRNGD is running and enough entropy is returned, automatic
312 seeding like with /dev/[u]random will be performed.
313 Positions tried are: /etc/entropy, /var/run/egd-pool.
316 *) Change the Unix RAND_poll() variant to be able to poll several
317 random devices, as specified by DEVRANDOM, until a sufficient amount
318 of data has been collected. We spend at most 10 ms on each file
319 (select timeout) and read in non-blocking mode. DEVRANDOM now
320 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
321 (previously it was just the string "/dev/urandom"), so on typical
322 platforms the 10 ms delay will never occur.
323 Also separate out the Unix variant to its own file, rand_unix.c.
324 For VMS, there's a currently-empty rand_vms.c.
327 *) Move OCSP client related routines to ocsp_cl.c. These
328 provide utility functions which an application needing
329 to issue a request to an OCSP responder and analyse the
330 response will typically need: as opposed to those which an
331 OCSP responder itself would need which will be added later.
333 OCSP_request_sign() signs an OCSP request with an API similar
334 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
335 response. OCSP_response_get1_basic() extracts basic response
336 from response. OCSP_resp_find_status(): finds and extracts status
337 information from an OCSP_CERTID structure (which will be created
338 when the request structure is built). These are built from lower
339 level functions which work on OCSP_SINGLERESP structures but
340 wont normally be used unless the application wishes to examine
341 extensions in the OCSP response for example.
343 Replace nonce routines with a pair of functions.
344 OCSP_request_add1_nonce() adds a nonce value and optionally
345 generates a random value. OCSP_check_nonce() checks the
346 validity of the nonce in an OCSP response.
349 *) Change function OCSP_request_add() to OCSP_request_add0_id().
350 This doesn't copy the supplied OCSP_CERTID and avoids the
351 need to free up the newly created id. Change return type
352 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
353 This can then be used to add extensions to the request.
354 Deleted OCSP_request_new(), since most of its functionality
355 is now in OCSP_REQUEST_new() (and the case insensitive name
356 clash) apart from the ability to set the request name which
357 will be added elsewhere.
360 *) Update OCSP API. Remove obsolete extensions argument from
361 various functions. Extensions are now handled using the new
362 OCSP extension code. New simple OCSP HTTP function which
363 can be used to send requests and parse the response.
366 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
367 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
368 uses the special reorder version of SET OF to sort the attributes
369 and reorder them to match the encoded order. This resolves a long
370 standing problem: a verify on a PKCS7 structure just after signing
371 it used to fail because the attribute order did not match the
372 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
373 it uses the received order. This is necessary to tolerate some broken
374 software that does not order SET OF. This is handled by encoding
375 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
376 to produce the required SET OF.
379 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
380 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
381 files to get correct declarations of the ASN.1 item variables.
384 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
385 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
386 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
387 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
388 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
389 ASN1_ITEM and no wrapper functions.
392 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
393 replace the old function pointer based I/O routines. Change most of
394 the *_d2i_bio() and *_d2i_fp() functions to use these.
397 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
398 lines, recognice more "algorithms" that can be deselected, and make
399 it complain about algorithm deselection that isn't recognised.
402 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
403 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
404 to use new functions. Add NO_ASN1_OLD which can be set to remove
405 some old style ASN1 functions: this can be used to determine if old
406 code will still work when these eventually go away.
409 *) New extension functions for OCSP structures, these follow the
410 same conventions as certificates and CRLs.
413 *) New function X509V3_add1_i2d(). This automatically encodes and
414 adds an extension. Its behaviour can be customised with various
415 flags to append, replace or delete. Various wrappers added for
416 certifcates and CRLs.
419 *) Fix to avoid calling the underlying ASN1 print routine when
420 an extension cannot be parsed. Correct a typo in the
421 OCSP_SERVICELOC extension. Tidy up print OCSP format.
424 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
425 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
426 when writing a 32767 byte record.
427 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
429 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
430 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
432 (RSA objects have a reference count access to which is protected
433 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
434 so they are meant to be shared between threads.)
435 [Bodo Moeller, Geoff Thorpe; original patch submitted by
436 "Reddie, Steven" <Steven.Reddie@ca.com>]
438 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
439 entries for variables.
442 *) Fix a deadlock in CRYPTO_mem_leaks().
445 *) Add functionality to apps/openssl.c for detecting locking
446 problems: As the program is single-threaded, all we have
447 to do is register a locking callback using an array for
448 storing which locks are currently held by the program.
451 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
452 SSL_get_ex_data_X509_STORE_idx(), which is used in
453 ssl_verify_cert_chain() and thus can be called at any time
454 during TLS/SSL handshakes so that thread-safety is essential.
455 Unfortunately, the ex_data design is not at all suited
456 for multi-threaded use, so it probably should be abolished.
459 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
460 [Broadcom, tweaked and integrated by Geoff Thorpe]
462 *) Move common extension printing code to new function
463 X509V3_print_extensions(). Reorganise OCSP print routines and
464 implement some needed OCSP ASN1 functions. Add OCSP extensions.
467 *) New function X509_signature_print() to remove duplication in some
471 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
472 set (this was treated exactly the same as SET OF previously). This
473 is used to reorder the STACK representing the structure to match the
474 encoding. This will be used to get round a problem where a PKCS7
475 structure which was signed could not be verified because the STACK
476 order did not reflect the encoded order.
479 *) Reimplement the OCSP ASN1 module using the new code.
482 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
483 for its ASN1 operations. The old style function pointers still exist
484 for now but they will eventually go away.
487 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
488 completely replaces the old ASN1 functionality with a table driven
489 encoder and decoder which interprets an ASN1_ITEM structure describing
490 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
491 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
492 has also been converted to the new form.
495 *) Change BN_mod_exp_recp so that negative moduli are tolerated
496 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
497 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
501 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
502 of not touching the result's sign bit.
505 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
509 *) Changed the LHASH code to use prototypes for callbacks, and created
510 macros to declare and implement thin (optionally static) functions
511 that provide type-safety and avoid function pointer casting for the
512 type-specific callbacks.
515 *) Use better test patterns in bntest.
518 *) Added Kerberos Cipher Suites to be used with TLS, as written in
520 [Veers Staats <staatsvr@asc.hpc.mil>,
521 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
523 *) rand_win.c fix for Borland C.
526 *) BN_rshift bugfix for n == 0.
529 *) Reformat the FAQ so the different questions and answers can be divided
530 in sections depending on the subject.
533 *) Have the zlib compression code load ZLIB.DLL dynamically under
537 *) New function BN_mod_sqrt for computing square roots modulo a prime
538 (using the probabilistic Tonelli-Shanks algorithm unless
539 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
540 be handled deterministically).
541 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
543 *) Store verify_result within SSL_SESSION also for client side to
544 avoid potential security hole. (Re-used sessions on the client side
545 always resulted in verify_result==X509_V_OK, not using the original
546 result of the server certificate verification.)
549 *) Make BN_mod_inverse faster by explicitly handling small quotients
550 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
551 512 bits], about 30% for larger ones [1024 or 2048 bits].)
554 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
555 SSL3_RT_APPLICATION_DATA, return 0.
556 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
560 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
561 releases, have been re-implemented by renaming the previous
562 implementations of ssl2_read and ssl3_read to ssl2_read_internal
563 and ssl3_read_internal, respectively, and adding 'peek' parameters
564 to them. The new ssl[23]_{read,peek} functions are calls to
565 ssl[23]_read_internal with the 'peek' flag set appropriately.
566 A 'peek' parameter has also been added to ssl3_read_bytes, which
567 does the actual work for ssl3_read_internal.
570 *) New function BN_kronecker.
573 *) Fix BN_gcd so that it works on negative inputs; the result is
574 positive unless both parameters are zero.
575 Previously something reasonably close to an infinite loop was
576 possible because numbers could be growing instead of shrinking
577 in the implementation of Euclid's algorithm.
580 *) Fix BN_is_word() and BN_is_one() macros to take into account the
581 sign of the number in question.
583 Fix BN_is_word(a,w) to work correctly for w == 0.
585 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
586 because its test if the absolute value of 'a' equals 'w'.
587 Note that BN_abs_is_word does *not* handle w == 0 reliably;
588 it exists mostly for use in the implementations of BN_is_zero(),
589 BN_is_one(), and BN_is_word().
592 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
593 the method-specific "init()" handler. Also clean up ex_data after
594 calling the method-specific "finish()" handler. Previously, this was
595 happening the other way round.
598 *) New function BN_swap.
601 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
602 the exponentiation functions are more likely to produce reasonable
603 results on negative inputs.
606 *) Change BN_mod_mul so that the result is always non-negative.
607 Previously, it could be negative if one of the factors was negative;
608 I don't think anyone really wanted that behaviour.
611 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
612 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
613 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
614 and add new functions:
627 These functions always generate non-negative results.
629 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
630 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
632 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
633 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
635 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
637 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
638 was actually never needed) and in BN_mul(). The removal in BN_mul()
639 required a small change in bn_mul_part_recursive() and the addition
640 of the functions bn_cmp_part_words(), bn_sub_part_words() and
641 bn_add_part_words(), which do the same thing as bn_cmp_words(),
642 bn_sub_words() and bn_add_words() except they take arrays with
646 *) In 'openssl passwd', verify passwords read from the terminal
647 unless the '-salt' option is used (which usually means that
648 verification would just waste user's time since the resulting
649 hash is going to be compared with some given password hash)
650 or the new '-noverify' option is used.
652 This is an incompatible change, but it does not affect
653 non-interactive use of 'openssl passwd' (passwords on the command
654 line, '-stdin' option, '-in ...' option) and thus should not
658 *) Remove all references to RSAref, since there's no more need for it.
661 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
662 The previous value, 12, was not always sufficient for BN_mod_exp().
665 *) Make DSO load along a path given through an environment variable
666 (SHLIB_PATH) with shl_load().
669 *) Constify the ENGINE code as a result of BIGNUM constification.
670 Also constify the RSA code and most things related to it. In a
671 few places, most notable in the depth of the ASN.1 code, ugly
672 casts back to non-const were required (to be solved at a later
676 *) Make it so the openssl application has all engines loaded by default.
679 *) Constify the BIGNUM routines a little more.
682 *) Make sure that shared libraries get the internal name engine with
683 the full version number and not just 0. This should mark the
684 shared libraries as not backward compatible. Of course, this should
685 be changed again when we can guarantee backward binary compatibility.
688 *) Add the following functions:
694 ENGINE_load_builtin_engines()
696 That way, an application can itself choose if external engines that
697 are built-in in OpenSSL shall ever be used or not. The benefit is
698 that applications won't have to be linked with libdl or other dso
699 libraries unless it's really needed.
701 Changed 'openssl engine' to load all engines on demand.
702 Changed the engine header files to avoid the duplication of some
703 declarations (they differed!).
706 *) 'openssl engine' can now list capabilities.
709 *) Better error reporting in 'openssl engine'.
712 *) Never call load_dh_param(NULL) in s_server.
715 *) Add engine application. It can currently list engines by name and
716 identity, and test if they are actually available.
719 *) Add support for shared libraries under Irix.
720 [Albert Chin-A-Young <china@thewrittenword.com>]
722 *) Improve RPM specification file by forcing symbolic linking and making
723 sure the installed documentation is also owned by root.root.
724 [Damien Miller <djm@mindrot.org>]
726 *) Add configuration option to build on Linux on both big-endian and
728 [Ralf Baechle <ralf@uni-koblenz.de>]
730 *) Give the OpenSSL applications more possibilities to make use of
731 keys (public as well as private) handled by engines.
734 *) Add OCSP code that comes from CertCo.
737 *) Add VMS support for the Rijndael code.
740 *) Added untested support for Nuron crypto accelerator.
743 *) Add support for external cryptographic devices. This code was
744 previously distributed separately as the "engine" branch.
745 [Geoff Thorpe, Richard Levitte]
747 *) Rework the filename-translation in the DSO code. It is now possible to
748 have far greater control over how a "name" is turned into a filename
749 depending on the operating environment and any oddities about the
750 different shared library filenames on each system.
753 *) Support threads on FreeBSD-elf in Configure.
756 *) Add the possibility to create shared libraries on HP-UX
759 *) Fix for SHA1 assembly problem with MASM: it produces
760 warnings about corrupt line number information when assembling
761 with debugging information. This is caused by the overlapping
763 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
766 NCONF_get_number() has no error checking at all. As a replacement,
767 NCONF_get_number_e() is defined (_e for "error checking") and is
768 promoted strongly. The old NCONF_get_number is kept around for
769 binary backward compatibility.
770 Make it possible for methods to load from something other than a BIO,
771 by providing a function pointer that is given a name instead of a BIO.
772 For example, this could be used to load configuration data from an
776 *) Fix typo in get_cert_by_subject() in by_dir.c
777 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
779 *) Rework the system to generate shared libraries:
781 - Make note of the expected extension for the shared libraries and
782 if there is a need for symbolic links from for example libcrypto.so.0
783 to libcrypto.so.0.9.7. There is extended info in Configure for
786 - Make as few rebuilds of the shared libraries as possible.
788 - Still avoid linking the OpenSSL programs with the shared libraries.
790 - When installing, install the shared libraries separately from the
794 *) Fix for non blocking accept BIOs. Added new I/O special reason
795 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
796 with non blocking I/O was not possible because no retry code was
797 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
801 *) Added the beginnings of Rijndael support.
804 *) Fix for bug in DirectoryString mask setting. Add support for
805 X509_NAME_print_ex() in 'req' and X509_print_ex() function
806 to allow certificate printing to more controllable, additional
807 'certopt' option to 'x509' to allow new printing options to be
811 *) Clean old EAY MD5 hack from e_os.h.
814 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
816 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
817 and not in SSL_clear because the latter is also used by the
818 accept/connect functions; previously, the settings made by
819 SSL_set_read_ahead would be lost during the handshake.
820 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
822 *) Correct util/mkdef.pl to be selective about disabled algorithms.
823 Previously, it would create entries for disableed algorithms no
827 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
829 *) In ssl23_get_client_hello, generate an error message when faced
830 with an initial SSL 3.0/TLS record that is too small to contain the
831 first two bytes of the ClientHello message, i.e. client_version.
832 (Note that this is a pathologic case that probably has never happened
833 in real life.) The previous approach was to use the version number
834 from the record header as a substitute; but our protocol choice
835 should not depend on that one because it is not authenticated
836 by the Finished messages.
839 *) More robust randomness gathering functions for Windows.
840 [Jeffrey Altman <jaltman@columbia.edu>]
842 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
843 not set then we don't setup the error code for issuer check errors
844 to avoid possibly overwriting other errors which the callback does
845 handle. If an application does set the flag then we assume it knows
846 what it is doing and can handle the new informational codes
850 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
851 a general "ANY" type, as such it should be able to decode anything
852 including tagged types. However it didn't check the class so it would
853 wrongly interpret tagged types in the same way as their universal
854 counterpart and unknown types were just rejected. Changed so that the
855 tagged and unknown types are handled in the same way as a SEQUENCE:
856 that is the encoding is stored intact. There is also a new type
857 "V_ASN1_OTHER" which is used when the class is not universal, in this
858 case we have no idea what the actual type is so we just lump them all
862 *) On VMS, stdout may very well lead to a file that is written to
863 in a record-oriented fashion. That means that every write() will
864 write a separate record, which will be read separately by the
865 programs trying to read from it. This can be very confusing.
867 The solution is to put a BIO filter in the way that will buffer
868 text until a linefeed is reached, and then write everything a
869 line at a time, so every record written will be an actual line,
870 not chunks of lines and not (usually doesn't happen, but I've
871 seen it once) several lines in one record. BIO_f_linebuffer() is
874 Currently, it's a VMS-only method, because that's where it has
875 been tested well enough.
878 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
879 it can return incorrect results.
880 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
881 but it was in 0.9.6-beta[12].)
884 *) Disable the check for content being present when verifying detached
885 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
886 include zero length content when signing messages.
889 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
890 BIO_ctrl (for BIO pairs).
893 *) Add DSO method for VMS.
896 *) Bug fix: Montgomery multiplication could produce results with the
900 *) Add RPM specification openssl.spec and modify it to build three
901 packages. The default package contains applications, application
902 documentation and run-time libraries. The devel package contains
903 include files, static libraries and function documentation. The
904 doc package contains the contents of the doc directory. The original
905 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
908 *) Add a large number of documentation files for many SSL routines.
909 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
911 *) Add a configuration entry for Sony News 4.
912 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
914 *) Don't set the two most significant bits to one when generating a
915 random number < q in the DSA library.
918 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
919 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
920 the underlying transport is blocking) if a handshake took place.
921 (The default behaviour is needed by applications such as s_client
922 and s_server that use select() to determine when to use SSL_read;
923 but for applications that know in advance when to expect data, it
924 just makes things more complicated.)
927 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
931 *) Add a few more EBCDIC conditionals that make `req' and `x509'
932 work better on such systems.
933 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
935 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
936 Update PKCS12_parse() so it copies the friendlyName and the
937 keyid to the certificates aux info.
940 *) Fix bug in PKCS7_verify() which caused an infinite loop
941 if there was more than one signature.
942 [Sven Uszpelkat <su@celocom.de>]
944 *) Major change in util/mkdef.pl to include extra information
945 about each symbol, as well as presentig variables as well
946 as functions. This change means that there's n more need
947 to rebuild the .num files when some algorithms are excluded.
950 *) Allow the verify time to be set by an application,
951 rather than always using the current time.
954 *) Phase 2 verify code reorganisation. The certificate
955 verify code now looks up an issuer certificate by a
956 number of criteria: subject name, authority key id
957 and key usage. It also verifies self signed certificates
958 by the same criteria. The main comparison function is
959 X509_check_issued() which performs these checks.
961 Lot of changes were necessary in order to support this
962 without completely rewriting the lookup code.
964 Authority and subject key identifier are now cached.
966 The LHASH 'certs' is X509_STORE has now been replaced
967 by a STACK_OF(X509_OBJECT). This is mainly because an
968 LHASH can't store or retrieve multiple objects with
971 As a result various functions (which were all internal
972 use only) have changed to handle the new X509_STORE
973 structure. This will break anything that messed round
974 with X509_STORE internally.
976 The functions X509_STORE_add_cert() now checks for an
977 exact match, rather than just subject name.
979 The X509_STORE API doesn't directly support the retrieval
980 of multiple certificates matching a given criteria, however
981 this can be worked round by performing a lookup first
982 (which will fill the cache with candidate certificates)
983 and then examining the cache for matches. This is probably
984 the best we can do without throwing out X509_LOOKUP
985 entirely (maybe later...).
987 The X509_VERIFY_CTX structure has been enhanced considerably.
989 All certificate lookup operations now go via a get_issuer()
990 callback. Although this currently uses an X509_STORE it
991 can be replaced by custom lookups. This is a simple way
992 to bypass the X509_STORE hackery necessary to make this
993 work and makes it possible to use more efficient techniques
994 in future. A very simple version which uses a simple
995 STACK for its trusted certificate store is also provided
996 using X509_STORE_CTX_trusted_stack().
998 The verify_cb() and verify() callbacks now have equivalents
999 in the X509_STORE_CTX structure.
1001 X509_STORE_CTX also has a 'flags' field which can be used
1002 to customise the verify behaviour.
1005 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1006 excludes S/MIME capabilities.
1009 *) When a certificate request is read in keep a copy of the
1010 original encoding of the signed data and use it when outputing
1011 again. Signatures then use the original encoding rather than
1012 a decoded, encoded version which may cause problems if the
1013 request is improperly encoded.
1016 *) For consistency with other BIO_puts implementations, call
1017 buffer_write(b, ...) directly in buffer_puts instead of calling
1020 In BIO_puts, increment b->num_write as in BIO_write.
1021 [Peter.Sylvester@EdelWeb.fr]
1023 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1024 BN_zero, we may not return a BIGNUM with an array consisting of
1028 *) Avoid calling abort() from within the library when problems are
1029 detected, except if preprocessor symbols have been defined
1030 (such as REF_CHECK, BN_DEBUG etc.).
1033 *) New openssl application 'rsautl'. This utility can be
1034 used for low level RSA operations. DER public key
1035 BIO/fp routines also added.
1038 *) New Configure entry and patches for compiling on QNX 4.
1039 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1041 *) A demo state-machine implementation was sponsored by
1042 Nuron (http://www.nuron.com/) and is now available in
1043 demos/state_machine.
1046 *) New options added to the 'dgst' utility for signature
1047 generation and verification.
1050 *) Unrecognized PKCS#7 content types are now handled via a
1051 catch all ASN1_TYPE structure. This allows unsupported
1052 types to be stored as a "blob" and an application can
1053 encode and decode it manually.
1056 *) Fix various signed/unsigned issues to make a_strex.c
1058 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1060 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1061 length if passed a buffer. ASN1_INTEGER_to_BN failed
1062 if passed a NULL BN and its argument was negative.
1063 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1065 *) Modification to PKCS#7 encoding routines to output definite
1066 length encoding. Since currently the whole structures are in
1067 memory there's not real point in using indefinite length
1068 constructed encoding. However if OpenSSL is compiled with
1069 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1072 *) Added BIO_vprintf() and BIO_vsnprintf().
1075 *) Added more prefixes to parse for in the the strings written
1076 through a logging bio, to cover all the levels that are available
1077 through syslog. The prefixes are now:
1079 PANIC, EMERG, EMR => LOG_EMERG
1080 ALERT, ALR => LOG_ALERT
1081 CRIT, CRI => LOG_CRIT
1082 ERROR, ERR => LOG_ERR
1083 WARNING, WARN, WAR => LOG_WARNING
1084 NOTICE, NOTE, NOT => LOG_NOTICE
1085 INFO, INF => LOG_INFO
1086 DEBUG, DBG => LOG_DEBUG
1088 and as before, if none of those prefixes are present at the
1089 beginning of the string, LOG_ERR is chosen.
1091 On Win32, the LOG_* levels are mapped according to this:
1093 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1094 LOG_WARNING => EVENTLOG_WARNING_TYPE
1095 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1099 *) Made it possible to reconfigure with just the configuration
1100 argument "reconf" or "reconfigure". The command line arguments
1101 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1102 and are retrieved from there when reconfiguring.
1106 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1108 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1111 *) The obj_dat.pl script was messing up the sorting of object
1112 names. The reason was that it compared the quoted version
1113 of strings as a result "OCSP" > "OCSP Signing" because
1114 " > SPACE. Changed script to store unquoted versions of
1115 names and add quotes on output. It was also omitting some
1116 names from the lookup table if they were given a default
1117 value (that is if SN is missing it is given the same
1118 value as LN and vice versa), these are now added on the
1119 grounds that if an object has a name we should be able to
1120 look it up. Finally added warning output when duplicate
1121 short or long names are found.
1124 *) Changes needed for Tandem NSK.
1125 [Scott Uroff <scott@xypro.com>]
1127 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1128 RSA_padding_check_SSLv23(), special padding was never detected
1129 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1130 version rollback attacks was not effective.
1132 In s23_clnt.c, don't use special rollback-attack detection padding
1133 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1134 client; similarly, in s23_srvr.c, don't do the rollback check if
1135 SSL 2.0 is the only protocol enabled in the server.
1138 *) Make it possible to get hexdumps of unprintable data with 'openssl
1139 asn1parse'. By implication, the functions ASN1_parse_dump() and
1140 BIO_dump_indent() are added.
1143 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1144 these print out strings and name structures based on various
1145 flags including RFC2253 support and proper handling of
1146 multibyte characters. Added options to the 'x509' utility
1147 to allow the various flags to be set.
1150 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1151 Also change the functions X509_cmp_current_time() and
1152 X509_gmtime_adj() work with an ASN1_TIME structure,
1153 this will enable certificates using GeneralizedTime in validity
1154 dates to be checked.
1157 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1158 negative public key encodings) on by default,
1159 NO_NEG_PUBKEY_BUG can be set to disable it.
1162 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1163 content octets. An i2c_ASN1_OBJECT is unnecessary because
1164 the encoding can be trivially obtained from the structure.
1167 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1168 not read locks (CRYPTO_r_[un]lock).
1171 *) A first attempt at creating official support for shared
1172 libraries through configuration. I've kept it so the
1173 default is static libraries only, and the OpenSSL programs
1174 are always statically linked for now, but there are
1175 preparations for dynamic linking in place.
1176 This has been tested on Linux and True64.
1179 *) Randomness polling function for Win9x, as described in:
1180 Peter Gutmann, Software Generation of Practically Strong
1184 *) Fix so PRNG is seeded in req if using an already existing
1188 *) New options to smime application. -inform and -outform
1189 allow alternative formats for the S/MIME message including
1190 PEM and DER. The -content option allows the content to be
1191 specified separately. This should allow things like Netscape
1192 form signing output easier to verify.
1195 *) Fix the ASN1 encoding of tags using the 'long form'.
1198 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1199 STRING types. These convert content octets to and from the
1200 underlying type. The actual tag and length octets are
1201 already assumed to have been read in and checked. These
1202 are needed because all other string types have virtually
1203 identical handling apart from the tag. By having versions
1204 of the ASN1 functions that just operate on content octets
1205 IMPLICIT tagging can be handled properly. It also allows
1206 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1207 and ASN1_INTEGER are identical apart from the tag.
1210 *) Change the handling of OID objects as follows:
1212 - New object identifiers are inserted in objects.txt, following
1213 the syntax given in objects.README.
1214 - objects.pl is used to process obj_mac.num and create a new
1216 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1219 This is currently kind of a hack, and the perl code in objects.pl
1220 isn't very elegant, but it works as I intended. The simplest way
1221 to check that it worked correctly is to look in obj_dat.h and
1222 check the array nid_objs and make sure the objects haven't moved
1223 around (this is important!). Additions are OK, as well as
1224 consistent name changes.
1227 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1230 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1231 The given file adds to whatever has already been seeded into the
1232 random pool through the RANDFILE configuration file option or
1233 environment variable, or the default random state file.
1236 *) mkstack.pl now sorts each macro group into lexical order.
1237 Previously the output order depended on the order the files
1238 appeared in the directory, resulting in needless rewriting
1242 *) Patches to make OpenSSL compile under Win32 again. Mostly
1243 work arounds for the VC++ problem that it treats func() as
1244 func(void). Also stripped out the parts of mkdef.pl that
1245 added extra typesafe functions: these no longer exist.
1248 *) Reorganisation of the stack code. The macros are now all
1249 collected in safestack.h . Each macro is defined in terms of
1250 a "stack macro" of the form SKM_<name>(type, a, b). The
1251 DEBUG_SAFESTACK is now handled in terms of function casts,
1252 this has the advantage of retaining type safety without the
1253 use of additional functions. If DEBUG_SAFESTACK is not defined
1254 then the non typesafe macros are used instead. Also modified the
1255 mkstack.pl script to handle the new form. Needs testing to see
1256 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1257 the default if no major problems. Similar behaviour for ASN1_SET_OF
1258 and PKCS12_STACK_OF.
1261 *) When some versions of IIS use the 'NET' form of private key the
1262 key derivation algorithm is different. Normally MD5(password) is
1263 used as a 128 bit RC4 key. In the modified case
1264 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1265 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1266 as the old Netscape_RSA functions except they have an additional
1267 'sgckey' parameter which uses the modified algorithm. Also added
1268 an -sgckey command line option to the rsa utility. Thanks to
1269 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1270 algorithm to openssl-dev.
1273 *) The evp_local.h macros were using 'c.##kname' which resulted in
1274 invalid expansion on some systems (SCO 5.0.5 for example).
1275 Corrected to 'c.kname'.
1276 [Phillip Porch <root@theporch.com>]
1278 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1279 a STACK of email addresses from a certificate or request, these look
1280 in the subject name and the subject alternative name extensions and
1281 omit any duplicate addresses.
1284 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1285 This makes DSA verification about 2 % faster.
1288 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1289 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1290 plus overhead for 1024 bit moduli).
1291 This makes exponentiations about 0.5 % faster for 1024 bit
1292 exponents (as measured by "openssl speed rsa2048").
1295 *) Rename memory handling macros to avoid conflicts with other
1297 Malloc => OPENSSL_malloc
1298 Malloc_locked => OPENSSL_malloc_locked
1299 Realloc => OPENSSL_realloc
1300 Free => OPENSSL_free
1303 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1304 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1307 *) CygWin32 support.
1308 [John Jarvie <jjarvie@newsguy.com>]
1310 *) The type-safe stack code has been rejigged. It is now only compiled
1311 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1312 by default all type-specific stack functions are "#define"d back to
1313 standard stack functions. This results in more streamlined output
1314 but retains the type-safety checking possibilities of the original
1318 *) The STACK code has been cleaned up, and certain type declarations
1319 that didn't make a lot of sense have been brought in line. This has
1320 also involved a cleanup of sorts in safestack.h to more correctly
1321 map type-safe stack functions onto their plain stack counterparts.
1322 This work has also resulted in a variety of "const"ifications of
1323 lots of the code, especially "_cmp" operations which should normally
1324 be prototyped with "const" parameters anyway.
1327 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1328 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1329 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1330 where all of 'md' is used each time the PRNG is used, but 'state'
1331 is used only indexed by a cyclic counter. As entropy may not be
1332 well distributed from the beginning, 'md' is important as a
1333 chaining variable. However, the output function chains only half
1334 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1335 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1336 in all of 'state' being rewritten, with the new values depending
1337 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1340 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1341 the handshake is continued after ssl_verify_cert_chain();
1342 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1343 can lead to 'unexplainable' connection aborts later.
1344 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1346 *) Major EVP API cipher revision.
1347 Add hooks for extra EVP features. This allows various cipher
1348 parameters to be set in the EVP interface. Support added for variable
1349 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1350 setting of RC2 and RC5 parameters.
1352 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1355 Remove lots of duplicated code from the EVP library. For example *every*
1356 cipher init() function handles the 'iv' in the same way according to the
1357 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1358 for CFB and OFB modes they zero ctx->num.
1360 New functionality allows removal of S/MIME code RC2 hack.
1362 Most of the routines have the same form and so can be declared in terms
1365 By shifting this to the top level EVP_CipherInit() it can be removed from
1366 all individual ciphers. If the cipher wants to handle IVs or keys
1367 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1370 Change lots of functions like EVP_EncryptUpdate() to now return a
1371 value: although software versions of the algorithms cannot fail
1372 any installed hardware versions can.
1375 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1376 this option is set, tolerate broken clients that send the negotiated
1377 protocol version number instead of the requested protocol version
1381 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1382 i.e. non-zero for export ciphersuites, zero otherwise.
1383 Previous versions had this flag inverted, inconsistent with
1384 rsa_tmp_cb (..._TMP_RSA_CB).
1385 [Bodo Moeller; problem reported by Amit Chopra]
1387 *) Add missing DSA library text string. Work around for some IIS
1388 key files with invalid SEQUENCE encoding.
1391 *) Add a document (doc/standards.txt) that list all kinds of standards
1392 and so on that are implemented in OpenSSL.
1395 *) Enhance c_rehash script. Old version would mishandle certificates
1396 with the same subject name hash and wouldn't handle CRLs at all.
1397 Added -fingerprint option to crl utility, to support new c_rehash
1401 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1404 *) Fix for SSL server purpose checking. Server checking was
1405 rejecting certificates which had extended key usage present
1406 but no ssl client purpose.
1407 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1409 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1410 is a little unclear about how a blank password is handled.
1411 Since the password in encoded as a BMPString with terminating
1412 double NULL a zero length password would end up as just the
1413 double NULL. However no password at all is different and is
1414 handled differently in the PKCS#12 key generation code. NS
1415 treats a blank password as zero length. MSIE treats it as no
1416 password on export: but it will try both on import. We now do
1417 the same: PKCS12_parse() tries zero length and no password if
1418 the password is set to "" or NULL (NULL is now a valid password:
1419 it wasn't before) as does the pkcs12 application.
1422 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1423 perror when PEM_read_bio_X509_REQ fails, the error message must
1424 be obtained from the error queue.
1427 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1428 it in ERR_remove_state if appropriate, and change ERR_get_state
1429 accordingly to avoid race conditions (this is necessary because
1430 thread_hash is no longer constant once set).
1433 *) Bugfix for linux-elf makefile.one.
1436 *) RSA_get_default_method() will now cause a default
1437 RSA_METHOD to be chosen if one doesn't exist already.
1438 Previously this was only set during a call to RSA_new()
1439 or RSA_new_method(NULL) meaning it was possible for
1440 RSA_get_default_method() to return NULL.
1443 *) Added native name translation to the existing DSO code
1444 that will convert (if the flag to do so is set) filenames
1445 that are sufficiently small and have no path information
1446 into a canonical native form. Eg. "blah" converted to
1447 "libblah.so" or "blah.dll" etc.
1450 *) New function ERR_error_string_n(e, buf, len) which is like
1451 ERR_error_string(e, buf), but writes at most 'len' bytes
1452 including the 0 terminator. For ERR_error_string_n, 'buf'
1454 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1456 *) CONF library reworked to become more general. A new CONF
1457 configuration file reader "class" is implemented as well as a
1458 new functions (NCONF_*, for "New CONF") to handle it. The now
1459 old CONF_* functions are still there, but are reimplemented to
1460 work in terms of the new functions. Also, a set of functions
1461 to handle the internal storage of the configuration data is
1462 provided to make it easier to write new configuration file
1463 reader "classes" (I can definitely see something reading a
1464 configuration file in XML format, for example), called _CONF_*,
1465 or "the configuration storage API"...
1467 The new configuration file reading functions are:
1469 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1470 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1472 NCONF_default, NCONF_WIN32
1474 NCONF_dump_fp, NCONF_dump_bio
1476 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1477 NCONF_new creates a new CONF object. This works in the same way
1478 as other interfaces in OpenSSL, like the BIO interface.
1479 NCONF_dump_* dump the internal storage of the configuration file,
1480 which is useful for debugging. All other functions take the same
1481 arguments as the old CONF_* functions wth the exception of the
1482 first that must be a `CONF *' instead of a `LHASH *'.
1484 To make it easer to use the new classes with the old CONF_* functions,
1485 the function CONF_set_default_method is provided.
1488 *) Add '-tls1' option to 'openssl ciphers', which was already
1489 mentioned in the documentation but had not been implemented.
1490 (This option is not yet really useful because even the additional
1491 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1494 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1495 OpenSSL-based applications) load shared libraries and bind to
1496 them in a portable way.
1497 [Geoff Thorpe, with contributions from Richard Levitte]
1499 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1501 *) Make sure _lrotl and _lrotr are only used with MSVC.
1503 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1504 (the default implementation of RAND_status).
1506 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1507 to '-clrext' (= clear extensions), as intended and documented.
1508 [Bodo Moeller; inconsistency pointed out by Michael Attili
1511 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1512 was larger than the MD block size.
1513 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1515 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1516 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1517 using the passed key: if the passed key was a private key the result
1518 of X509_print(), for example, would be to print out all the private key
1522 *) des_quad_cksum() byte order bug fix.
1523 [Ulf Möller, using the problem description in krb4-0.9.7, where
1524 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1526 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1528 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1530 *) For easily testing in shell scripts whether some command
1531 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1532 returns with exit code 0 iff no command of the given name is available.
1533 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1534 the output goes to stdout and nothing is printed to stderr.
1535 Additional arguments are always ignored.
1537 Since for each cipher there is a command of the same name,
1538 the 'no-cipher' compilation switches can be tested this way.
1540 ('openssl no-XXX' is not able to detect pseudo-commands such
1541 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1544 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1547 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1548 is set; it will be thrown away anyway because each handshake creates
1550 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1551 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1552 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1553 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1556 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1557 'Q' and 'R' lose their special meanings (quit/renegotiate).
1558 This is part of what -quiet does; unlike -quiet, -ign_eof
1559 does not suppress any output.
1562 *) Add compatibility options to the purpose and trust code. The
1563 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1564 accepts a certificate or CA, this was the previous behaviour,
1565 with all the associated security issues.
1567 X509_TRUST_COMPAT is the old trust behaviour: only and
1568 automatically trust self signed roots in certificate store. A
1569 new trust setting X509_TRUST_DEFAULT is used to specify that
1570 a purpose has no associated trust setting and it should instead
1571 use the value in the default purpose.
1574 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1575 and fix a memory leak.
1578 *) In util/mkerr.pl (which implements 'make errors'), preserve
1579 reason strings from the previous version of the .c file, as
1580 the default to have only downcase letters (and digits) in
1581 automatically generated reasons codes is not always appropriate.
1584 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1585 using strerror. Previously, ERR_reason_error_string() returned
1586 library names as reason strings for SYSerr; but SYSerr is a special
1587 case where small numbers are errno values, not library numbers.
1590 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1591 converts DSA parameters into DH parameters. (When creating parameters,
1592 DSA_generate_parameters is used.)
1595 *) Include 'length' (recommended exponent length) in C code generated
1596 by 'openssl dhparam -C'.
1599 *) The second argument to set_label in perlasm was already being used
1600 so couldn't be used as a "file scope" flag. Moved to third argument
1604 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1605 instead of RAND_bytes for encryption IVs and salts.
1608 *) Include RAND_status() into RAND_METHOD instead of implementing
1609 it only for md_rand.c Otherwise replacing the PRNG by calling
1610 RAND_set_rand_method would be impossible.
1613 *) Don't let DSA_generate_key() enter an infinite loop if the random
1614 number generation fails.
1617 *) New 'rand' application for creating pseudo-random output.
1620 *) Added configuration support for Linux/IA64
1621 [Rolf Haberrecker <rolf@suse.de>]
1623 *) Assembler module support for Mingw32.
1626 *) Shared library support for HPUX (in shlib/).
1627 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1629 *) Shared library support for Solaris gcc.
1630 [Lutz Behnke <behnke@trustcenter.de>]
1632 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1634 *) PKCS7_encrypt() was adding text MIME headers twice because they
1635 were added manually and by SMIME_crlf_copy().
1638 *) In bntest.c don't call BN_rand with zero bits argument.
1639 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1641 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1642 case was implemented. This caused BN_div_recp() to fail occasionally.
1645 *) Add an optional second argument to the set_label() in the perl
1646 assembly language builder. If this argument exists and is set
1647 to 1 it signals that the assembler should use a symbol whose
1648 scope is the entire file, not just the current function. This
1649 is needed with MASM which uses the format label:: for this scope.
1650 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1652 *) Change the ASN1 types so they are typedefs by default. Before
1653 almost all types were #define'd to ASN1_STRING which was causing
1654 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1658 *) Change names of new functions to the new get1/get0 naming
1659 convention: After 'get1', the caller owns a reference count
1660 and has to call ..._free; 'get0' returns a pointer to some
1661 data structure without incrementing reference counters.
1662 (Some of the existing 'get' functions increment a reference
1663 counter, some don't.)
1664 Similarly, 'set1' and 'add1' functions increase reference
1665 counters or duplicate objects.
1668 *) Allow for the possibility of temp RSA key generation failure:
1669 the code used to assume it always worked and crashed on failure.
1672 *) Fix potential buffer overrun problem in BIO_printf().
1673 [Ulf Möller, using public domain code by Patrick Powell; problem
1674 pointed out by David Sacerdote <das33@cornell.edu>]
1676 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1677 RAND_egd() and RAND_status(). In the command line application,
1678 the EGD socket can be specified like a seed file using RANDFILE
1682 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1683 Some CAs (e.g. Verisign) distribute certificates in this form.
1686 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1687 list to exclude them. This means that no special compilation option
1688 is needed to use anonymous DH: it just needs to be included in the
1692 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1693 EVP_MD_type. The old functionality is available in a new macro called
1694 EVP_MD_md(). Change code that uses it and update docs.
1697 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1698 where the 'void *' argument is replaced by a function pointer argument.
1699 Previously 'void *' was abused to point to functions, which works on
1700 many platforms, but is not correct. As these functions are usually
1701 called by macros defined in OpenSSL header files, most source code
1702 should work without changes.
1705 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1706 sections with information on -D... compiler switches used for
1707 compiling the library so that applications can see them. To enable
1708 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1709 must be defined. E.g.,
1710 #define OPENSSL_ALGORITHM_DEFINES
1711 #include <openssl/opensslconf.h>
1712 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1713 [Richard Levitte, Ulf and Bodo Möller]
1715 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1719 *) Change the 'other' type in certificate aux info to a STACK_OF
1720 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1721 the required ASN1 format: arbitrary types determined by an OID.
1724 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1725 argument to 'req'. This is not because the function is newer or
1726 better than others it just uses the work 'NEW' in the certificate
1727 request header lines. Some software needs this.
1730 *) Reorganise password command line arguments: now passwords can be
1731 obtained from various sources. Delete the PEM_cb function and make
1732 it the default behaviour: i.e. if the callback is NULL and the
1733 usrdata argument is not NULL interpret it as a null terminated pass
1734 phrase. If usrdata and the callback are NULL then the pass phrase
1735 is prompted for as usual.
1738 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1739 the support is automatically enabled. The resulting binaries will
1740 autodetect the card and use it if present.
1741 [Ben Laurie and Compaq Inc.]
1743 *) Work around for Netscape hang bug. This sends certificate request
1744 and server done in one record. Since this is perfectly legal in the
1745 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1746 the bugs/SSLv3 entry for more info.
1749 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1752 *) Add -rand argument to smime and pkcs12 applications and read/write
1756 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1759 *) Add command line password options to the remaining applications.
1762 *) Bug fix for BN_div_recp() for numerators with an even number of
1766 *) More tests in bntest.c, and changed test_bn output.
1769 *) ./config recognizes MacOS X now.
1772 *) Bug fix for BN_div() when the first words of num and divsor are
1773 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1776 *) Add support for various broken PKCS#8 formats, and command line
1777 options to produce them.
1780 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1781 get temporary BIGNUMs from a BN_CTX.
1784 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1788 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1789 include a #define from the old name to the new. The original intent
1790 was that statically linked binaries could for example just call
1791 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1792 link with digests. This never worked becayse SSLeay_add_all_digests()
1793 and SSLeay_add_all_ciphers() were in the same source file so calling
1794 one would link with the other. They are now in separate source files.
1797 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1800 *) Use a less unusual form of the Miller-Rabin primality test (it used
1801 a binary algorithm for exponentiation integrated into the Miller-Rabin
1802 loop, our standard modexp algorithms are faster).
1805 *) Support for the EBCDIC character set completed.
1806 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1808 *) Source code cleanups: use const where appropriate, eliminate casts,
1809 use void * instead of char * in lhash.
1812 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1813 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1814 this the server could overwrite ephemeral keys that the client
1818 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1819 using 50 iterations of the Rabin-Miller test.
1821 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1822 iterations of the Rabin-Miller test as required by the appendix
1823 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1824 As BN_is_prime_fasttest includes trial division, DSA parameter
1825 generation becomes much faster.
1827 This implies a change for the callback functions in DSA_is_prime
1828 and DSA_generate_parameters: The callback function is called once
1829 for each positive witness in the Rabin-Miller test, not just
1830 occasionally in the inner loop; and the parameters to the
1831 callback function now provide an iteration count for the outer
1832 loop rather than for the current invocation of the inner loop.
1833 DSA_generate_parameters additionally can call the callback
1834 function with an 'iteration count' of -1, meaning that a
1835 candidate has passed the trial division test (when q is generated
1836 from an application-provided seed, trial division is skipped).
1839 *) New function BN_is_prime_fasttest that optionally does trial
1840 division before starting the Rabin-Miller test and has
1841 an additional BN_CTX * argument (whereas BN_is_prime always
1842 has to allocate at least one BN_CTX).
1843 'callback(1, -1, cb_arg)' is called when a number has passed the
1844 trial division stage.
1847 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1851 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1854 *) New function BN_pseudo_rand().
1857 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1858 bignum version of BN_from_montgomery() with the working code from
1859 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1863 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1864 made it impossible to use the same SSL_SESSION data structure in
1865 SSL2 clients in multiple threads.
1868 *) The return value of RAND_load_file() no longer counts bytes obtained
1869 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1870 to seed the PRNG (previously an explicit byte count was required).
1871 [Ulf Möller, Bodo Möller]
1873 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1874 used (char *) instead of (void *) and had casts all over the place.
1877 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1880 *) Retain source code compatibility for BN_prime_checks macro:
1881 BN_is_prime(..., BN_prime_checks, ...) now uses
1882 BN_prime_checks_for_size to determine the appropriate number of
1883 Rabin-Miller iterations.
1886 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1887 DH_CHECK_P_NOT_SAFE_PRIME.
1888 (Check if this is true? OpenPGP calls them "strong".)
1891 *) Merge the functionality of "dh" and "gendh" programs into a new program
1892 "dhparam". The old programs are retained for now but will handle DH keys
1893 (instead of parameters) in future.
1896 *) Make the ciphers, s_server and s_client programs check the return values
1897 when a new cipher list is set.
1900 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1901 ciphers. Before when the 56bit ciphers were enabled the sorting was
1904 The syntax for the cipher sorting has been extended to support sorting by
1905 cipher-strength (using the strength_bits hard coded in the tables).
1906 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1908 Fix a bug in the cipher-command parser: when supplying a cipher command
1909 string with an "undefined" symbol (neither command nor alphanumeric
1910 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1911 an error is flagged.
1913 Due to the strength-sorting extension, the code of the
1914 ssl_create_cipher_list() function was completely rearranged. I hope that
1915 the readability was also increased :-)
1916 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1918 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1919 for the first serial number and places 2 in the serial number file. This
1920 avoids problems when the root CA is created with serial number zero and
1921 the first user certificate has the same issuer name and serial number
1925 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1926 the new code. Add documentation for this stuff.
1929 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1930 X509_*() to X509at_*() on the grounds that they don't handle X509
1931 structures and behave in an analagous way to the X509v3 functions:
1932 they shouldn't be called directly but wrapper functions should be used
1935 So we also now have some wrapper functions that call the X509at functions
1936 when passed certificate requests. (TO DO: similar things can be done with
1937 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1938 things. Some of these need some d2i or i2d and print functionality
1939 because they handle more complex structures.)
1942 *) Add missing #ifndefs that caused missing symbols when building libssl
1943 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1944 NO_RSA in ssl/s2*.c.
1945 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1947 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1948 has a return value which indicates the quality of the random data
1949 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1950 error queue. New function RAND_pseudo_bytes() generates output that is
1951 guaranteed to be unique but not unpredictable. RAND_add is like
1952 RAND_seed, but takes an extra argument for an entropy estimate
1953 (RAND_seed always assumes full entropy).
1956 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1957 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1958 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1959 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1960 false-positive rate of at most 2^-80 for random input.
1963 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1966 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1967 in the 0.9.5 release), this returns the chain
1968 from an X509_CTX structure with a dup of the stack and all
1969 the X509 reference counts upped: so the stack will exist
1970 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1973 Also make SSL_SESSION_print() print out the verify return
1977 *) Add manpage for the pkcs12 command. Also change the default
1978 behaviour so MAC iteration counts are used unless the new
1979 -nomaciter option is used. This improves file security and
1980 only older versions of MSIE (4.0 for example) need it.
1983 *) Honor the no-xxx Configure options when creating .DEF files.
1986 *) Add PKCS#10 attributes to field table: challengePassword,
1987 unstructuredName and unstructuredAddress. These are taken from
1988 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1989 international characters are used.
1991 More changes to X509_ATTRIBUTE code: allow the setting of types
1992 based on strings. Remove the 'loc' parameter when adding
1993 attributes because these will be a SET OF encoding which is sorted
1997 *) Initial changes to the 'req' utility to allow request generation
1998 automation. This will allow an application to just generate a template
1999 file containing all the field values and have req construct the
2002 Initial support for X509_ATTRIBUTE handling. Stacks of these are
2003 used all over the place including certificate requests and PKCS#7
2004 structures. They are currently handled manually where necessary with
2005 some primitive wrappers for PKCS#7. The new functions behave in a
2006 manner analogous to the X509 extension functions: they allow
2007 attributes to be looked up by NID and added.
2009 Later something similar to the X509V3 code would be desirable to
2010 automatically handle the encoding, decoding and printing of the
2011 more complex types. The string types like challengePassword can
2012 be handled by the string table functions.
2014 Also modified the multi byte string table handling. Now there is
2015 a 'global mask' which masks out certain types. The table itself
2016 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2017 is useful when for example there is only one permissible type
2018 (as in countryName) and using the mask might result in no valid
2022 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2023 SSL_get_peer_finished to allow applications to obtain the latest
2024 Finished messages sent to the peer or expected from the peer,
2025 respectively. (SSL_get_peer_finished is usually the Finished message
2026 actually received from the peer, otherwise the protocol will be aborted.)
2028 As the Finished message are message digests of the complete handshake
2029 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2030 be used for external authentication procedures when the authentication
2031 provided by SSL/TLS is not desired or is not enough.
2034 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2035 the host supports BWX extension and if Compaq C is present on the
2036 $PATH. Just exploiting of the BWX extension results in 20-30%
2037 performance kick for some algorithms, e.g. DES and RC4 to mention
2038 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2042 *) Add support for MS "fast SGC". This is arguably a violation of the
2043 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2044 weak crypto and after checking the certificate is SGC a second one
2045 with strong crypto. MS SGC stops the first handshake after receiving
2046 the server certificate message and sends a second client hello. Since
2047 a server will typically do all the time consuming operations before
2048 expecting any further messages from the client (server key exchange
2049 is the most expensive) there is little difference between the two.
2051 To get OpenSSL to support MS SGC we have to permit a second client
2052 hello message after we have sent server done. In addition we have to
2053 reset the MAC if we do get this second client hello.
2056 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2057 if a DER encoded private key is RSA or DSA traditional format. Changed
2058 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2059 format DER encoded private key. Newer code should use PKCS#8 format which
2060 has the key type encoded in the ASN1 structure. Added DER private key
2061 support to pkcs8 application.
2064 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2065 ciphersuites has been selected (as required by the SSL 3/TLS 1
2066 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2067 is set, we interpret this as a request to violate the specification
2068 (the worst that can happen is a handshake failure, and 'correct'
2069 behaviour would result in a handshake failure anyway).
2072 *) In SSL_CTX_add_session, take into account that there might be multiple
2073 SSL_SESSION structures with the same session ID (e.g. when two threads
2074 concurrently obtain them from an external cache).
2075 The internal cache can handle only one SSL_SESSION with a given ID,
2076 so if there's a conflict, we now throw out the old one to achieve
2080 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2081 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2082 some routines that use cipher OIDs: some ciphers do not have OIDs
2083 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2087 *) Simplify the trust setting structure and code. Now we just have
2088 two sequences of OIDs for trusted and rejected settings. These will
2089 typically have values the same as the extended key usage extension
2090 and any application specific purposes.
2092 The trust checking code now has a default behaviour: it will just
2093 check for an object with the same NID as the passed id. Functions can
2094 be provided to override either the default behaviour or the behaviour
2095 for a given id. SSL client, server and email already have functions
2096 in place for compatibility: they check the NID and also return "trusted"
2097 if the certificate is self signed.
2100 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2101 traditional format into an EVP_PKEY structure.
2104 *) Add a password callback function PEM_cb() which either prompts for
2105 a password if usr_data is NULL or otherwise assumes it is a null
2106 terminated password. Allow passwords to be passed on command line
2107 environment or config files in a few more utilities.
2110 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2111 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2112 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2113 Update documentation.
2116 *) Support for ASN1 "NULL" type. This could be handled before by using
2117 ASN1_TYPE but there wasn't any function that would try to read a NULL
2118 and produce an error if it couldn't. For compatibility we also have
2119 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2120 don't allocate anything because they don't need to.
2123 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2125 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2127 *) Rebuild of the memory allocation routines used by OpenSSL code and
2128 possibly others as well. The purpose is to make an interface that
2129 provide hooks so anyone can build a separate set of allocation and
2130 deallocation routines to be used by OpenSSL, for example memory
2131 pool implementations, or something else, which was previously hard
2132 since Malloc(), Realloc() and Free() were defined as macros having
2133 the values malloc, realloc and free, respectively (except for Win32
2134 compilations). The same is provided for memory debugging code.
2135 OpenSSL already comes with functionality to find memory leaks, but
2136 this gives people a chance to debug other memory problems.
2138 With these changes, a new set of functions and macros have appeared:
2140 CRYPTO_set_mem_debug_functions() [F]
2141 CRYPTO_get_mem_debug_functions() [F]
2142 CRYPTO_dbg_set_options() [F]
2143 CRYPTO_dbg_get_options() [F]
2144 CRYPTO_malloc_debug_init() [M]
2146 The memory debug functions are NULL by default, unless the library
2147 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2148 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2149 gives the standard debugging functions that come with OpenSSL) or
2150 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2151 provided by the library user) must be used. When the standard
2152 debugging functions are used, CRYPTO_dbg_set_options can be used to
2153 request additional information:
2154 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2155 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2157 Also, things like CRYPTO_set_mem_functions will always give the
2158 expected result (the new set of functions is used for allocation
2159 and deallocation) at all times, regardless of platform and compiler
2162 To finish it up, some functions that were never use in any other
2163 way than through macros have a new API and new semantic:
2166 CRYPTO_dbg_realloc()
2169 All macros of value have retained their old syntax.
2170 [Richard Levitte and Bodo Moeller]
2172 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2173 ordering of SMIMECapabilities wasn't in "strength order" and there
2174 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2178 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2179 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2180 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2182 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2183 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2184 functionality to handle multipart/signed properly) and a utility
2185 called 'smime' to call all this stuff. This is based on code I
2186 originally wrote for Celo who have kindly allowed it to be
2187 included in OpenSSL.
2190 *) Add variants des_set_key_checked and des_set_key_unchecked of
2191 des_set_key (aka des_key_sched). Global variable des_check_key
2192 decides which of these is called by des_set_key; this way
2193 des_check_key behaves as it always did, but applications and
2194 the library itself, which was buggy for des_check_key == 1,
2195 have a cleaner way to pick the version they need.
2198 *) New function PKCS12_newpass() which changes the password of a
2202 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2203 dynamic mix. In both cases the ids can be used as an index into the
2204 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2205 functions so they accept a list of the field values and the
2206 application doesn't need to directly manipulate the X509_TRUST
2210 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2214 *) Modify the way the V3 extension code looks up extensions. This now
2215 works in a similar way to the object code: we have some "standard"
2216 extensions in a static table which is searched with OBJ_bsearch()
2217 and the application can add dynamic ones if needed. The file
2218 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2219 updated whenever a new extension is added to the core code and kept
2220 in ext_nid order. There is a simple program 'tabtest.c' which checks
2221 this. New extensions are not added too often so this file can readily
2222 be maintained manually.
2224 There are two big advantages in doing things this way. The extensions
2225 can be looked up immediately and no longer need to be "added" using
2226 X509V3_add_standard_extensions(): this function now does nothing.
2227 [Side note: I get *lots* of email saying the extension code doesn't
2228 work because people forget to call this function]
2229 Also no dynamic allocation is done unless new extensions are added:
2230 so if we don't add custom extensions there is no need to call
2231 X509V3_EXT_cleanup().
2234 *) Modify enc utility's salting as follows: make salting the default. Add a
2235 magic header, so unsalted files fail gracefully instead of just decrypting
2236 to garbage. This is because not salting is a big security hole, so people
2237 should be discouraged from doing it.
2240 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2241 digest to be passed on the command line but it only used this
2242 parameter when signing a certificate. Modified so all relevant
2243 operations are affected by the digest parameter including the
2244 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2245 DSA key was used because it didn't fix the digest.
2248 *) Initial certificate chain verify code. Currently tests the untrusted
2249 certificates for consistency with the verify purpose (which is set
2250 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2252 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2253 this is because it will reject chains with invalid extensions whereas
2254 every previous version of OpenSSL and SSLeay made no checks at all.
2256 Trust code: checks the root CA for the relevant trust settings. Trust
2257 settings have an initial value consistent with the verify purpose: e.g.
2258 if the verify purpose is for SSL client use it expects the CA to be
2259 trusted for SSL client use. However the default value can be changed to
2260 permit custom trust settings: one example of this would be to only trust
2261 certificates from a specific "secure" set of CAs.
2263 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2264 which should be used for version portability: especially since the
2265 verify structure is likely to change more often now.
2267 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2268 to set them. If not set then assume SSL clients will verify SSL servers
2271 Two new options to the verify program: -untrusted allows a set of
2272 untrusted certificates to be passed in and -purpose which sets the
2273 intended purpose of the certificate. If a purpose is set then the
2274 new chain verify code is used to check extension consistency.
2277 *) Support for the authority information access extension.
2280 *) Modify RSA and DSA PEM read routines to transparently handle
2281 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2282 public keys in a format compatible with certificate
2283 SubjectPublicKeyInfo structures. Unfortunately there were already
2284 functions called *_PublicKey_* which used various odd formats so
2285 these are retained for compatibility: however the DSA variants were
2286 never in a public release so they have been deleted. Changed dsa/rsa
2287 utilities to handle the new format: note no releases ever handled public
2288 keys so we should be OK.
2290 The primary motivation for this change is to avoid the same fiasco
2291 that dogs private keys: there are several incompatible private key
2292 formats some of which are standard and some OpenSSL specific and
2293 require various evil hacks to allow partial transparent handling and
2294 even then it doesn't work with DER formats. Given the option anything
2295 other than PKCS#8 should be dumped: but the other formats have to
2296 stay in the name of compatibility.
2298 With public keys and the benefit of hindsight one standard format
2299 is used which works with EVP_PKEY, RSA or DSA structures: though
2300 it clearly returns an error if you try to read the wrong kind of key.
2302 Added a -pubkey option to the 'x509' utility to output the public key.
2303 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2304 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2305 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2306 that do the same as the EVP_PKEY_assign_*() except they up the
2307 reference count of the added key (they don't "swallow" the
2311 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2312 CRLs would fail if the file contained no certificates or no CRLs:
2313 added a new function to read in both types and return the number
2314 read: this means that if none are read it will be an error. The
2315 DER versions of the certificate and CRL reader would always fail
2316 because it isn't possible to mix certificates and CRLs in DER format
2317 without choking one or the other routine. Changed this to just read
2318 a certificate: this is the best we can do. Also modified the code
2319 in apps/verify.c to take notice of return codes: it was previously
2320 attempting to read in certificates from NULL pointers and ignoring
2321 any errors: this is one reason why the cert and CRL reader seemed
2322 to work. It doesn't check return codes from the default certificate
2323 routines: these may well fail if the certificates aren't installed.
2326 *) Code to support otherName option in GeneralName.
2329 *) First update to verify code. Change the verify utility
2330 so it warns if it is passed a self signed certificate:
2331 for consistency with the normal behaviour. X509_verify
2332 has been modified to it will now verify a self signed
2333 certificate if *exactly* the same certificate appears
2334 in the store: it was previously impossible to trust a
2335 single self signed certificate. This means that:
2336 openssl verify ss.pem
2337 now gives a warning about a self signed certificate but
2338 openssl verify -CAfile ss.pem ss.pem
2342 *) For servers, store verify_result in SSL_SESSION data structure
2343 (and add it to external session representation).
2344 This is needed when client certificate verifications fails,
2345 but an application-provided verification callback (set by
2346 SSL_CTX_set_cert_verify_callback) allows accepting the session
2347 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2348 but returns 1): When the session is reused, we have to set
2349 ssl->verify_result to the appropriate error code to avoid
2351 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2353 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2354 case in PKCS7_dataInit() where the signed PKCS7 structure
2355 didn't contain any existing data because it was being created.
2356 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2358 *) Add a salt to the key derivation routines in enc.c. This
2359 forms the first 8 bytes of the encrypted file. Also add a
2360 -S option to allow a salt to be input on the command line.
2363 *) New function X509_cmp(). Oddly enough there wasn't a function
2364 to compare two certificates. We do this by working out the SHA1
2365 hash and comparing that. X509_cmp() will be needed by the trust
2369 *) SSL_get1_session() is like SSL_get_session(), but increments
2370 the reference count in the SSL_SESSION returned.
2371 [Geoff Thorpe <geoff@eu.c2.net>]
2373 *) Fix for 'req': it was adding a null to request attributes.
2374 Also change the X509_LOOKUP and X509_INFO code to handle
2375 certificate auxiliary information.
2378 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2382 *) Add the possibility to add extra information to the memory leak
2383 detecting output, to form tracebacks, showing from where each
2384 allocation was originated: CRYPTO_push_info("constant string") adds
2385 the string plus current file name and line number to a per-thread
2386 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2387 is like calling CYRPTO_pop_info() until the stack is empty.
2388 Also updated memory leak detection code to be multi-thread-safe.
2391 *) Add options -text and -noout to pkcs7 utility and delete the
2392 encryption options which never did anything. Update docs.
2395 *) Add options to some of the utilities to allow the pass phrase
2396 to be included on either the command line (not recommended on
2397 OSes like Unix) or read from the environment. Update the
2398 manpages and fix a few bugs.
2401 *) Add a few manpages for some of the openssl commands.
2404 *) Fix the -revoke option in ca. It was freeing up memory twice,
2405 leaking and not finding already revoked certificates.
2408 *) Extensive changes to support certificate auxiliary information.
2409 This involves the use of X509_CERT_AUX structure and X509_AUX
2410 functions. An X509_AUX function such as PEM_read_X509_AUX()
2411 can still read in a certificate file in the usual way but it
2412 will also read in any additional "auxiliary information". By
2413 doing things this way a fair degree of compatibility can be
2414 retained: existing certificates can have this information added
2415 using the new 'x509' options.
2417 Current auxiliary information includes an "alias" and some trust
2418 settings. The trust settings will ultimately be used in enhanced
2419 certificate chain verification routines: currently a certificate
2420 can only be trusted if it is self signed and then it is trusted
2424 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2425 The problem was that one of the replacement routines had not been working
2426 since SSLeay releases. For now the offending routine has been replaced
2427 with non-optimised assembler. Even so, this now gives around 95%
2428 performance improvement for 1024 bit RSA signs.
2431 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2432 handling. Most clients have the effective key size in bits equal to
2433 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2434 A few however don't do this and instead use the size of the decrypted key
2435 to determine the RC2 key length and the AlgorithmIdentifier to determine
2436 the effective key length. In this case the effective key length can still
2437 be 40 bits but the key length can be 168 bits for example. This is fixed
2438 by manually forcing an RC2 key into the EVP_PKEY structure because the
2439 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2440 the key length and effective key length are equal.
2443 *) Add a bunch of functions that should simplify the creation of
2444 X509_NAME structures. Now you should be able to do:
2445 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2446 and have it automatically work out the correct field type and fill in
2447 the structures. The more adventurous can try:
2448 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2449 and it will (hopefully) work out the correct multibyte encoding.
2452 *) Change the 'req' utility to use the new field handling and multibyte
2453 copy routines. Before the DN field creation was handled in an ad hoc
2454 way in req, ca, and x509 which was rather broken and didn't support
2455 BMPStrings or UTF8Strings. Since some software doesn't implement
2456 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2457 using the dirstring_type option. See the new comment in the default
2458 openssl.cnf for more info.
2461 *) Make crypto/rand/md_rand.c more robust:
2462 - Assure unique random numbers after fork().
2463 - Make sure that concurrent threads access the global counter and
2464 md serializably so that we never lose entropy in them
2465 or use exactly the same state in multiple threads.
2466 Access to the large state is not always serializable because
2467 the additional locking could be a performance killer, and
2468 md should be large enough anyway.
2471 *) New file apps/app_rand.c with commonly needed functionality
2472 for handling the random seed file.
2474 Use the random seed file in some applications that previously did not:
2476 dsaparam -genkey (which also ignored its '-rand' option),
2479 x509 (when signing).
2480 Except on systems with /dev/urandom, it is crucial to have a random
2481 seed file at least for key creation, DSA signing, and for DH exchanges;
2482 for RSA signatures we could do without one.
2484 gendh and gendsa (unlike genrsa) used to read only the first byte
2485 of each file listed in the '-rand' option. The function as previously
2486 found in genrsa is now in app_rand.c and is used by all programs
2487 that support '-rand'.
2490 *) In RAND_write_file, use mode 0600 for creating files;
2491 don't just chmod when it may be too late.
2494 *) Report an error from X509_STORE_load_locations
2495 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2498 *) New function ASN1_mbstring_copy() this copies a string in either
2499 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2500 into an ASN1_STRING type. A mask of permissible types is passed
2501 and it chooses the "minimal" type to use or an error if not type
2505 *) Add function equivalents to the various macros in asn1.h. The old
2506 macros are retained with an M_ prefix. Code inside the library can
2507 use the M_ macros. External code (including the openssl utility)
2508 should *NOT* in order to be "shared library friendly".
2511 *) Add various functions that can check a certificate's extensions
2512 to see if it usable for various purposes such as SSL client,
2513 server or S/MIME and CAs of these types. This is currently
2514 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2515 verification. Also added a -purpose flag to x509 utility to
2516 print out all the purposes.
2519 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2523 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2524 for, obtain and decode and extension and obtain its critical flag.
2525 This allows all the necessary extension code to be handled in a
2526 single function call.
2529 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2530 platforms. See crypto/rc4/rc4_enc.c for further details.
2533 *) New -noout option to asn1parse. This causes no output to be produced
2534 its main use is when combined with -strparse and -out to extract data
2535 from a file (which may not be in ASN.1 format).
2538 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2539 when producing the local key id.
2540 [Richard Levitte <levitte@stacken.kth.se>]
2542 *) New option -dhparam in s_server. This allows a DH parameter file to be
2543 stated explicitly. If it is not stated then it tries the first server
2544 certificate file. The previous behaviour hard coded the filename
2548 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2549 a public key to be input or output. For example:
2550 openssl rsa -in key.pem -pubout -out pubkey.pem
2551 Also added necessary DSA public key functions to handle this.
2554 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2555 in the message. This was handled by allowing
2556 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2557 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2559 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2560 to the end of the strings whereas this didn't. This would cause problems
2561 if strings read with d2i_ASN1_bytes() were later modified.
2562 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2564 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2565 data and it contains EOF it will end up returning an error. This is
2566 caused by input 46 bytes long. The cause is due to the way base64
2567 BIOs find the start of base64 encoded data. They do this by trying a
2568 trial decode on each line until they find one that works. When they
2569 do a flag is set and it starts again knowing it can pass all the
2570 data directly through the decoder. Unfortunately it doesn't reset
2571 the context it uses. This means that if EOF is reached an attempt
2572 is made to pass two EOFs through the context and this causes the
2573 resulting error. This can also cause other problems as well. As is
2574 usual with these problems it takes *ages* to find and the fix is
2575 trivial: move one line.
2576 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2578 *) Ugly workaround to get s_client and s_server working under Windows. The
2579 old code wouldn't work because it needed to select() on sockets and the
2580 tty (for keypresses and to see if data could be written). Win32 only
2581 supports select() on sockets so we select() with a 1s timeout on the
2582 sockets and then see if any characters are waiting to be read, if none
2583 are present then we retry, we also assume we can always write data to
2584 the tty. This isn't nice because the code then blocks until we've
2585 received a complete line of data and it is effectively polling the
2586 keyboard at 1s intervals: however it's quite a bit better than not
2587 working at all :-) A dedicated Windows application might handle this
2588 with an event loop for example.
2591 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2592 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2593 will be called when RSA_sign() and RSA_verify() are used. This is useful
2594 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2595 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2596 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2597 This necessitated the support of an extra signature type NID_md5_sha1
2598 for SSL signatures and modifications to the SSL library to use it instead
2599 of calling RSA_public_decrypt() and RSA_private_encrypt().
2602 *) Add new -verify -CAfile and -CApath options to the crl program, these
2603 will lookup a CRL issuers certificate and verify the signature in a
2604 similar way to the verify program. Tidy up the crl program so it
2605 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2606 less strict. It will now permit CRL extensions even if it is not
2607 a V2 CRL: this will allow it to tolerate some broken CRLs.
2610 *) Initialize all non-automatic variables each time one of the openssl
2611 sub-programs is started (this is necessary as they may be started
2612 multiple times from the "OpenSSL>" prompt).
2613 [Lennart Bang, Bodo Moeller]
2615 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2616 removing all other RSA functionality (this is what NO_RSA does). This
2617 is so (for example) those in the US can disable those operations covered
2618 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2622 *) Non-copying interface to BIO pairs.
2623 (still largely untested)
2626 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2627 ASCII string. This was handled independently in various places before.
2630 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2631 UTF8 strings a character at a time.
2634 *) Use client_version from client hello to select the protocol
2635 (s23_srvr.c) and for RSA client key exchange verification
2636 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2639 *) Add various utility functions to handle SPKACs, these were previously
2640 handled by poking round in the structure internals. Added new function
2641 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2642 print, verify and generate SPKACs. Based on an original idea from
2643 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2646 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2649 *) Allow the config file extension section to be overwritten on the
2650 command line. Based on an original idea from Massimiliano Pala
2651 <madwolf@comune.modena.it>. The new option is called -extensions
2652 and can be applied to ca, req and x509. Also -reqexts to override
2653 the request extensions in req and -crlexts to override the crl extensions
2657 *) Add new feature to the SPKAC handling in ca. Now you can include
2658 the same field multiple times by preceding it by "XXXX." for example:
2661 this is the same syntax as used in the req config file.
2664 *) Allow certificate extensions to be added to certificate requests. These
2665 are specified in a 'req_extensions' option of the req section of the
2666 config file. They can be printed out with the -text option to req but
2667 are otherwise ignored at present.
2670 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2671 data read consists of only the final block it would not decrypted because
2672 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2673 A misplaced 'break' also meant the decrypted final block might not be
2674 copied until the next read.
2677 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2678 a few extra parameters to the DH structure: these will be useful if
2679 for example we want the value of 'q' or implement X9.42 DH.
2682 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2683 provides hooks that allow the default DSA functions or functions on a
2684 "per key" basis to be replaced. This allows hardware acceleration and
2685 hardware key storage to be handled without major modification to the
2686 library. Also added low level modexp hooks and CRYPTO_EX structure and
2687 associated functions.
2690 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2691 as "read only": it can't be written to and the buffer it points to will
2692 not be freed. Reading from a read only BIO is much more efficient than
2693 a normal memory BIO. This was added because there are several times when
2694 an area of memory needs to be read from a BIO. The previous method was
2695 to create a memory BIO and write the data to it, this results in two
2696 copies of the data and an O(n^2) reading algorithm. There is a new
2697 function BIO_new_mem_buf() which creates a read only memory BIO from
2698 an area of memory. Also modified the PKCS#7 routines to use read only
2702 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2703 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2704 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2705 but a retry condition occured while trying to read the rest.
2708 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2709 NID_pkcs7_encrypted by default: this was wrong since this should almost
2710 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2711 the encrypted data type: this is a more sensible place to put it and it
2712 allows the PKCS#12 code to be tidied up that duplicated this
2716 *) Changed obj_dat.pl script so it takes its input and output files on
2717 the command line. This should avoid shell escape redirection problems
2721 *) Initial support for certificate extension requests, these are included
2722 in things like Xenroll certificate requests. Included functions to allow
2723 extensions to be obtained and added.
2726 *) -crlf option to s_client and s_server for sending newlines as
2727 CRLF (as required by many protocols).
2730 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2732 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2733 [Ralf S. Engelschall]
2735 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2736 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2738 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2742 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2743 DH parameters/keys (q is lost during that conversion, but the resulting
2744 DH parameters contain its length).
2746 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2747 much faster than DH_generate_parameters (which creates parameters
2748 where p = 2*q + 1), and also the smaller q makes DH computations
2749 much more efficient (160-bit exponentiation instead of 1024-bit
2750 exponentiation); so this provides a convenient way to support DHE
2751 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2752 utter importance to use
2753 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2755 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2756 when such DH parameters are used, because otherwise small subgroup
2757 attacks may become possible!
2760 *) Avoid memory leak in i2d_DHparams.
2763 *) Allow the -k option to be used more than once in the enc program:
2764 this allows the same encrypted message to be read by multiple recipients.
2767 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2768 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2769 it will always use the numerical form of the OID, even if it has a short
2773 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2774 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2775 otherwise bn_mod_exp was called. In the case of hardware keys for example
2776 no private key components need be present and it might store extra data
2777 in the RSA structure, which cannot be accessed from bn_mod_exp.
2778 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2779 private key operations.
2782 *) Added support for SPARC Linux.
2785 *) pem_password_cb function type incompatibly changed from
2786 typedef int pem_password_cb(char *buf, int size, int rwflag);
2788 ....(char *buf, int size, int rwflag, void *userdata);
2789 so that applications can pass data to their callbacks:
2790 The PEM[_ASN1]_{read,write}... functions and macros now take an
2791 additional void * argument, which is just handed through whenever
2792 the password callback is called.
2793 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2795 New function SSL_CTX_set_default_passwd_cb_userdata.
2797 Compatibility note: As many C implementations push function arguments
2798 onto the stack in reverse order, the new library version is likely to
2799 interoperate with programs that have been compiled with the old
2800 pem_password_cb definition (PEM_whatever takes some data that
2801 happens to be on the stack as its last argument, and the callback
2802 just ignores this garbage); but there is no guarantee whatsoever that
2805 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2806 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2807 problems not only on Windows, but also on some Unix platforms.
2808 To avoid problematic command lines, these definitions are now in an
2809 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2810 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2813 *) MIPS III/IV assembler module is reimplemented.
2816 *) More DES library cleanups: remove references to srand/rand and
2817 delete an unused file.
2820 *) Add support for the the free Netwide assembler (NASM) under Win32,
2821 since not many people have MASM (ml) and it can be hard to obtain.
2822 This is currently experimental but it seems to work OK and pass all
2823 the tests. Check out INSTALL.W32 for info.
2826 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2827 without temporary keys kept an extra copy of the server key,
2828 and connections with temporary keys did not free everything in case
2832 *) New function RSA_check_key and new openssl rsa option -check
2833 for verifying the consistency of RSA keys.
2834 [Ulf Moeller, Bodo Moeller]
2836 *) Various changes to make Win32 compile work:
2837 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2838 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2839 comparison" warnings.
2840 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2843 *) Add a debugging option to PKCS#5 v2 key generation function: when
2844 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2845 derived keys are printed to stderr.
2848 *) Copy the flags in ASN1_STRING_dup().
2849 [Roman E. Pavlov <pre@mo.msk.ru>]
2851 *) The x509 application mishandled signing requests containing DSA
2852 keys when the signing key was also DSA and the parameters didn't match.
2854 It was supposed to omit the parameters when they matched the signing key:
2855 the verifying software was then supposed to automatically use the CA's
2856 parameters if they were absent from the end user certificate.
2858 Omitting parameters is no longer recommended. The test was also
2859 the wrong way round! This was probably due to unusual behaviour in
2860 EVP_cmp_parameters() which returns 1 if the parameters match.
2861 This meant that parameters were omitted when they *didn't* match and
2862 the certificate was useless. Certificates signed with 'ca' didn't have
2864 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2866 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2867 The interface is as follows:
2868 Applications can use
2869 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2870 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2871 "off" is now the default.
2872 The library internally uses
2873 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2874 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2875 to disable memory-checking temporarily.
2877 Some inconsistent states that previously were possible (and were
2878 even the default) are now avoided.
2880 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2881 with each memory chunk allocated; this is occasionally more helpful
2882 than just having a counter.
2884 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2886 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2890 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2891 which largely parallels "options", but is for changing API behaviour,
2892 whereas "options" are about protocol behaviour.
2893 Initial "mode" flags are:
2895 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2896 a single record has been written.
2897 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2898 retries use the same buffer location.
2899 (But all of the contents must be
2903 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
2906 *) Fix problems with no-hmac etc.
2907 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2909 *) New functions RSA_get_default_method(), RSA_set_method() and
2910 RSA_get_method(). These allows replacement of RSA_METHODs without having
2911 to mess around with the internals of an RSA structure.
2914 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2915 Also really enable memory leak checks in openssl.c and in some
2917 [Chad C. Mulligan, Bodo Moeller]
2919 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2920 up the length of negative integers. This has now been simplified to just
2921 store the length when it is first determined and use it later, rather
2922 than trying to keep track of where data is copied and updating it to
2924 [Steve Henson, reported by Brien Wheeler
2925 <bwheeler@authentica-security.com>]
2927 *) Add a new function PKCS7_signatureVerify. This allows the verification
2928 of a PKCS#7 signature but with the signing certificate passed to the
2929 function itself. This contrasts with PKCS7_dataVerify which assumes the
2930 certificate is present in the PKCS#7 structure. This isn't always the
2931 case: certificates can be omitted from a PKCS#7 structure and be
2932 distributed by "out of band" means (such as a certificate database).
2935 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2936 function prototypes in pem.h, also change util/mkdef.pl to add the
2937 necessary function names.
2940 *) mk1mf.pl (used by Windows builds) did not properly read the
2941 options set by Configure in the top level Makefile, and Configure
2942 was not even able to write more than one option correctly.
2943 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2946 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2947 file to be loaded from a BIO or FILE pointer. The BIO version will
2948 for example allow memory BIOs to contain config info.
2951 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2952 Whoever hopes to achieve shared-library compatibility across versions
2953 must use this, not the compile-time macro.
2954 (Exercise 0.9.4: Which is the minimum library version required by
2956 Note: All this applies only to multi-threaded programs, others don't
2960 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2961 through a BIO pair triggered the default case, i.e.
2962 SSLerr(...,SSL_R_UNKNOWN_STATE).
2965 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2966 can use the SSL library even if none of the specific BIOs is
2970 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2971 for the encoded length.
2972 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2974 *) Add initial documentation of the X509V3 functions.
2977 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2978 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2979 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2980 secure PKCS#8 private key format with a high iteration count.
2983 *) Fix determination of Perl interpreter: A perl or perl5
2984 _directory_ in $PATH was also accepted as the interpreter.
2985 [Ralf S. Engelschall]
2987 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2988 wrong with it but it was very old and did things like calling
2989 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2993 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2994 to use the new extension code.
2997 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2998 with macros. This should make it easier to change their form, add extra
2999 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
3003 *) Add to configuration table a new entry that can specify an alternative
3004 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
3005 according to Mark Crispin <MRC@Panda.COM>.
3009 *) DES CBC did not update the IV. Weird.
3012 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3013 Changing the behaviour of the former might break existing programs --
3014 where IV updating is needed, des_ncbc_encrypt can be used.
3017 *) When bntest is run from "make test" it drives bc to check its
3018 calculations, as well as internally checking them. If an internal check
3019 fails, it needs to cause bc to give a non-zero result or make test carries
3020 on without noticing the failure. Fixed.
3023 *) DES library cleanups.
3026 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3027 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3028 ciphers. NOTE: although the key derivation function has been verified
3029 against some published test vectors it has not been extensively tested
3030 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3034 *) Instead of "mkdir -p", which is not fully portable, use new
3035 Perl script "util/mkdir-p.pl".
3038 *) Rewrite the way password based encryption (PBE) is handled. It used to
3039 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3040 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3041 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3042 the 'parameter' field of the AlgorithmIdentifier is passed to the
3043 underlying key generation function so it must do its own ASN1 parsing.
3044 This has also changed the EVP_PBE_CipherInit() function which now has a
3045 'parameter' argument instead of literal salt and iteration count values
3046 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3049 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3050 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3051 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3052 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3053 value was just used as a "magic string" and not used directly its
3054 value doesn't matter.
3057 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3061 *) "linux-sparc64" configuration (ultrapenguin).
3062 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3063 "linux-sparc" configuration.
3064 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3066 *) config now generates no-xxx options for missing ciphers.
3069 *) Support the EBCDIC character set (work in progress).
3070 File ebcdic.c not yet included because it has a different license.
3071 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3073 *) Support BS2000/OSD-POSIX.
3074 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3076 *) Make callbacks for key generation use void * instead of char *.
3079 *) Make S/MIME samples compile (not yet tested).
3082 *) Additional typesafe stacks.
3085 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3089 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3091 *) New configuration variant "sco5-gcc".
3093 *) Updated some demos.
3094 [Sean O Riordain, Wade Scholine]
3096 *) Add missing BIO_free at exit of pkcs12 application.
3099 *) Fix memory leak in conf.c.
3102 *) Updates for Win32 to assembler version of MD5.
3105 *) Set #! path to perl in apps/der_chop to where we found it
3106 instead of using a fixed path.
3109 *) SHA library changes for irix64-mips4-cc.
3112 *) Improvements for VMS support.
3116 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3118 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3119 This also avoids the problems with SC4.2 and unpatched SC5.
3120 [Andy Polyakov <appro@fy.chalmers.se>]
3122 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3123 These are required because of the typesafe stack would otherwise break
3124 existing code. If old code used a structure member which used to be STACK
3125 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3126 sk_num or sk_value it would produce an error because the num, data members
3127 are not present in STACK_OF. Now it just produces a warning. sk_set
3128 replaces the old method of assigning a value to sk_value
3129 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3130 that does this will no longer work (and should use sk_set instead) but
3131 this could be regarded as a "questionable" behaviour anyway.
3134 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3135 correctly handle encrypted S/MIME data.
3138 *) Change type of various DES function arguments from des_cblock
3139 (which means, in function argument declarations, pointer to char)
3140 to des_cblock * (meaning pointer to array with 8 char elements),
3141 which allows the compiler to do more typechecking; it was like
3142 that back in SSLeay, but with lots of ugly casts.
3144 Introduce new type const_des_cblock.
3147 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3148 problems: find RecipientInfo structure that matches recipient certificate
3149 and initialise the ASN1 structures properly based on passed cipher.
3152 *) Belatedly make the BN tests actually check the results.
3155 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3156 to and from BNs: it was completely broken. New compilation option
3157 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3158 key elements as negative integers.
3161 *) Reorganize and speed up MD5.
3162 [Andy Polyakov <appro@fy.chalmers.se>]
3165 [Richard Levitte <richard@levitte.org>]
3167 *) New option -out to asn1parse to allow the parsed structure to be
3168 output to a file. This is most useful when combined with the -strparse
3169 option to examine the output of things like OCTET STRINGS.
3172 *) Make SSL library a little more fool-proof by not requiring any longer
3173 that SSL_set_{accept,connect}_state be called before
3174 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3175 in many applications because usually everything *appeared* to work as
3176 intended anyway -- now it really works as intended).
3179 *) Move openssl.cnf out of lib/.
3182 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3183 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3184 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3185 [Ralf S. Engelschall]
3187 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3188 handle PKCS#7 enveloped data properly.
3189 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3191 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3192 copying pointers. The cert_st handling is changed by this in
3193 various ways (and thus what used to be known as ctx->default_cert
3194 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3195 any longer when s->cert does not give us what we need).
3196 ssl_cert_instantiate becomes obsolete by this change.
3197 As soon as we've got the new code right (possibly it already is?),
3198 we have solved a couple of bugs of the earlier code where s->cert
3199 was used as if it could not have been shared with other SSL structures.
3201 Note that using the SSL API in certain dirty ways now will result
3202 in different behaviour than observed with earlier library versions:
3203 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3204 does not influence s as it used to.
3206 In order to clean up things more thoroughly, inside SSL_SESSION
3207 we don't use CERT any longer, but a new structure SESS_CERT
3208 that holds per-session data (if available); currently, this is
3209 the peer's certificate chain and, for clients, the server's certificate
3210 and temporary key. CERT holds only those values that can have
3211 meaningful defaults in an SSL_CTX.
3214 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3215 from the internal representation. Various PKCS#7 fixes: remove some
3216 evil casts and set the enc_dig_alg field properly based on the signing
3220 *) Allow PKCS#12 password to be set from the command line or the
3221 environment. Let 'ca' get its config file name from the environment
3222 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3226 *) Allow certificate policies extension to use an IA5STRING for the
3227 organization field. This is contrary to the PKIX definition but
3228 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3232 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3233 without disallowing inline assembler and the like for non-pedantic builds.
3236 *) Support Borland C++ builder.
3237 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3242 *) SHA-1 cleanups and performance enhancements.
3243 [Andy Polyakov <appro@fy.chalmers.se>]
3245 *) Sparc v8plus assembler for the bignum library.
3246 [Andy Polyakov <appro@fy.chalmers.se>]
3248 *) Accept any -xxx and +xxx compiler options in Configure.
3251 *) Update HPUX configuration.
3254 *) Add missing sk_<type>_unshift() function to safestack.h
3255 [Ralf S. Engelschall]
3257 *) New function SSL_CTX_use_certificate_chain_file that sets the
3258 "extra_cert"s in addition to the certificate. (This makes sense
3259 only for "PEM" format files, as chains as a whole are not
3263 *) Support verify_depth from the SSL API.
3264 x509_vfy.c had what can be considered an off-by-one-error:
3265 Its depth (which was not part of the external interface)
3266 was actually counting the number of certificates in a chain;
3267 now it really counts the depth.
3270 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3271 instead of X509err, which often resulted in confusing error
3272 messages since the error codes are not globally unique
3273 (e.g. an alleged error in ssl3_accept when a certificate
3274 didn't match the private key).
3276 *) New function SSL_CTX_set_session_id_context that allows to set a default
3277 value (so that you don't need SSL_set_session_id_context for each
3278 connection using the SSL_CTX).
3281 *) OAEP decoding bug fix.
3284 *) Support INSTALL_PREFIX for package builders, as proposed by
3288 *) New Configure options "threads" and "no-threads". For systems
3289 where the proper compiler options are known (currently Solaris
3290 and Linux), "threads" is the default.
3293 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3296 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3297 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3298 such as /usr/local/bin.
3301 *) "make linux-shared" to build shared libraries.
3302 [Niels Poppe <niels@netbox.org>]
3304 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3307 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3308 extension adding in x509 utility.
3311 *) Remove NOPROTO sections and error code comments.
3314 *) Partial rewrite of the DEF file generator to now parse the ANSI
3318 *) New Configure options --prefix=DIR and --openssldir=DIR.
3321 *) Complete rewrite of the error code script(s). It is all now handled
3322 by one script at the top level which handles error code gathering,
3323 header rewriting and C source file generation. It should be much better
3324 than the old method: it now uses a modified version of Ulf's parser to
3325 read the ANSI prototypes in all header files (thus the old K&R definitions
3326 aren't needed for error creation any more) and do a better job of
3327 translating function codes into names. The old 'ASN1 error code imbedded
3328 in a comment' is no longer necessary and it doesn't use .err files which
3329 have now been deleted. Also the error code call doesn't have to appear all
3330 on one line (which resulted in some large lines...).
3333 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3336 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3337 0 (which usually indicates a closed connection), but continue reading.
3340 *) Fix some race conditions.
3343 *) Add support for CRL distribution points extension. Add Certificate
3344 Policies and CRL distribution points documentation.
3347 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3350 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3351 8 of keying material. Merlin has also confirmed interop with this fix
3352 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3353 [Merlin Hughes <merlin@baltimore.ie>]
3355 *) Fix lots of warnings.
3356 [Richard Levitte <levitte@stacken.kth.se>]
3358 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3359 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3360 [Richard Levitte <levitte@stacken.kth.se>]
3362 *) Fix problems with sizeof(long) == 8.
3363 [Andy Polyakov <appro@fy.chalmers.se>]
3365 *) Change functions to ANSI C.
3368 *) Fix typos in error codes.
3369 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3371 *) Remove defunct assembler files from Configure.
3374 *) SPARC v8 assembler BIGNUM implementation.
3375 [Andy Polyakov <appro@fy.chalmers.se>]
3377 *) Support for Certificate Policies extension: both print and set.
3378 Various additions to support the r2i method this uses.
3381 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3382 return a const string when you are expecting an allocated buffer.
3385 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3386 types DirectoryString and DisplayText.
3389 *) Add code to allow r2i extensions to access the configuration database,
3390 add an LHASH database driver and add several ctx helper functions.
3393 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3394 fail when they extended the size of a BIGNUM.
3397 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3398 support typesafe stack.
3401 *) Fix typo in SSL_[gs]et_options().
3402 [Nils Frostberg <nils@medcom.se>]
3404 *) Delete various functions and files that belonged to the (now obsolete)
3405 old X509V3 handling code.
3408 *) New Configure option "rsaref".
3411 *) Don't auto-generate pem.h.
3414 *) Introduce type-safe ASN.1 SETs.
3417 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3418 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3420 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3421 that links with OpenSSL (well at least cause lots of warnings), but fear
3422 not: the conversion is trivial, and it eliminates loads of evil casts. A
3423 few STACKed things have been converted already. Feel free to convert more.
3424 In the fullness of time, I'll do away with the STACK type altogether.
3427 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3428 specified in <certfile> by updating the entry in the index.txt file.
3429 This way one no longer has to edit the index.txt file manually for
3430 revoking a certificate. The -revoke option does the gory details now.
3431 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3433 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3434 `-text' option at all and this way the `-noout -text' combination was
3435 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3436 [Ralf S. Engelschall]
3438 *) Make sure a corresponding plain text error message exists for the
3439 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3440 verify callback function determined that a certificate was revoked.
3441 [Ralf S. Engelschall]
3443 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3444 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3445 all available cipers including rc5, which was forgotten until now.
3446 In order to let the testing shell script know which algorithms
3447 are available, a new (up to now undocumented) command
3448 "openssl list-cipher-commands" is used.
3451 *) Bugfix: s_client occasionally would sleep in select() when
3452 it should have checked SSL_pending() first.
3455 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3456 the raw DSA values prior to ASN.1 encoding.
3459 *) Tweaks to Configure
3460 [Niels Poppe <niels@netbox.org>]
3462 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3466 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3469 *) New config option to avoid instructions that are illegal on the 80386.
3470 The default code is faster, but requires at least a 486.
3473 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3474 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3475 same as SSL2_VERSION anyway.
3478 *) New "-showcerts" option for s_client.
3481 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3482 application. Various cleanups and fixes.
3485 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3486 modify error routines to work internally. Add error codes and PBE init
3487 to library startup routines.
3490 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3491 packing functions to asn1 and evp. Changed function names and error
3492 codes along the way.
3495 *) PKCS12 integration: and so it begins... First of several patches to
3496 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3497 objects to objects.h
3500 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3501 and display support for Thawte strong extranet extension.
3504 *) Add LinuxPPC support.
3505 [Jeff Dubrule <igor@pobox.org>]
3507 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3508 bn_div_words in alpha.s.
3509 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3511 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3512 OAEP isn't supported when OpenSSL is built with RSAref.
3513 [Ulf Moeller <ulf@fitug.de>]
3515 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3516 so they no longer are missing under -DNOPROTO.
3517 [Soren S. Jorvang <soren@t.dk>]
3520 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3522 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3523 doesn't work when the session is reused. Coming soon!
3526 *) Fix a security hole, that allows sessions to be reused in the wrong
3527 context thus bypassing client cert protection! All software that uses
3528 client certs and session caches in multiple contexts NEEDS PATCHING to
3529 allow session reuse! A fuller solution is in the works.
3530 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3532 *) Some more source tree cleanups (removed obsolete files
3533 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3534 permission on "config" script to be executable) and a fix for the INSTALL
3536 [Ulf Moeller <ulf@fitug.de>]
3538 *) Remove some legacy and erroneous uses of malloc, free instead of
3540 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3542 *) Make rsa_oaep_test return non-zero on error.
3543 [Ulf Moeller <ulf@fitug.de>]
3545 *) Add support for native Solaris shared libraries. Configure
3546 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3547 if someone would make that last step automatic.
3548 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3550 *) ctx_size was not built with the right compiler during "make links". Fixed.
3553 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3554 except NULL ciphers". This means the default cipher list will no longer
3555 enable NULL ciphers. They need to be specifically enabled e.g. with
3556 the string "DEFAULT:eNULL".
3559 *) Fix to RSA private encryption routines: if p < q then it would
3560 occasionally produce an invalid result. This will only happen with
3561 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3564 *) Be less restrictive and allow also `perl util/perlpath.pl
3565 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3566 because this way one can also use an interpreter named `perl5' (which is
3567 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3568 installed as `perl').
3569 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3571 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3572 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3574 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3575 advapi32.lib to Win32 build and change the pem test comparision
3576 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3577 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3578 and crypto/des/ede_cbcm_enc.c.
3581 *) DES quad checksum was broken on big-endian architectures. Fixed.
3584 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3585 Win32 test batch file so it (might) work again. The Win32 test batch file
3586 is horrible: I feel ill....
3589 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3590 in e_os.h. Audit of header files to check ANSI and non ANSI
3591 sections: 10 functions were absent from non ANSI section and not exported
3592 from Windows DLLs. Fixed up libeay.num for new functions.
3595 *) Make `openssl version' output lines consistent.
3596 [Ralf S. Engelschall]
3598 *) Fix Win32 symbol export lists for BIO functions: Added
3599 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3600 to ms/libeay{16,32}.def.
3601 [Ralf S. Engelschall]
3603 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3604 fine under Unix and passes some trivial tests I've now added. But the
3605 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3606 added to make sure no one expects that this stuff really works in the
3607 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3608 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3610 [Ralf S. Engelschall]
3612 *) Fix the generation of two part addresses in perl.
3613 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3615 *) Add config entry for Linux on MIPS.
3616 [John Tobey <jtobey@channel1.com>]
3618 *) Make links whenever Configure is run, unless we are on Windoze.
3621 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3622 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3626 *) Add a useful kludge to allow package maintainers to specify compiler and
3627 other platforms details on the command line without having to patch the
3628 Configure script everytime: One now can use ``perl Configure
3629 <id>:<details>'', i.e. platform ids are allowed to have details appended
3630 to them (seperated by colons). This is treated as there would be a static
3631 pre-configured entry in Configure's %table under key <id> with value
3632 <details> and ``perl Configure <id>'' is called. So, when you want to
3633 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3634 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3635 now, which overrides the FreeBSD-elf entry on-the-fly.
3636 [Ralf S. Engelschall]
3638 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3641 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3642 on the `perl Configure ...' command line. This way one can compile
3643 OpenSSL libraries with Position Independent Code (PIC) which is needed
3644 for linking it into DSOs.
3645 [Ralf S. Engelschall]
3647 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3651 *) Cleaned up the LICENSE document: The official contact for any license
3652 questions now is the OpenSSL core team under openssl-core@openssl.org.
3653 And add a paragraph about the dual-license situation to make sure people
3654 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3655 to the OpenSSL toolkit.
3656 [Ralf S. Engelschall]
3658 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3659 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3660 Additonally cleaned up the `make links' target: Remove unnecessary
3661 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3662 to speed processing and no longer clutter the display with confusing
3663 stuff. Instead only the actually done links are displayed.
3664 [Ralf S. Engelschall]
3666 *) Permit null encryption ciphersuites, used for authentication only. It used
3667 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3668 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3672 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3673 signed attributes when verifying signatures (this would break them),
3674 the detached data encoding was wrong and public keys obtained using
3675 X509_get_pubkey() weren't freed.
3678 *) Add text documentation for the BUFFER functions. Also added a work around
3679 to a Win95 console bug. This was triggered by the password read stuff: the
3680 last character typed gets carried over to the next fread(). If you were
3681 generating a new cert request using 'req' for example then the last
3682 character of the passphrase would be CR which would then enter the first
3686 *) Added the new `Includes OpenSSL Cryptography Software' button as
3687 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3688 button and can be used by applications based on OpenSSL to show the
3689 relationship to the OpenSSL project.
3690 [Ralf S. Engelschall]
3692 *) Remove confusing variables in function signatures in files
3693 ssl/ssl_lib.c and ssl/ssl.h.
3694 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3696 *) Don't install bss_file.c under PREFIX/include/
3697 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3699 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3700 functions that return function pointers and has support for NT specific
3701 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3702 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3703 unsigned to signed types: this was killing the Win32 compile.
3706 *) Add new certificate file to stack functions,
3707 SSL_add_dir_cert_subjects_to_stack() and
3708 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3709 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3710 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3711 This means that Apache-SSL and similar packages don't have to mess around
3712 to add as many CAs as they want to the preferred list.
3715 *) Experiment with doxygen documentation. Currently only partially applied to
3717 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3718 openssl.doxy as the configuration file.
3721 *) Get rid of remaining C++-style comments which strict C compilers hate.
3722 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3724 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3725 compiled in by default: it has problems with large keys.
3728 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3729 DH private keys and/or callback functions which directly correspond to
3730 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3731 is needed for applications which have to configure certificates on a
3732 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3734 For the RSA certificate situation is makes no difference, but
3735 for the DSA certificate situation this fixes the "no shared cipher"
3736 problem where the OpenSSL cipher selection procedure failed because the
3737 temporary keys were not overtaken from the context and the API provided
3738 no way to reconfigure them.
3739 The new functions now let applications reconfigure the stuff and they
3740 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3741 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3742 non-public-API function ssl_cert_instantiate() is used as a helper
3743 function and also to reduce code redundancy inside ssl_rsa.c.
3744 [Ralf S. Engelschall]
3746 *) Move s_server -dcert and -dkey options out of the undocumented feature
3747 area because they are useful for the DSA situation and should be
3748 recognized by the users.
3749 [Ralf S. Engelschall]
3751 *) Fix the cipher decision scheme for export ciphers: the export bits are
3752 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3753 SSL_EXP_MASK. So, the original variable has to be used instead of the
3754 already masked variable.
3755 [Richard Levitte <levitte@stacken.kth.se>]
3757 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3758 [Richard Levitte <levitte@stacken.kth.se>]
3760 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3761 from `int' to `unsigned int' because it's a length and initialized by
3762 EVP_DigestFinal() which expects an `unsigned int *'.
3763 [Richard Levitte <levitte@stacken.kth.se>]
3765 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3766 script. Instead use the usual Shell->Perl transition trick.
3767 [Ralf S. Engelschall]
3769 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3770 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3771 -noout -modulus' as it's already the case for `openssl rsa -noout
3772 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3773 currently the public key is printed (a decision which was already done by
3774 `openssl dsa -modulus' in the past) which serves a similar purpose.
3775 Additionally the NO_RSA no longer completely removes the whole -modulus
3776 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3778 [Ralf S. Engelschall]
3780 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3781 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3782 [Arne Ansper <arne@ats.cyber.ee>]
3784 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3785 to be added. Now both 'req' and 'ca' can use new objects defined in the
3789 *) Add cool BIO that does syslog (or event log on NT).
3790 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3792 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3793 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3794 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3795 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3798 *) Add preliminary config info for new extension code.
3801 *) Make RSA_NO_PADDING really use no padding.
3802 [Ulf Moeller <ulf@fitug.de>]
3804 *) Generate errors when private/public key check is done.
3807 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3808 for some CRL extensions and new objects added.
3811 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3812 key usage extension and fuller support for authority key id.
3815 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3816 padding method for RSA, which is recommended for new applications in PKCS
3817 #1 v2.0 (RFC 2437, October 1998).
3818 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3819 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3820 against Bleichbacher's attack on RSA.
3821 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3824 *) Updates to the new SSL compression code
3825 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3827 *) Fix so that the version number in the master secret, when passed
3828 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3829 (because the server will not accept higher), that the version number
3830 is 0x03,0x01, not 0x03,0x00
3831 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3833 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3834 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3835 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3838 *) Support for RAW extensions where an arbitrary extension can be
3839 created by including its DER encoding. See apps/openssl.cnf for
3843 *) Make sure latest Perl versions don't interpret some generated C array
3844 code as Perl array code in the crypto/err/err_genc.pl script.
3845 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3847 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3848 not many people have the assembler. Various Win32 compilation fixes and
3849 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3853 *) Modify configure script 'Configure' to automatically create crypto/date.h
3854 file under Win32 and also build pem.h from pem.org. New script
3855 util/mkfiles.pl to create the MINFO file on environments that can't do a
3856 'make files': perl util/mkfiles.pl >MINFO should work.
3859 *) Major rework of DES function declarations, in the pursuit of correctness
3860 and purity. As a result, many evil casts evaporated, and some weirdness,
3861 too. You may find this causes warnings in your code. Zapping your evil
3862 casts will probably fix them. Mostly.
3865 *) Fix for a typo in asn1.h. Bug fix to object creation script
3866 obj_dat.pl. It considered a zero in an object definition to mean
3867 "end of object": none of the objects in objects.h have any zeros
3868 so it wasn't spotted.
3869 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3871 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3872 Masking (CBCM). In the absence of test vectors, the best I have been able
3873 to do is check that the decrypt undoes the encrypt, so far. Send me test
3874 vectors if you have them.
3877 *) Correct calculation of key length for export ciphers (too much space was
3878 allocated for null ciphers). This has not been tested!
3881 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3882 message is now correct (it understands "crypto" and "ssl" on its
3883 command line). There is also now an "update" option. This will update
3884 the util/ssleay.num and util/libeay.num files with any new functions.
3886 perl util/mkdef.pl crypto ssl update
3887 it will update them.
3890 *) Overhauled the Perl interface (perl/*):
3891 - ported BN stuff to OpenSSL's different BN library
3892 - made the perl/ source tree CVS-aware
3893 - renamed the package from SSLeay to OpenSSL (the files still contain
3894 their history because I've copied them in the repository)
3895 - removed obsolete files (the test scripts will be replaced
3896 by better Test::Harness variants in the future)
3897 [Ralf S. Engelschall]
3899 *) First cut for a very conservative source tree cleanup:
3900 1. merge various obsolete readme texts into doc/ssleay.txt
3901 where we collect the old documents and readme texts.
3902 2. remove the first part of files where I'm already sure that we no
3903 longer need them because of three reasons: either they are just temporary
3904 files which were left by Eric or they are preserved original files where
3905 I've verified that the diff is also available in the CVS via "cvs diff
3906 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3907 the crypto/md/ stuff).
3908 [Ralf S. Engelschall]
3910 *) More extension code. Incomplete support for subject and issuer alt
3911 name, issuer and authority key id. Change the i2v function parameters
3912 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3913 what that's for :-) Fix to ASN1 macro which messed up
3914 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3917 *) Preliminary support for ENUMERATED type. This is largely copied from the
3921 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3922 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3924 *) Make sure `make rehash' target really finds the `openssl' program.
3925 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3927 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3928 like to hear about it if this slows down other processors.
3931 *) Add CygWin32 platform information to Configure script.
3932 [Alan Batie <batie@aahz.jf.intel.com>]
3934 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3935 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3937 *) New program nseq to manipulate netscape certificate sequences
3940 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3944 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3945 but the BN code had some problems that would cause failures when
3946 doing certificate verification and some other functions.
3947 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3949 *) Add ASN1 and PEM code to support netscape certificate sequences.
3952 *) Add ASN1 and PEM code to support netscape certificate sequences.
3955 *) Add several PKIX and private extended key usage OIDs.
3958 *) Modify the 'ca' program to handle the new extension code. Modify
3959 openssl.cnf for new extension format, add comments.
3962 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3963 and add a sample to openssl.cnf so req -x509 now adds appropriate
3967 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3968 error code, add initial support to X509_print() and x509 application.
3971 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3972 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3973 stuff is currently isolated and isn't even compiled yet.
3976 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3977 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3978 Removed the versions check from X509 routines when loading extensions:
3979 this allows certain broken certificates that don't set the version
3980 properly to be processed.
3983 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3984 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3985 can still be regenerated with "make depend".
3988 *) Spelling mistake in C version of CAST-128.
3989 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3991 *) Changes to the error generation code. The perl script err-code.pl
3992 now reads in the old error codes and retains the old numbers, only
3993 adding new ones if necessary. It also only changes the .err files if new
3994 codes are added. The makefiles have been modified to only insert errors
3995 when needed (to avoid needlessly modifying header files). This is done
3996 by only inserting errors if the .err file is newer than the auto generated
3997 C file. To rebuild all the error codes from scratch (the old behaviour)
3998 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3999 or delete all the .err files.
4002 *) CAST-128 was incorrectly implemented for short keys. The C version has
4003 been fixed, but is untested. The assembler versions are also fixed, but
4004 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
4005 to regenerate it if needed.
4006 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4007 Hagino <itojun@kame.net>]
4009 *) File was opened incorrectly in randfile.c.
4010 [Ulf Möller <ulf@fitug.de>]
4012 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4013 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4014 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4015 al: it's just almost always a UTCTime. Note this patch adds new error
4016 codes so do a "make errors" if there are problems.
4019 *) Correct Linux 1 recognition in config.
4020 [Ulf Möller <ulf@fitug.de>]
4022 *) Remove pointless MD5 hash when using DSA keys in ca.
4023 [Anonymous <nobody@replay.com>]
4025 *) Generate an error if given an empty string as a cert directory. Also
4026 generate an error if handed NULL (previously returned 0 to indicate an
4027 error, but didn't set one).
4028 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4030 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4033 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4034 parameters. This was causing a warning which killed off the Win32 compile.
4037 *) Remove C++ style comments from crypto/bn/bn_local.h.
4038 [Neil Costigan <neil.costigan@celocom.com>]
4040 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4041 based on a text string, looking up short and long names and finally
4042 "dot" format. The "dot" format stuff didn't work. Added new function
4043 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4044 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4045 OID is not part of the table.
4048 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4049 X509_LOOKUP_by_alias().
4052 *) Sort openssl functions by name.
4055 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4056 encryption from sample DSA keys (in case anyone is interested the password
4060 *) Make _all_ *_free functions accept a NULL pointer.
4061 [Frans Heymans <fheymans@isaserver.be>]
4063 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4065 [Anonymous <nobody@replay.com>]
4067 *) s_server should send the CAfile as acceptable CAs, not its own cert.
4068 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4070 *) Don't blow it for numeric -newkey arguments to apps/req.
4071 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4073 *) Temp key "for export" tests were wrong in s3_srvr.c.
4074 [Anonymous <nobody@replay.com>]
4076 *) Add prototype for temp key callback functions
4077 SSL_CTX_set_tmp_{rsa,dh}_callback().
4080 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4081 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4084 *) X509_name_add_entry() freed the wrong thing after an error.
4085 [Arne Ansper <arne@ats.cyber.ee>]
4087 *) rsa_eay.c would attempt to free a NULL context.
4088 [Arne Ansper <arne@ats.cyber.ee>]
4090 *) BIO_s_socket() had a broken should_retry() on Windoze.
4091 [Arne Ansper <arne@ats.cyber.ee>]
4093 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4094 [Arne Ansper <arne@ats.cyber.ee>]
4096 *) Make sure the already existing X509_STORE->depth variable is initialized
4097 in X509_STORE_new(), but document the fact that this variable is still
4098 unused in the certificate verification process.
4099 [Ralf S. Engelschall]
4101 *) Fix the various library and apps files to free up pkeys obtained from
4102 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4105 *) Fix reference counting in X509_PUBKEY_get(). This makes
4106 demos/maurice/example2.c work, amongst others, probably.
4107 [Steve Henson and Ben Laurie]
4109 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4110 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4111 are no longer created. This way we have a single and consistent command
4112 line interface `openssl <command>', similar to `cvs <command>'.
4113 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4115 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4116 BIT STRING wrapper always have zero unused bits.
4119 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4122 *) Make the top-level INSTALL documentation easier to understand.
4125 *) Makefiles updated to exit if an error occurs in a sub-directory
4126 make (including if user presses ^C) [Paul Sutton]
4128 *) Make Montgomery context stuff explicit in RSA data structure.
4131 *) Fix build order of pem and err to allow for generated pem.h.
4134 *) Fix renumbering bug in X509_NAME_delete_entry().
4137 *) Enhanced the err-ins.pl script so it makes the error library number
4138 global and can add a library name. This is needed for external ASN1 and
4139 other error libraries.
4142 *) Fixed sk_insert which never worked properly.
4145 *) Fix ASN1 macros so they can handle indefinite length construted
4146 EXPLICIT tags. Some non standard certificates use these: they can now
4150 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4151 into a single doc/ssleay.txt bundle. This way the information is still
4152 preserved but no longer messes up this directory. Now it's new room for
4153 the new set of documenation files.
4154 [Ralf S. Engelschall]
4156 *) SETs were incorrectly DER encoded. This was a major pain, because they
4157 shared code with SEQUENCEs, which aren't coded the same. This means that
4158 almost everything to do with SETs or SEQUENCEs has either changed name or
4159 number of arguments.
4160 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4162 *) Fix test data to work with the above.
4165 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4166 was already fixed by Eric for 0.9.1 it seems.
4167 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4169 *) Autodetect FreeBSD3.
4172 *) Fix various bugs in Configure. This affects the following platforms:
4176 unixware-2.0-pentium
4180 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4181 before they are needed.
4184 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4188 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4190 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4191 changed SSLeay to OpenSSL in version strings.
4192 [Ralf S. Engelschall]
4194 *) Some fixups to the top-level documents.
4197 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4198 because the symlink to include/ was missing.
4199 [Ralf S. Engelschall]
4201 *) Incorporated the popular no-RSA/DSA-only patches
4202 which allow to compile a RSA-free SSLeay.
4203 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4205 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4206 when "ssleay" is still not found.
4207 [Ralf S. Engelschall]
4209 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4210 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4212 *) Updated the README file.
4213 [Ralf S. Engelschall]
4215 *) Added various .cvsignore files in the CVS repository subdirs
4216 to make a "cvs update" really silent.
4217 [Ralf S. Engelschall]
4219 *) Recompiled the error-definition header files and added
4220 missing symbols to the Win32 linker tables.
4221 [Ralf S. Engelschall]
4223 *) Cleaned up the top-level documents;
4224 o new files: CHANGES and LICENSE
4225 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4226 o merged COPYRIGHT into LICENSE
4227 o removed obsolete TODO file
4228 o renamed MICROSOFT to INSTALL.W32
4229 [Ralf S. Engelschall]
4231 *) Removed dummy files from the 0.9.1b source tree:
4232 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4233 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4234 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4235 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4236 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4237 [Ralf S. Engelschall]
4239 *) Added various platform portability fixes.
4242 *) The Genesis of the OpenSSL rpject:
4243 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4244 Young and Tim J. Hudson created while they were working for C2Net until
4246 [The OpenSSL Project]
4249 Changes between 0.9.0b and 0.9.1b [not released]
4251 *) Updated a few CA certificates under certs/
4254 *) Changed some BIGNUM api stuff.
4257 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4258 DGUX x86, Linux Alpha, etc.
4261 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4262 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4266 *) Add -strparse option to asn1pars program which parses nested
4268 [Dr Stephen Henson <shenson@bigfoot.com>]
4270 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4273 *) DSA fix for "ca" program.
4276 *) Added "-genkey" option to "dsaparam" program.
4279 *) Added RIPE MD160 (rmd160) message digest.
4282 *) Added -a (all) option to "ssleay version" command.
4285 *) Added PLATFORM define which is the id given to Configure.
4288 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4291 *) Extended the ASN.1 parser routines.
4294 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4297 *) Added a BN_CTX to the BN library.
4300 *) Fixed the weak key values in DES library
4303 *) Changed API in EVP library for cipher aliases.
4306 *) Added support for RC2/64bit cipher.
4309 *) Converted the lhash library to the crypto/mem.c functions.
4312 *) Added more recognized ASN.1 object ids.
4315 *) Added more RSA padding checks for SSL/TLS.
4318 *) Added BIO proxy/filter functionality.
4321 *) Added extra_certs to SSL_CTX which can be used
4322 send extra CA certificates to the client in the CA cert chain sending
4323 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4326 *) Now Fortezza is denied in the authentication phase because
4327 this is key exchange mechanism is not supported by SSLeay at all.
4330 *) Additional PKCS1 checks.
4333 *) Support the string "TLSv1" for all TLS v1 ciphers.
4336 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4337 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4340 *) Fixed a few memory leaks.
4343 *) Fixed various code and comment typos.
4346 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4347 bytes sent in the client random.
4348 [Edward Bishop <ebishop@spyglass.com>]