5 Changes between 0.9.4 and 0.9.5 [xx XXX 2000]
7 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10 *) Use BN_prime_checks_size(BN_num_bits(w)) rounds of Miller-Rabin when
11 generating DSA primes.
14 *) New function BN_pseudo_rand().
17 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
18 bignum version of BN_from_montgomery() with the working code from
19 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
23 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
24 made it impossible to use the same SSL_SESSION data structure in
25 SSL2 clients in multiple threads.
28 *) The return value of RAND_load_file() no longer counts bytes obtained
29 by stat(). RAND_load_file(..., -1) is new and uses the complete file
30 to seed the PRNG (previously an explicit byte count was required).
31 [Ulf Möller, Bodo Möller]
33 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
34 used (char *) instead of (void *) and had casts all over the place.
37 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
40 *) Retain source code compatibility for BN_prime_checks macro.
43 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
44 DH_CHECK_P_NOT_SAFE_PRIME.
45 (Check if this is true? OpenPGP calls them "strong".)
48 *) Merge the functionality of "dh" and "gendh" programs into a new program
49 "dhparam". The old programs are retained for now but will handle DH keys
50 (instead of parameters) in future.
53 *) Make the ciphers, s_server and s_client programs check the return values
54 when a new cipher list is set.
57 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
58 ciphers. Before when the 56bit ciphers were enabled the sorting was
61 The syntax for the cipher sorting has been extended to support sorting by
62 cipher-strength (using the strength_bits hard coded in the tables).
63 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
65 Fix a bug in the cipher-command parser: when supplying a cipher command
66 string with an "undefined" symbol (neither command nor alphanumeric
67 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
70 Due to the strength-sorting extension, the code of the
71 ssl_create_cipher_list() function was completely rearranged. I hope that
72 the readability was also increased :-)
73 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
75 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
76 for the first serial number and places 2 in the serial number file. This
77 avoids problems when the root CA is created with serial number zero and
78 the first user certificate has the same issuer name and serial number
82 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
83 the new code. Add documentation for this stuff.
86 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
87 X509_*() to X509at_*() on the grounds that they don't handle X509
88 structures and behave in an analagous way to the X509v3 functions:
89 they shouldn't be called directly but wrapper functions should be used
92 So we also now have some wrapper functions that call the X509at functions
93 when passed certificate requests. (TO DO: similar things can be done with
94 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
95 things. Some of these need some d2i or i2d and print functionality
96 because they handle more complex structures.)
99 *) Add missing #ifndefs that caused missing symbols when building libssl
100 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
102 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
104 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
105 has a return value which indicates the quality of the random data
106 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
107 error queue. New function RAND_pseudo_bytes() generates output that is
108 guaranteed to be unique but not unpredictable.
111 *) Do more iterations of Rabin-Miller probable prime test (specifically,
112 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
113 instead of only 2 for all lengths; see BN_prime_checks_size definition
114 in crypto/bn/bn_prime.c for the complete table). This guarantees a
115 false-positive rate of at most 2^-80 (actually less because we are
116 additionally doing trial division) for random input.
119 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
122 *) New function X509_CTX_rget_chain(), this returns the chain
123 from an X509_CTX structure with a dup of the stack and all
124 the X509 reference counts upped: so the stack will exist
125 after X509_CTX_cleanup() has been called. Modify pkcs12.c
128 Also make SSL_SESSION_print() print out the verify return
132 *) Add manpage for the pkcs12 command. Also change the default
133 behaviour so MAC iteration counts are used unless the new
134 -nomaciter option is used. This improves file security and
135 only older versions of MSIE (4.0 for example) need it.
138 *) Honor the no-xxx Configure options when creating .DEF files.
141 *) Add PKCS#10 attributes to field table: challengePassword,
142 unstructuredName and unstructuredAddress. These are taken from
143 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
144 international characters are used.
146 More changes to X509_ATTRIBUTE code: allow the setting of types
147 based on strings. Remove the 'loc' parameter when adding
148 attributes because these will be a SET OF encoding which is sorted
152 *) Initial changes to the 'req' utility to allow request generation
153 automation. This will allow an application to just generate a template
154 file containing all the field values and have req construct the
157 Initial support for X509_ATTRIBUTE handling. Stacks of these are
158 used all over the place including certificate requests and PKCS#7
159 structures. They are currently handled manually where necessary with
160 some primitive wrappers for PKCS#7. The new functions behave in a
161 manner analogous to the X509 extension functions: they allow
162 attributes to be looked up by NID and added.
164 Later something similar to the X509V3 code would be desirable to
165 automatically handle the encoding, decoding and printing of the
166 more complex types. The string types like challengePassword can
167 be handled by the string table functions.
169 Also modified the multi byte string table handling. Now there is
170 a 'global mask' which masks out certain types. The table itself
171 can use the flag STABLE_NO_MASK to ignore the mask setting: this
172 is useful when for example there is only one permissible type
173 (as in countryName) and using the mask might result in no valid
177 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
178 SSL_get_peer_finished to allow applications to obtain the latest
179 Finished messages sent to the peer or expected from the peer,
180 respectively. (SSL_get_peer_finished is usually the Finished message
181 actually received from the peer, otherwise the protocol will be aborted.)
183 As the Finished message are message digests of the complete handshake
184 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
185 be used for external authentication procedures when the authentication
186 provided by SSL/TLS is not desired or is not enough.
189 *) Enhanced support for Alpha Linux is added. Now ./config checks if
190 the host supports BWX extension and if Compaq C is present on the
191 $PATH. Just exploiting of the BWX extension results in 20-30%
192 performance kick for some algorithms, e.g. DES and RC4 to mention
193 a couple. Compaq C in turn generates ~20% faster code for MD5 and
197 *) Add support for MS "fast SGC". This is arguably a violation of the
198 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
199 weak crypto and after checking the certificate is SGC a second one
200 with strong crypto. MS SGC stops the first handshake after receiving
201 the server certificate message and sends a second client hello. Since
202 a server will typically do all the time consuming operations before
203 expecting any further messages from the client (server key exchange
204 is the most expensive) there is little difference between the two.
206 To get OpenSSL to support MS SGC we have to permit a second client
207 hello message after we have sent server done. In addition we have to
208 reset the MAC if we do get this second client hello and include the
212 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
213 if a DER encoded private key is RSA or DSA traditional format. Changed
214 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
215 format DER encoded private key. Newer code should use PKCS#8 format which
216 has the key type encoded in the ASN1 structure. Added DER private key
217 support to pkcs8 application.
220 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
221 ciphersuites has been selected (as required by the SSL 3/TLS 1
222 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
223 is set, we interpret this as a request to violate the specification
224 (the worst that can happen is a handshake failure, and 'correct'
225 behaviour would result in a handshake failure anyway).
228 *) In SSL_CTX_add_session, take into account that there might be multiple
229 SSL_SESSION structures with the same session ID (e.g. when two threads
230 concurrently obtain them from an external cache).
231 The internal cache can handle only one SSL_SESSION with a given ID,
232 so if there's a conflict, we now throw out the old one to achieve
236 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
237 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
238 some routines that use cipher OIDs: some ciphers do not have OIDs
239 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
243 *) Simplify the trust setting structure and code. Now we just have
244 two sequences of OIDs for trusted and rejected settings. These will
245 typically have values the same as the extended key usage extension
246 and any application specific purposes.
248 The trust checking code now has a default behaviour: it will just
249 check for an object with the same NID as the passed id. Functions can
250 be provided to override either the default behaviour or the behaviour
251 for a given id. SSL client, server and email already have functions
252 in place for compatibility: they check the NID and also return "trusted"
253 if the certificate is self signed.
256 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
257 traditional format into an EVP_PKEY structure.
260 *) Add a password callback function PEM_cb() which either prompts for
261 a password if usr_data is NULL or otherwise assumes it is a null
262 terminated password. Allow passwords to be passed on command line
263 environment or config files in a few more utilities.
266 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
267 keys. Add some short names for PKCS#8 PBE algorithms and allow them
268 to be specified on the command line for the pkcs8 and pkcs12 utilities.
269 Update documentation.
272 *) Support for ASN1 "NULL" type. This could be handled before by using
273 ASN1_TYPE but there wasn't any function that would try to read a NULL
274 and produce an error if it couldn't. For compatibility we also have
275 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
276 don't allocate anything because they don't need to.
279 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
281 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
283 *) Rebuild of the memory allocation routines used by OpenSSL code and
284 possibly others as well. The purpose is to make an interface that
285 provide hooks so anyone can build a separate set of allocation and
286 deallocation routines to be used by OpenSSL, for example memory
287 pool implementations, or something else, which was previously hard
288 since Malloc(), Realloc() and Free() were defined as macros having
289 the values malloc, realloc and free, respectively (except for Win32
290 compilations). The same is provided for memory debugging code.
291 OpenSSL already comes with functionality to find memory leaks, but
292 this gives people a chance to debug other memory problems.
294 With these changes, a new set of functions and macros have appeared:
296 CRYPTO_set_mem_debug_functions() [F]
297 CRYPTO_get_mem_debug_functions() [F]
298 CRYPTO_dbg_set_options() [F]
299 CRYPTO_dbg_get_options() [F]
300 CRYPTO_malloc_debug_init() [M]
302 The memory debug functions are NULL by default, unless the library
303 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
304 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
305 gives the standard debugging functions that come with OpenSSL) or
306 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
307 provided by the library user) must be used. When the standard
308 debugging functions are used, CRYPTO_dbg_set_options can be used to
309 request additional information:
310 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
311 the CRYPTO_MDEBUG_xxx macro when compiling the library.
313 Also, things like CRYPTO_set_mem_functions will always give the
314 expected result (the new set of functions is used for allocation
315 and deallocation) at all times, regardless of platform and compiler
318 To finish it up, some functions that were never use in any other
319 way than through macros have a new API and new semantic:
325 All macros of value have retained their old syntax.
326 [Richard Levitte and Bodo Moeller]
328 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
329 ordering of SMIMECapabilities wasn't in "strength order" and there
330 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
334 *) Some ASN1 types with illegal zero length encoding (INTEGER,
335 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
336 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
338 *) Merge in my S/MIME library for OpenSSL. This provides a simple
339 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
340 functionality to handle multipart/signed properly) and a utility
341 called 'smime' to call all this stuff. This is based on code I
342 originally wrote for Celo who have kindly allowed it to be
346 *) Add variants des_set_key_checked and des_set_key_unchecked of
347 des_set_key (aka des_key_sched). Global variable des_check_key
348 decides which of these is called by des_set_key; this way
349 des_check_key behaves as it always did, but applications and
350 the library itself, which was buggy for des_check_key == 1,
351 have a cleaner way to pick the version they need.
354 *) New function PKCS12_newpass() which changes the password of a
358 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
359 dynamic mix. In both cases the ids can be used as an index into the
360 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
361 functions so they accept a list of the field values and the
362 application doesn't need to directly manipulate the X509_TRUST
366 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
370 *) Modify the way the V3 extension code looks up extensions. This now
371 works in a similar way to the object code: we have some "standard"
372 extensions in a static table which is searched with OBJ_bsearch()
373 and the application can add dynamic ones if needed. The file
374 crypto/x509v3/ext_dat.h now has the info: this file needs to be
375 updated whenever a new extension is added to the core code and kept
376 in ext_nid order. There is a simple program 'tabtest.c' which checks
377 this. New extensions are not added too often so this file can readily
378 be maintained manually.
380 There are two big advantages in doing things this way. The extensions
381 can be looked up immediately and no longer need to be "added" using
382 X509V3_add_standard_extensions(): this function now does nothing.
383 [Side note: I get *lots* of email saying the extension code doesn't
384 work because people forget to call this function]
385 Also no dynamic allocation is done unless new extensions are added:
386 so if we don't add custom extensions there is no need to call
387 X509V3_EXT_cleanup().
390 *) Modify enc utility's salting as follows: make salting the default. Add a
391 magic header, so unsalted files fail gracefully instead of just decrypting
392 to garbage. This is because not salting is a big security hole, so people
393 should be discouraged from doing it.
396 *) Fixes and enhancements to the 'x509' utility. It allowed a message
397 digest to be passed on the command line but it only used this
398 parameter when signing a certificate. Modified so all relevant
399 operations are affected by the digest parameter including the
400 -fingerprint and -x509toreq options. Also -x509toreq choked if a
401 DSA key was used because it didn't fix the digest.
404 *) Initial certificate chain verify code. Currently tests the untrusted
405 certificates for consistency with the verify purpose (which is set
406 when the X509_STORE_CTX structure is set up) and checks the pathlength.
408 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
409 this is because it will reject chains with invalid extensions whereas
410 every previous version of OpenSSL and SSLeay made no checks at all.
412 Trust code: checks the root CA for the relevant trust settings. Trust
413 settings have an initial value consistent with the verify purpose: e.g.
414 if the verify purpose is for SSL client use it expects the CA to be
415 trusted for SSL client use. However the default value can be changed to
416 permit custom trust settings: one example of this would be to only trust
417 certificates from a specific "secure" set of CAs.
419 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
420 which should be used for version portability: especially since the
421 verify structure is likely to change more often now.
423 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
424 to set them. If not set then assume SSL clients will verify SSL servers
427 Two new options to the verify program: -untrusted allows a set of
428 untrusted certificates to be passed in and -purpose which sets the
429 intended purpose of the certificate. If a purpose is set then the
430 new chain verify code is used to check extension consistency.
433 *) Support for the authority information access extension.
436 *) Modify RSA and DSA PEM read routines to transparently handle
437 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
438 public keys in a format compatible with certificate
439 SubjectPublicKeyInfo structures. Unfortunately there were already
440 functions called *_PublicKey_* which used various odd formats so
441 these are retained for compatibility: however the DSA variants were
442 never in a public release so they have been deleted. Changed dsa/rsa
443 utilities to handle the new format: note no releases ever handled public
444 keys so we should be OK.
446 The primary motivation for this change is to avoid the same fiasco
447 that dogs private keys: there are several incompatible private key
448 formats some of which are standard and some OpenSSL specific and
449 require various evil hacks to allow partial transparent handling and
450 even then it doesn't work with DER formats. Given the option anything
451 other than PKCS#8 should be dumped: but the other formats have to
452 stay in the name of compatibility.
454 With public keys and the benefit of hindsight one standard format
455 is used which works with EVP_PKEY, RSA or DSA structures: though
456 it clearly returns an error if you try to read the wrong kind of key.
458 Added a -pubkey option to the 'x509' utility to output the public key.
459 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() and add
460 EVP_PKEY_rset_*() functions that do the same as the EVP_PKEY_assign_*()
461 except they up the reference count of the added key (they don't "swallow"
465 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
466 CRLs would fail if the file contained no certificates or no CRLs:
467 added a new function to read in both types and return the number
468 read: this means that if none are read it will be an error. The
469 DER versions of the certificate and CRL reader would always fail
470 because it isn't possible to mix certificates and CRLs in DER format
471 without choking one or the other routine. Changed this to just read
472 a certificate: this is the best we can do. Also modified the code
473 in apps/verify.c to take notice of return codes: it was previously
474 attempting to read in certificates from NULL pointers and ignoring
475 any errors: this is one reason why the cert and CRL reader seemed
476 to work. It doesn't check return codes from the default certificate
477 routines: these may well fail if the certificates aren't installed.
480 *) Code to support otherName option in GeneralName.
483 *) First update to verify code. Change the verify utility
484 so it warns if it is passed a self signed certificate:
485 for consistency with the normal behaviour. X509_verify
486 has been modified to it will now verify a self signed
487 certificate if *exactly* the same certificate appears
488 in the store: it was previously impossible to trust a
489 single self signed certificate. This means that:
490 openssl verify ss.pem
491 now gives a warning about a self signed certificate but
492 openssl verify -CAfile ss.pem ss.pem
496 *) For servers, store verify_result in SSL_SESSION data structure
497 (and add it to external session representation).
498 This is needed when client certificate verifications fails,
499 but an application-provided verification callback (set by
500 SSL_CTX_set_cert_verify_callback) allows accepting the session
501 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
502 but returns 1): When the session is reused, we have to set
503 ssl->verify_result to the appropriate error code to avoid
505 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
507 *) Fix a bug in the new PKCS#7 code: it didn't consider the
508 case in PKCS7_dataInit() where the signed PKCS7 structure
509 didn't contain any existing data because it was being created.
510 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
512 *) Add a salt to the key derivation routines in enc.c. This
513 forms the first 8 bytes of the encrypted file. Also add a
514 -S option to allow a salt to be input on the command line.
517 *) New function X509_cmp(). Oddly enough there wasn't a function
518 to compare two certificates. We do this by working out the SHA1
519 hash and comparing that. X509_cmp() will be needed by the trust
523 *) SSL_get1_session() is like SSL_get_session(), but increments
524 the reference count in the SSL_SESSION returned.
525 [Geoff Thorpe <geoff@eu.c2.net>]
527 *) Fix for 'req': it was adding a null to request attributes.
528 Also change the X509_LOOKUP and X509_INFO code to handle
529 certificate auxiliary information.
532 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
536 *) Add the possibility to add extra information to the memory leak
537 detecting output, to form tracebacks, showing from where each
538 allocation was originated: CRYPTO_push_info("constant string") adds
539 the string plus current file name and line number to a per-thread
540 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
541 is like calling CYRPTO_pop_info() until the stack is empty.
542 Also updated memory leak detection code to be multi-thread-safe.
545 *) Add options -text and -noout to pkcs7 utility and delete the
546 encryption options which never did anything. Update docs.
549 *) Add options to some of the utilities to allow the pass phrase
550 to be included on either the command line (not recommended on
551 OSes like Unix) or read from the environment. Update the
552 manpages and fix a few bugs.
555 *) Add a few manpages for some of the openssl commands.
558 *) Fix the -revoke option in ca. It was freeing up memory twice,
559 leaking and not finding already revoked certificates.
562 *) Extensive changes to support certificate auxiliary information.
563 This involves the use of X509_CERT_AUX structure and X509_AUX
564 functions. An X509_AUX function such as PEM_read_X509_AUX()
565 can still read in a certificate file in the usual way but it
566 will also read in any additional "auxiliary information". By
567 doing things this way a fair degree of compatibility can be
568 retained: existing certificates can have this information added
569 using the new 'x509' options.
571 Current auxiliary information includes an "alias" and some trust
572 settings. The trust settings will ultimately be used in enhanced
573 certificate chain verification routines: currently a certificate
574 can only be trusted if it is self signed and then it is trusted
578 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
579 The problem was that one of the replacement routines had not been working
580 since SSLeay releases. For now the offending routine has been replaced
581 with non-optimised assembler. Even so, this now gives around 95%
582 performance improvement for 1024 bit RSA signs.
585 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
586 handling. Most clients have the effective key size in bits equal to
587 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
588 A few however don't do this and instead use the size of the decrypted key
589 to determine the RC2 key length and the AlgorithmIdentifier to determine
590 the effective key length. In this case the effective key length can still
591 be 40 bits but the key length can be 168 bits for example. This is fixed
592 by manually forcing an RC2 key into the EVP_PKEY structure because the
593 EVP code can't currently handle unusual RC2 key sizes: it always assumes
594 the key length and effective key length are equal.
597 *) Add a bunch of functions that should simplify the creation of
598 X509_NAME structures. Now you should be able to do:
599 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
600 and have it automatically work out the correct field type and fill in
601 the structures. The more adventurous can try:
602 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
603 and it will (hopefully) work out the correct multibyte encoding.
606 *) Change the 'req' utility to use the new field handling and multibyte
607 copy routines. Before the DN field creation was handled in an ad hoc
608 way in req, ca, and x509 which was rather broken and didn't support
609 BMPStrings or UTF8Strings. Since some software doesn't implement
610 BMPStrings or UTF8Strings yet, they can be enabled using the config file
611 using the dirstring_type option. See the new comment in the default
612 openssl.cnf for more info.
615 *) Make crypto/rand/md_rand.c more robust:
616 - Assure unique random numbers after fork().
617 - Make sure that concurrent threads access the global counter and
618 md serializably so that we never lose entropy in them
619 or use exactly the same state in multiple threads.
620 Access to the large state is not always serializable because
621 the additional locking could be a performance killer, and
622 md should be large enough anyway.
625 *) New file apps/app_rand.c with commonly needed functionality
626 for handling the random seed file.
628 Use the random seed file in some applications that previously did not:
630 dsaparam -genkey (which also ignored its '-rand' option),
634 Except on systems with /dev/urandom, it is crucial to have a random
635 seed file at least for key creation, DSA signing, and for DH exchanges;
636 for RSA signatures we could do without one.
638 gendh and gendsa (unlike genrsa) used to read only the first byte
639 of each file listed in the '-rand' option. The function as previously
640 found in genrsa is now in app_rand.c and is used by all programs
641 that support '-rand'.
644 *) In RAND_write_file, use mode 0600 for creating files;
645 don't just chmod when it may be too late.
648 *) Report an error from X509_STORE_load_locations
649 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
652 *) New function ASN1_mbstring_copy() this copies a string in either
653 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
654 into an ASN1_STRING type. A mask of permissible types is passed
655 and it chooses the "minimal" type to use or an error if not type
659 *) Add function equivalents to the various macros in asn1.h. The old
660 macros are retained with an M_ prefix. Code inside the library can
661 use the M_ macros. External code (including the openssl utility)
662 should *NOT* in order to be "shared library friendly".
665 *) Add various functions that can check a certificate's extensions
666 to see if it usable for various purposes such as SSL client,
667 server or S/MIME and CAs of these types. This is currently
668 VERY EXPERIMENTAL but will ultimately be used for certificate chain
669 verification. Also added a -purpose flag to x509 utility to
670 print out all the purposes.
673 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
677 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
678 for, obtain and decode and extension and obtain its critical flag.
679 This allows all the necessary extension code to be handled in a
680 single function call.
683 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
684 platforms. See crypto/rc4/rc4_enc.c for further details.
687 *) New -noout option to asn1parse. This causes no output to be produced
688 its main use is when combined with -strparse and -out to extract data
689 from a file (which may not be in ASN.1 format).
692 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
693 when producing the local key id.
694 [Richard Levitte <levitte@stacken.kth.se>]
696 *) New option -dhparam in s_server. This allows a DH parameter file to be
697 stated explicitly. If it is not stated then it tries the first server
698 certificate file. The previous behaviour hard coded the filename
702 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
703 a public key to be input or output. For example:
704 openssl rsa -in key.pem -pubout -out pubkey.pem
705 Also added necessary DSA public key functions to handle this.
708 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
709 in the message. This was handled by allowing
710 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
711 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
713 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
714 to the end of the strings whereas this didn't. This would cause problems
715 if strings read with d2i_ASN1_bytes() were later modified.
716 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
718 *) Fix for base64 decode bug. When a base64 bio reads only one line of
719 data and it contains EOF it will end up returning an error. This is
720 caused by input 46 bytes long. The cause is due to the way base64
721 BIOs find the start of base64 encoded data. They do this by trying a
722 trial decode on each line until they find one that works. When they
723 do a flag is set and it starts again knowing it can pass all the
724 data directly through the decoder. Unfortunately it doesn't reset
725 the context it uses. This means that if EOF is reached an attempt
726 is made to pass two EOFs through the context and this causes the
727 resulting error. This can also cause other problems as well. As is
728 usual with these problems it takes *ages* to find and the fix is
729 trivial: move one line.
730 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
732 *) Ugly workaround to get s_client and s_server working under Windows. The
733 old code wouldn't work because it needed to select() on sockets and the
734 tty (for keypresses and to see if data could be written). Win32 only
735 supports select() on sockets so we select() with a 1s timeout on the
736 sockets and then see if any characters are waiting to be read, if none
737 are present then we retry, we also assume we can always write data to
738 the tty. This isn't nice because the code then blocks until we've
739 received a complete line of data and it is effectively polling the
740 keyboard at 1s intervals: however it's quite a bit better than not
741 working at all :-) A dedicated Windows application might handle this
742 with an event loop for example.
745 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
746 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
747 will be called when RSA_sign() and RSA_verify() are used. This is useful
748 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
749 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
750 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
751 This necessitated the support of an extra signature type NID_md5_sha1
752 for SSL signatures and modifications to the SSL library to use it instead
753 of calling RSA_public_decrypt() and RSA_private_encrypt().
756 *) Add new -verify -CAfile and -CApath options to the crl program, these
757 will lookup a CRL issuers certificate and verify the signature in a
758 similar way to the verify program. Tidy up the crl program so it
759 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
760 less strict. It will now permit CRL extensions even if it is not
761 a V2 CRL: this will allow it to tolerate some broken CRLs.
764 *) Initialize all non-automatic variables each time one of the openssl
765 sub-programs is started (this is necessary as they may be started
766 multiple times from the "OpenSSL>" prompt).
767 [Lennart Bang, Bodo Moeller]
769 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
770 removing all other RSA functionality (this is what NO_RSA does). This
771 is so (for example) those in the US can disable those operations covered
772 by the RSA patent while allowing storage and parsing of RSA keys and RSA
776 *) Non-copying interface to BIO pairs.
777 (still largely untested)
780 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
781 ASCII string. This was handled independently in various places before.
784 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
785 UTF8 strings a character at a time.
788 *) Use client_version from client hello to select the protocol
789 (s23_srvr.c) and for RSA client key exchange verification
790 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
793 *) Add various utility functions to handle SPKACs, these were previously
794 handled by poking round in the structure internals. Added new function
795 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
796 print, verify and generate SPKACs. Based on an original idea from
797 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
800 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
803 *) Allow the config file extension section to be overwritten on the
804 command line. Based on an original idea from Massimiliano Pala
805 <madwolf@comune.modena.it>. The new option is called -extensions
806 and can be applied to ca, req and x509. Also -reqexts to override
807 the request extensions in req and -crlexts to override the crl extensions
811 *) Add new feature to the SPKAC handling in ca. Now you can include
812 the same field multiple times by preceding it by "XXXX." for example:
815 this is the same syntax as used in the req config file.
818 *) Allow certificate extensions to be added to certificate requests. These
819 are specified in a 'req_extensions' option of the req section of the
820 config file. They can be printed out with the -text option to req but
821 are otherwise ignored at present.
824 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
825 data read consists of only the final block it would not decrypted because
826 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
827 A misplaced 'break' also meant the decrypted final block might not be
828 copied until the next read.
831 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
832 a few extra parameters to the DH structure: these will be useful if
833 for example we want the value of 'q' or implement X9.42 DH.
836 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
837 provides hooks that allow the default DSA functions or functions on a
838 "per key" basis to be replaced. This allows hardware acceleration and
839 hardware key storage to be handled without major modification to the
840 library. Also added low level modexp hooks and CRYPTO_EX structure and
841 associated functions.
844 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
845 as "read only": it can't be written to and the buffer it points to will
846 not be freed. Reading from a read only BIO is much more efficient than
847 a normal memory BIO. This was added because there are several times when
848 an area of memory needs to be read from a BIO. The previous method was
849 to create a memory BIO and write the data to it, this results in two
850 copies of the data and an O(n^2) reading algorithm. There is a new
851 function BIO_new_mem_buf() which creates a read only memory BIO from
852 an area of memory. Also modified the PKCS#7 routines to use read only
856 *) Bugfix: ssl23_get_client_hello did not work properly when called in
857 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
858 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
859 but a retry condition occured while trying to read the rest.
862 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
863 NID_pkcs7_encrypted by default: this was wrong since this should almost
864 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
865 the encrypted data type: this is a more sensible place to put it and it
866 allows the PKCS#12 code to be tidied up that duplicated this
870 *) Changed obj_dat.pl script so it takes its input and output files on
871 the command line. This should avoid shell escape redirection problems
875 *) Initial support for certificate extension requests, these are included
876 in things like Xenroll certificate requests. Included functions to allow
877 extensions to be obtained and added.
880 *) -crlf option to s_client and s_server for sending newlines as
881 CRLF (as required by many protocols).
884 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
886 *) Install libRSAglue.a when OpenSSL is built with RSAref.
887 [Ralf S. Engelschall]
889 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
890 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
892 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
896 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
897 DH parameters/keys (q is lost during that conversion, but the resulting
898 DH parameters contain its length).
900 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
901 much faster than DH_generate_parameters (which creates parameters
902 where p = 2*q + 1), and also the smaller q makes DH computations
903 much more efficient (160-bit exponentiation instead of 1024-bit
904 exponentiation); so this provides a convenient way to support DHE
905 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
906 utter importance to use
907 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
909 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
910 when such DH parameters are used, because otherwise small subgroup
911 attacks may become possible!
914 *) Avoid memory leak in i2d_DHparams.
917 *) Allow the -k option to be used more than once in the enc program:
918 this allows the same encrypted message to be read by multiple recipients.
921 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
922 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
923 it will always use the numerical form of the OID, even if it has a short
927 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
928 method only got called if p,q,dmp1,dmq1,iqmp components were present,
929 otherwise bn_mod_exp was called. In the case of hardware keys for example
930 no private key components need be present and it might store extra data
931 in the RSA structure, which cannot be accessed from bn_mod_exp.
932 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
933 private key operations.
936 *) Added support for SPARC Linux.
939 *) pem_password_cb function type incompatibly changed from
940 typedef int pem_password_cb(char *buf, int size, int rwflag);
942 ....(char *buf, int size, int rwflag, void *userdata);
943 so that applications can pass data to their callbacks:
944 The PEM[_ASN1]_{read,write}... functions and macros now take an
945 additional void * argument, which is just handed through whenever
946 the password callback is called.
947 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
949 New function SSL_CTX_set_default_passwd_cb_userdata.
951 Compatibility note: As many C implementations push function arguments
952 onto the stack in reverse order, the new library version is likely to
953 interoperate with programs that have been compiled with the old
954 pem_password_cb definition (PEM_whatever takes some data that
955 happens to be on the stack as its last argument, and the callback
956 just ignores this garbage); but there is no guarantee whatsoever that
959 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
960 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
961 problems not only on Windows, but also on some Unix platforms.
962 To avoid problematic command lines, these definitions are now in an
963 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
964 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
967 *) MIPS III/IV assembler module is reimplemented.
970 *) More DES library cleanups: remove references to srand/rand and
971 delete an unused file.
974 *) Add support for the the free Netwide assembler (NASM) under Win32,
975 since not many people have MASM (ml) and it can be hard to obtain.
976 This is currently experimental but it seems to work OK and pass all
977 the tests. Check out INSTALL.W32 for info.
980 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
981 without temporary keys kept an extra copy of the server key,
982 and connections with temporary keys did not free everything in case
986 *) New function RSA_check_key and new openssl rsa option -check
987 for verifying the consistency of RSA keys.
988 [Ulf Moeller, Bodo Moeller]
990 *) Various changes to make Win32 compile work:
991 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
992 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
993 comparison" warnings.
994 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
997 *) Add a debugging option to PKCS#5 v2 key generation function: when
998 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
999 derived keys are printed to stderr.
1002 *) Copy the flags in ASN1_STRING_dup().
1003 [Roman E. Pavlov <pre@mo.msk.ru>]
1005 *) The x509 application mishandled signing requests containing DSA
1006 keys when the signing key was also DSA and the parameters didn't match.
1008 It was supposed to omit the parameters when they matched the signing key:
1009 the verifying software was then supposed to automatically use the CA's
1010 parameters if they were absent from the end user certificate.
1012 Omitting parameters is no longer recommended. The test was also
1013 the wrong way round! This was probably due to unusual behaviour in
1014 EVP_cmp_parameters() which returns 1 if the parameters match.
1015 This meant that parameters were omitted when they *didn't* match and
1016 the certificate was useless. Certificates signed with 'ca' didn't have
1018 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
1020 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
1021 The interface is as follows:
1022 Applications can use
1023 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
1024 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
1025 "off" is now the default.
1026 The library internally uses
1027 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
1028 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
1029 to disable memory-checking temporarily.
1031 Some inconsistent states that previously were possible (and were
1032 even the default) are now avoided.
1034 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
1035 with each memory chunk allocated; this is occasionally more helpful
1036 than just having a counter.
1038 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
1040 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
1044 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
1045 which largely parallels "options", but is for changing API behaviour,
1046 whereas "options" are about protocol behaviour.
1047 Initial "mode" flags are:
1049 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
1050 a single record has been written.
1051 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
1052 retries use the same buffer location.
1053 (But all of the contents must be
1057 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
1060 *) Fix problems with no-hmac etc.
1061 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
1063 *) New functions RSA_get_default_method(), RSA_set_method() and
1064 RSA_get_method(). These allows replacement of RSA_METHODs without having
1065 to mess around with the internals of an RSA structure.
1068 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
1069 Also really enable memory leak checks in openssl.c and in some
1071 [Chad C. Mulligan, Bodo Moeller]
1073 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
1074 up the length of negative integers. This has now been simplified to just
1075 store the length when it is first determined and use it later, rather
1076 than trying to keep track of where data is copied and updating it to
1078 [Steve Henson, reported by Brien Wheeler
1079 <bwheeler@authentica-security.com>]
1081 *) Add a new function PKCS7_signatureVerify. This allows the verification
1082 of a PKCS#7 signature but with the signing certificate passed to the
1083 function itself. This contrasts with PKCS7_dataVerify which assumes the
1084 certificate is present in the PKCS#7 structure. This isn't always the
1085 case: certificates can be omitted from a PKCS#7 structure and be
1086 distributed by "out of band" means (such as a certificate database).
1089 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
1090 function prototypes in pem.h, also change util/mkdef.pl to add the
1091 necessary function names.
1094 *) mk1mf.pl (used by Windows builds) did not properly read the
1095 options set by Configure in the top level Makefile, and Configure
1096 was not even able to write more than one option correctly.
1097 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
1100 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
1101 file to be loaded from a BIO or FILE pointer. The BIO version will
1102 for example allow memory BIOs to contain config info.
1105 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
1106 Whoever hopes to achieve shared-library compatibility across versions
1107 must use this, not the compile-time macro.
1108 (Exercise 0.9.4: Which is the minimum library version required by
1110 Note: All this applies only to multi-threaded programs, others don't
1114 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
1115 through a BIO pair triggered the default case, i.e.
1116 SSLerr(...,SSL_R_UNKNOWN_STATE).
1119 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
1120 can use the SSL library even if none of the specific BIOs is
1124 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
1125 for the encoded length.
1126 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
1128 *) Add initial documentation of the X509V3 functions.
1131 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
1132 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
1133 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
1134 secure PKCS#8 private key format with a high iteration count.
1137 *) Fix determination of Perl interpreter: A perl or perl5
1138 _directory_ in $PATH was also accepted as the interpreter.
1139 [Ralf S. Engelschall]
1141 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
1142 wrong with it but it was very old and did things like calling
1143 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
1147 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
1148 to use the new extension code.
1151 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
1152 with macros. This should make it easier to change their form, add extra
1153 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
1157 *) Add to configuration table a new entry that can specify an alternative
1158 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
1159 according to Mark Crispin <MRC@Panda.COM>.
1163 *) DES CBC did not update the IV. Weird.
1166 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
1167 Changing the behaviour of the former might break existing programs --
1168 where IV updating is needed, des_ncbc_encrypt can be used.
1171 *) When bntest is run from "make test" it drives bc to check its
1172 calculations, as well as internally checking them. If an internal check
1173 fails, it needs to cause bc to give a non-zero result or make test carries
1174 on without noticing the failure. Fixed.
1177 *) DES library cleanups.
1180 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
1181 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
1182 ciphers. NOTE: although the key derivation function has been verified
1183 against some published test vectors it has not been extensively tested
1184 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
1188 *) Instead of "mkdir -p", which is not fully portable, use new
1189 Perl script "util/mkdir-p.pl".
1192 *) Rewrite the way password based encryption (PBE) is handled. It used to
1193 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
1194 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
1195 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
1196 the 'parameter' field of the AlgorithmIdentifier is passed to the
1197 underlying key generation function so it must do its own ASN1 parsing.
1198 This has also changed the EVP_PBE_CipherInit() function which now has a
1199 'parameter' argument instead of literal salt and iteration count values
1200 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
1203 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
1204 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
1205 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
1206 KEY" because this clashed with PKCS#8 unencrypted string. Since this
1207 value was just used as a "magic string" and not used directly its
1208 value doesn't matter.
1211 *) Introduce some semblance of const correctness to BN. Shame C doesn't
1215 *) "linux-sparc64" configuration (ultrapenguin).
1216 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
1217 "linux-sparc" configuration.
1218 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
1220 *) config now generates no-xxx options for missing ciphers.
1223 *) Support the EBCDIC character set (work in progress).
1224 File ebcdic.c not yet included because it has a different license.
1225 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1227 *) Support BS2000/OSD-POSIX.
1228 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
1230 *) Make callbacks for key generation use void * instead of char *.
1233 *) Make S/MIME samples compile (not yet tested).
1236 *) Additional typesafe stacks.
1239 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
1243 Changes between 0.9.3 and 0.9.3a [29 May 1999]
1245 *) New configuration variant "sco5-gcc".
1247 *) Updated some demos.
1248 [Sean O Riordain, Wade Scholine]
1250 *) Add missing BIO_free at exit of pkcs12 application.
1253 *) Fix memory leak in conf.c.
1256 *) Updates for Win32 to assembler version of MD5.
1259 *) Set #! path to perl in apps/der_chop to where we found it
1260 instead of using a fixed path.
1263 *) SHA library changes for irix64-mips4-cc.
1266 *) Improvements for VMS support.
1270 Changes between 0.9.2b and 0.9.3 [24 May 1999]
1272 *) Bignum library bug fix. IRIX 6 passes "make test" now!
1273 This also avoids the problems with SC4.2 and unpatched SC5.
1274 [Andy Polyakov <appro@fy.chalmers.se>]
1276 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
1277 These are required because of the typesafe stack would otherwise break
1278 existing code. If old code used a structure member which used to be STACK
1279 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
1280 sk_num or sk_value it would produce an error because the num, data members
1281 are not present in STACK_OF. Now it just produces a warning. sk_set
1282 replaces the old method of assigning a value to sk_value
1283 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
1284 that does this will no longer work (and should use sk_set instead) but
1285 this could be regarded as a "questionable" behaviour anyway.
1288 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
1289 correctly handle encrypted S/MIME data.
1292 *) Change type of various DES function arguments from des_cblock
1293 (which means, in function argument declarations, pointer to char)
1294 to des_cblock * (meaning pointer to array with 8 char elements),
1295 which allows the compiler to do more typechecking; it was like
1296 that back in SSLeay, but with lots of ugly casts.
1298 Introduce new type const_des_cblock.
1301 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
1302 problems: find RecipientInfo structure that matches recipient certificate
1303 and initialise the ASN1 structures properly based on passed cipher.
1306 *) Belatedly make the BN tests actually check the results.
1309 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
1310 to and from BNs: it was completely broken. New compilation option
1311 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
1312 key elements as negative integers.
1315 *) Reorganize and speed up MD5.
1316 [Andy Polyakov <appro@fy.chalmers.se>]
1319 [Richard Levitte <richard@levitte.org>]
1321 *) New option -out to asn1parse to allow the parsed structure to be
1322 output to a file. This is most useful when combined with the -strparse
1323 option to examine the output of things like OCTET STRINGS.
1326 *) Make SSL library a little more fool-proof by not requiring any longer
1327 that SSL_set_{accept,connect}_state be called before
1328 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
1329 in many applications because usually everything *appeared* to work as
1330 intended anyway -- now it really works as intended).
1333 *) Move openssl.cnf out of lib/.
1336 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
1337 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
1338 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
1339 [Ralf S. Engelschall]
1341 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
1342 handle PKCS#7 enveloped data properly.
1343 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
1345 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
1346 copying pointers. The cert_st handling is changed by this in
1347 various ways (and thus what used to be known as ctx->default_cert
1348 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
1349 any longer when s->cert does not give us what we need).
1350 ssl_cert_instantiate becomes obsolete by this change.
1351 As soon as we've got the new code right (possibly it already is?),
1352 we have solved a couple of bugs of the earlier code where s->cert
1353 was used as if it could not have been shared with other SSL structures.
1355 Note that using the SSL API in certain dirty ways now will result
1356 in different behaviour than observed with earlier library versions:
1357 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
1358 does not influence s as it used to.
1360 In order to clean up things more thoroughly, inside SSL_SESSION
1361 we don't use CERT any longer, but a new structure SESS_CERT
1362 that holds per-session data (if available); currently, this is
1363 the peer's certificate chain and, for clients, the server's certificate
1364 and temporary key. CERT holds only those values that can have
1365 meaningful defaults in an SSL_CTX.
1368 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
1369 from the internal representation. Various PKCS#7 fixes: remove some
1370 evil casts and set the enc_dig_alg field properly based on the signing
1374 *) Allow PKCS#12 password to be set from the command line or the
1375 environment. Let 'ca' get its config file name from the environment
1376 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
1380 *) Allow certificate policies extension to use an IA5STRING for the
1381 organization field. This is contrary to the PKIX definition but
1382 VeriSign uses it and IE5 only recognises this form. Document 'x509'
1386 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
1387 without disallowing inline assembler and the like for non-pedantic builds.
1390 *) Support Borland C++ builder.
1391 [Janez Jere <jj@void.si>, modified by Ulf Möller]
1396 *) SHA-1 cleanups and performance enhancements.
1397 [Andy Polyakov <appro@fy.chalmers.se>]
1399 *) Sparc v8plus assembler for the bignum library.
1400 [Andy Polyakov <appro@fy.chalmers.se>]
1402 *) Accept any -xxx and +xxx compiler options in Configure.
1405 *) Update HPUX configuration.
1408 *) Add missing sk_<type>_unshift() function to safestack.h
1409 [Ralf S. Engelschall]
1411 *) New function SSL_CTX_use_certificate_chain_file that sets the
1412 "extra_cert"s in addition to the certificate. (This makes sense
1413 only for "PEM" format files, as chains as a whole are not
1417 *) Support verify_depth from the SSL API.
1418 x509_vfy.c had what can be considered an off-by-one-error:
1419 Its depth (which was not part of the external interface)
1420 was actually counting the number of certificates in a chain;
1421 now it really counts the depth.
1424 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
1425 instead of X509err, which often resulted in confusing error
1426 messages since the error codes are not globally unique
1427 (e.g. an alleged error in ssl3_accept when a certificate
1428 didn't match the private key).
1430 *) New function SSL_CTX_set_session_id_context that allows to set a default
1431 value (so that you don't need SSL_set_session_id_context for each
1432 connection using the SSL_CTX).
1435 *) OAEP decoding bug fix.
1438 *) Support INSTALL_PREFIX for package builders, as proposed by
1442 *) New Configure options "threads" and "no-threads". For systems
1443 where the proper compiler options are known (currently Solaris
1444 and Linux), "threads" is the default.
1447 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
1450 *) Install various scripts to $(OPENSSLDIR)/misc, not to
1451 $(INSTALLTOP)/bin -- they shouldn't clutter directories
1452 such as /usr/local/bin.
1455 *) "make linux-shared" to build shared libraries.
1456 [Niels Poppe <niels@netbox.org>]
1458 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
1461 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
1462 extension adding in x509 utility.
1465 *) Remove NOPROTO sections and error code comments.
1468 *) Partial rewrite of the DEF file generator to now parse the ANSI
1472 *) New Configure options --prefix=DIR and --openssldir=DIR.
1475 *) Complete rewrite of the error code script(s). It is all now handled
1476 by one script at the top level which handles error code gathering,
1477 header rewriting and C source file generation. It should be much better
1478 than the old method: it now uses a modified version of Ulf's parser to
1479 read the ANSI prototypes in all header files (thus the old K&R definitions
1480 aren't needed for error creation any more) and do a better job of
1481 translating function codes into names. The old 'ASN1 error code imbedded
1482 in a comment' is no longer necessary and it doesn't use .err files which
1483 have now been deleted. Also the error code call doesn't have to appear all
1484 on one line (which resulted in some large lines...).
1487 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
1490 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
1491 0 (which usually indicates a closed connection), but continue reading.
1494 *) Fix some race conditions.
1497 *) Add support for CRL distribution points extension. Add Certificate
1498 Policies and CRL distribution points documentation.
1501 *) Move the autogenerated header file parts to crypto/opensslconf.h.
1504 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
1505 8 of keying material. Merlin has also confirmed interop with this fix
1506 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
1507 [Merlin Hughes <merlin@baltimore.ie>]
1509 *) Fix lots of warnings.
1510 [Richard Levitte <levitte@stacken.kth.se>]
1512 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
1513 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
1514 [Richard Levitte <levitte@stacken.kth.se>]
1516 *) Fix problems with sizeof(long) == 8.
1517 [Andy Polyakov <appro@fy.chalmers.se>]
1519 *) Change functions to ANSI C.
1522 *) Fix typos in error codes.
1523 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
1525 *) Remove defunct assembler files from Configure.
1528 *) SPARC v8 assembler BIGNUM implementation.
1529 [Andy Polyakov <appro@fy.chalmers.se>]
1531 *) Support for Certificate Policies extension: both print and set.
1532 Various additions to support the r2i method this uses.
1535 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
1536 return a const string when you are expecting an allocated buffer.
1539 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
1540 types DirectoryString and DisplayText.
1543 *) Add code to allow r2i extensions to access the configuration database,
1544 add an LHASH database driver and add several ctx helper functions.
1547 *) Fix an evil bug in bn_expand2() which caused various BN functions to
1548 fail when they extended the size of a BIGNUM.
1551 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
1552 support typesafe stack.
1555 *) Fix typo in SSL_[gs]et_options().
1556 [Nils Frostberg <nils@medcom.se>]
1558 *) Delete various functions and files that belonged to the (now obsolete)
1559 old X509V3 handling code.
1562 *) New Configure option "rsaref".
1565 *) Don't auto-generate pem.h.
1568 *) Introduce type-safe ASN.1 SETs.
1571 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
1572 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
1574 *) Introduce type-safe STACKs. This will almost certainly break lots of code
1575 that links with OpenSSL (well at least cause lots of warnings), but fear
1576 not: the conversion is trivial, and it eliminates loads of evil casts. A
1577 few STACKed things have been converted already. Feel free to convert more.
1578 In the fullness of time, I'll do away with the STACK type altogether.
1581 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
1582 specified in <certfile> by updating the entry in the index.txt file.
1583 This way one no longer has to edit the index.txt file manually for
1584 revoking a certificate. The -revoke option does the gory details now.
1585 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
1587 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
1588 `-text' option at all and this way the `-noout -text' combination was
1589 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
1590 [Ralf S. Engelschall]
1592 *) Make sure a corresponding plain text error message exists for the
1593 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
1594 verify callback function determined that a certificate was revoked.
1595 [Ralf S. Engelschall]
1597 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
1598 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
1599 all available cipers including rc5, which was forgotten until now.
1600 In order to let the testing shell script know which algorithms
1601 are available, a new (up to now undocumented) command
1602 "openssl list-cipher-commands" is used.
1605 *) Bugfix: s_client occasionally would sleep in select() when
1606 it should have checked SSL_pending() first.
1609 *) New functions DSA_do_sign and DSA_do_verify to provide access to
1610 the raw DSA values prior to ASN.1 encoding.
1613 *) Tweaks to Configure
1614 [Niels Poppe <niels@netbox.org>]
1616 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
1620 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
1623 *) New config option to avoid instructions that are illegal on the 80386.
1624 The default code is faster, but requires at least a 486.
1627 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
1628 SSL2_SERVER_VERSION (not used at all) macros, which are now the
1629 same as SSL2_VERSION anyway.
1632 *) New "-showcerts" option for s_client.
1635 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
1636 application. Various cleanups and fixes.
1639 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
1640 modify error routines to work internally. Add error codes and PBE init
1641 to library startup routines.
1644 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
1645 packing functions to asn1 and evp. Changed function names and error
1646 codes along the way.
1649 *) PKCS12 integration: and so it begins... First of several patches to
1650 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
1651 objects to objects.h
1654 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
1655 and display support for Thawte strong extranet extension.
1658 *) Add LinuxPPC support.
1659 [Jeff Dubrule <igor@pobox.org>]
1661 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
1662 bn_div_words in alpha.s.
1663 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
1665 *) Make sure the RSA OAEP test is skipped under -DRSAref because
1666 OAEP isn't supported when OpenSSL is built with RSAref.
1667 [Ulf Moeller <ulf@fitug.de>]
1669 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
1670 so they no longer are missing under -DNOPROTO.
1671 [Soren S. Jorvang <soren@t.dk>]
1674 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
1676 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
1677 doesn't work when the session is reused. Coming soon!
1680 *) Fix a security hole, that allows sessions to be reused in the wrong
1681 context thus bypassing client cert protection! All software that uses
1682 client certs and session caches in multiple contexts NEEDS PATCHING to
1683 allow session reuse! A fuller solution is in the works.
1684 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
1686 *) Some more source tree cleanups (removed obsolete files
1687 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
1688 permission on "config" script to be executable) and a fix for the INSTALL
1690 [Ulf Moeller <ulf@fitug.de>]
1692 *) Remove some legacy and erroneous uses of malloc, free instead of
1694 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
1696 *) Make rsa_oaep_test return non-zero on error.
1697 [Ulf Moeller <ulf@fitug.de>]
1699 *) Add support for native Solaris shared libraries. Configure
1700 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
1701 if someone would make that last step automatic.
1702 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
1704 *) ctx_size was not built with the right compiler during "make links". Fixed.
1707 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
1708 except NULL ciphers". This means the default cipher list will no longer
1709 enable NULL ciphers. They need to be specifically enabled e.g. with
1710 the string "DEFAULT:eNULL".
1713 *) Fix to RSA private encryption routines: if p < q then it would
1714 occasionally produce an invalid result. This will only happen with
1715 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
1718 *) Be less restrictive and allow also `perl util/perlpath.pl
1719 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
1720 because this way one can also use an interpreter named `perl5' (which is
1721 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
1722 installed as `perl').
1723 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
1725 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
1726 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
1728 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
1729 advapi32.lib to Win32 build and change the pem test comparision
1730 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
1731 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
1732 and crypto/des/ede_cbcm_enc.c.
1735 *) DES quad checksum was broken on big-endian architectures. Fixed.
1738 *) Comment out two functions in bio.h that aren't implemented. Fix up the
1739 Win32 test batch file so it (might) work again. The Win32 test batch file
1740 is horrible: I feel ill....
1743 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
1744 in e_os.h. Audit of header files to check ANSI and non ANSI
1745 sections: 10 functions were absent from non ANSI section and not exported
1746 from Windows DLLs. Fixed up libeay.num for new functions.
1749 *) Make `openssl version' output lines consistent.
1750 [Ralf S. Engelschall]
1752 *) Fix Win32 symbol export lists for BIO functions: Added
1753 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
1754 to ms/libeay{16,32}.def.
1755 [Ralf S. Engelschall]
1757 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
1758 fine under Unix and passes some trivial tests I've now added. But the
1759 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
1760 added to make sure no one expects that this stuff really works in the
1761 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
1762 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
1764 [Ralf S. Engelschall]
1766 *) Fix the generation of two part addresses in perl.
1767 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
1769 *) Add config entry for Linux on MIPS.
1770 [John Tobey <jtobey@channel1.com>]
1772 *) Make links whenever Configure is run, unless we are on Windoze.
1775 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
1776 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
1780 *) Add a useful kludge to allow package maintainers to specify compiler and
1781 other platforms details on the command line without having to patch the
1782 Configure script everytime: One now can use ``perl Configure
1783 <id>:<details>'', i.e. platform ids are allowed to have details appended
1784 to them (seperated by colons). This is treated as there would be a static
1785 pre-configured entry in Configure's %table under key <id> with value
1786 <details> and ``perl Configure <id>'' is called. So, when you want to
1787 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
1788 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
1789 now, which overrides the FreeBSD-elf entry on-the-fly.
1790 [Ralf S. Engelschall]
1792 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
1795 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
1796 on the `perl Configure ...' command line. This way one can compile
1797 OpenSSL libraries with Position Independent Code (PIC) which is needed
1798 for linking it into DSOs.
1799 [Ralf S. Engelschall]
1801 *) Remarkably, export ciphers were totally broken and no-one had noticed!
1805 *) Cleaned up the LICENSE document: The official contact for any license
1806 questions now is the OpenSSL core team under openssl-core@openssl.org.
1807 And add a paragraph about the dual-license situation to make sure people
1808 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
1809 to the OpenSSL toolkit.
1810 [Ralf S. Engelschall]
1812 *) General source tree makefile cleanups: Made `making xxx in yyy...'
1813 display consistent in the source tree and replaced `/bin/rm' by `rm'.
1814 Additonally cleaned up the `make links' target: Remove unnecessary
1815 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
1816 to speed processing and no longer clutter the display with confusing
1817 stuff. Instead only the actually done links are displayed.
1818 [Ralf S. Engelschall]
1820 *) Permit null encryption ciphersuites, used for authentication only. It used
1821 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
1822 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
1826 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
1827 signed attributes when verifying signatures (this would break them),
1828 the detached data encoding was wrong and public keys obtained using
1829 X509_get_pubkey() weren't freed.
1832 *) Add text documentation for the BUFFER functions. Also added a work around
1833 to a Win95 console bug. This was triggered by the password read stuff: the
1834 last character typed gets carried over to the next fread(). If you were
1835 generating a new cert request using 'req' for example then the last
1836 character of the passphrase would be CR which would then enter the first
1840 *) Added the new `Includes OpenSSL Cryptography Software' button as
1841 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
1842 button and can be used by applications based on OpenSSL to show the
1843 relationship to the OpenSSL project.
1844 [Ralf S. Engelschall]
1846 *) Remove confusing variables in function signatures in files
1847 ssl/ssl_lib.c and ssl/ssl.h.
1848 [Lennart Bong <lob@kulthea.stacken.kth.se>]
1850 *) Don't install bss_file.c under PREFIX/include/
1851 [Lennart Bong <lob@kulthea.stacken.kth.se>]
1853 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
1854 functions that return function pointers and has support for NT specific
1855 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
1856 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
1857 unsigned to signed types: this was killing the Win32 compile.
1860 *) Add new certificate file to stack functions,
1861 SSL_add_dir_cert_subjects_to_stack() and
1862 SSL_add_file_cert_subjects_to_stack(). These largely supplant
1863 SSL_load_client_CA_file(), and can be used to add multiple certs easily
1864 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
1865 This means that Apache-SSL and similar packages don't have to mess around
1866 to add as many CAs as they want to the preferred list.
1869 *) Experiment with doxygen documentation. Currently only partially applied to
1871 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
1872 openssl.doxy as the configuration file.
1875 *) Get rid of remaining C++-style comments which strict C compilers hate.
1876 [Ralf S. Engelschall, pointed out by Carlos Amengual]
1878 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
1879 compiled in by default: it has problems with large keys.
1882 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
1883 DH private keys and/or callback functions which directly correspond to
1884 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
1885 is needed for applications which have to configure certificates on a
1886 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
1888 For the RSA certificate situation is makes no difference, but
1889 for the DSA certificate situation this fixes the "no shared cipher"
1890 problem where the OpenSSL cipher selection procedure failed because the
1891 temporary keys were not overtaken from the context and the API provided
1892 no way to reconfigure them.
1893 The new functions now let applications reconfigure the stuff and they
1894 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
1895 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
1896 non-public-API function ssl_cert_instantiate() is used as a helper
1897 function and also to reduce code redundancy inside ssl_rsa.c.
1898 [Ralf S. Engelschall]
1900 *) Move s_server -dcert and -dkey options out of the undocumented feature
1901 area because they are useful for the DSA situation and should be
1902 recognized by the users.
1903 [Ralf S. Engelschall]
1905 *) Fix the cipher decision scheme for export ciphers: the export bits are
1906 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
1907 SSL_EXP_MASK. So, the original variable has to be used instead of the
1908 already masked variable.
1909 [Richard Levitte <levitte@stacken.kth.se>]
1911 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
1912 [Richard Levitte <levitte@stacken.kth.se>]
1914 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
1915 from `int' to `unsigned int' because it's a length and initialized by
1916 EVP_DigestFinal() which expects an `unsigned int *'.
1917 [Richard Levitte <levitte@stacken.kth.se>]
1919 *) Don't hard-code path to Perl interpreter on shebang line of Configure
1920 script. Instead use the usual Shell->Perl transition trick.
1921 [Ralf S. Engelschall]
1923 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
1924 (in addition to RSA certificates) to match the behaviour of `openssl dsa
1925 -noout -modulus' as it's already the case for `openssl rsa -noout
1926 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
1927 currently the public key is printed (a decision which was already done by
1928 `openssl dsa -modulus' in the past) which serves a similar purpose.
1929 Additionally the NO_RSA no longer completely removes the whole -modulus
1930 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
1932 [Ralf S. Engelschall]
1934 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
1935 BIO. See the source (crypto/evp/bio_ok.c) for more info.
1936 [Arne Ansper <arne@ats.cyber.ee>]
1938 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
1939 to be added. Now both 'req' and 'ca' can use new objects defined in the
1943 *) Add cool BIO that does syslog (or event log on NT).
1944 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
1946 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
1947 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
1948 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
1949 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
1952 *) Add preliminary config info for new extension code.
1955 *) Make RSA_NO_PADDING really use no padding.
1956 [Ulf Moeller <ulf@fitug.de>]
1958 *) Generate errors when private/public key check is done.
1961 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
1962 for some CRL extensions and new objects added.
1965 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
1966 key usage extension and fuller support for authority key id.
1969 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
1970 padding method for RSA, which is recommended for new applications in PKCS
1971 #1 v2.0 (RFC 2437, October 1998).
1972 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
1973 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
1974 against Bleichbacher's attack on RSA.
1975 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
1978 *) Updates to the new SSL compression code
1979 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
1981 *) Fix so that the version number in the master secret, when passed
1982 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
1983 (because the server will not accept higher), that the version number
1984 is 0x03,0x01, not 0x03,0x00
1985 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
1987 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
1988 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
1989 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
1992 *) Support for RAW extensions where an arbitrary extension can be
1993 created by including its DER encoding. See apps/openssl.cnf for
1997 *) Make sure latest Perl versions don't interpret some generated C array
1998 code as Perl array code in the crypto/err/err_genc.pl script.
1999 [Lars Weber <3weber@informatik.uni-hamburg.de>]
2001 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
2002 not many people have the assembler. Various Win32 compilation fixes and
2003 update to the INSTALL.W32 file with (hopefully) more accurate Win32
2007 *) Modify configure script 'Configure' to automatically create crypto/date.h
2008 file under Win32 and also build pem.h from pem.org. New script
2009 util/mkfiles.pl to create the MINFO file on environments that can't do a
2010 'make files': perl util/mkfiles.pl >MINFO should work.
2013 *) Major rework of DES function declarations, in the pursuit of correctness
2014 and purity. As a result, many evil casts evaporated, and some weirdness,
2015 too. You may find this causes warnings in your code. Zapping your evil
2016 casts will probably fix them. Mostly.
2019 *) Fix for a typo in asn1.h. Bug fix to object creation script
2020 obj_dat.pl. It considered a zero in an object definition to mean
2021 "end of object": none of the objects in objects.h have any zeros
2022 so it wasn't spotted.
2023 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
2025 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
2026 Masking (CBCM). In the absence of test vectors, the best I have been able
2027 to do is check that the decrypt undoes the encrypt, so far. Send me test
2028 vectors if you have them.
2031 *) Correct calculation of key length for export ciphers (too much space was
2032 allocated for null ciphers). This has not been tested!
2035 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
2036 message is now correct (it understands "crypto" and "ssl" on its
2037 command line). There is also now an "update" option. This will update
2038 the util/ssleay.num and util/libeay.num files with any new functions.
2040 perl util/mkdef.pl crypto ssl update
2041 it will update them.
2044 *) Overhauled the Perl interface (perl/*):
2045 - ported BN stuff to OpenSSL's different BN library
2046 - made the perl/ source tree CVS-aware
2047 - renamed the package from SSLeay to OpenSSL (the files still contain
2048 their history because I've copied them in the repository)
2049 - removed obsolete files (the test scripts will be replaced
2050 by better Test::Harness variants in the future)
2051 [Ralf S. Engelschall]
2053 *) First cut for a very conservative source tree cleanup:
2054 1. merge various obsolete readme texts into doc/ssleay.txt
2055 where we collect the old documents and readme texts.
2056 2. remove the first part of files where I'm already sure that we no
2057 longer need them because of three reasons: either they are just temporary
2058 files which were left by Eric or they are preserved original files where
2059 I've verified that the diff is also available in the CVS via "cvs diff
2060 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
2061 the crypto/md/ stuff).
2062 [Ralf S. Engelschall]
2064 *) More extension code. Incomplete support for subject and issuer alt
2065 name, issuer and authority key id. Change the i2v function parameters
2066 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
2067 what that's for :-) Fix to ASN1 macro which messed up
2068 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
2071 *) Preliminary support for ENUMERATED type. This is largely copied from the
2075 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
2076 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2078 *) Make sure `make rehash' target really finds the `openssl' program.
2079 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
2081 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
2082 like to hear about it if this slows down other processors.
2085 *) Add CygWin32 platform information to Configure script.
2086 [Alan Batie <batie@aahz.jf.intel.com>]
2088 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
2089 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
2091 *) New program nseq to manipulate netscape certificate sequences
2094 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
2098 *) Fixes to BN code. Previously the default was to define BN_RECURSION
2099 but the BN code had some problems that would cause failures when
2100 doing certificate verification and some other functions.
2101 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
2103 *) Add ASN1 and PEM code to support netscape certificate sequences.
2106 *) Add ASN1 and PEM code to support netscape certificate sequences.
2109 *) Add several PKIX and private extended key usage OIDs.
2112 *) Modify the 'ca' program to handle the new extension code. Modify
2113 openssl.cnf for new extension format, add comments.
2116 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
2117 and add a sample to openssl.cnf so req -x509 now adds appropriate
2121 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
2122 error code, add initial support to X509_print() and x509 application.
2125 *) Takes a deep breath and start addding X509 V3 extension support code. Add
2126 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
2127 stuff is currently isolated and isn't even compiled yet.
2130 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
2131 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
2132 Removed the versions check from X509 routines when loading extensions:
2133 this allows certain broken certificates that don't set the version
2134 properly to be processed.
2137 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
2138 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
2139 can still be regenerated with "make depend".
2142 *) Spelling mistake in C version of CAST-128.
2143 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
2145 *) Changes to the error generation code. The perl script err-code.pl
2146 now reads in the old error codes and retains the old numbers, only
2147 adding new ones if necessary. It also only changes the .err files if new
2148 codes are added. The makefiles have been modified to only insert errors
2149 when needed (to avoid needlessly modifying header files). This is done
2150 by only inserting errors if the .err file is newer than the auto generated
2151 C file. To rebuild all the error codes from scratch (the old behaviour)
2152 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
2153 or delete all the .err files.
2156 *) CAST-128 was incorrectly implemented for short keys. The C version has
2157 been fixed, but is untested. The assembler versions are also fixed, but
2158 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
2159 to regenerate it if needed.
2160 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
2161 Hagino <itojun@kame.net>]
2163 *) File was opened incorrectly in randfile.c.
2164 [Ulf Möller <ulf@fitug.de>]
2166 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
2167 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
2168 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
2169 al: it's just almost always a UTCTime. Note this patch adds new error
2170 codes so do a "make errors" if there are problems.
2173 *) Correct Linux 1 recognition in config.
2174 [Ulf Möller <ulf@fitug.de>]
2176 *) Remove pointless MD5 hash when using DSA keys in ca.
2177 [Anonymous <nobody@replay.com>]
2179 *) Generate an error if given an empty string as a cert directory. Also
2180 generate an error if handed NULL (previously returned 0 to indicate an
2181 error, but didn't set one).
2182 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
2184 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
2187 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
2188 parameters. This was causing a warning which killed off the Win32 compile.
2191 *) Remove C++ style comments from crypto/bn/bn_local.h.
2192 [Neil Costigan <neil.costigan@celocom.com>]
2194 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
2195 based on a text string, looking up short and long names and finally
2196 "dot" format. The "dot" format stuff didn't work. Added new function
2197 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
2198 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
2199 OID is not part of the table.
2202 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
2203 X509_LOOKUP_by_alias().
2206 *) Sort openssl functions by name.
2209 *) Get the gendsa program working (hopefully) and add it to app list. Remove
2210 encryption from sample DSA keys (in case anyone is interested the password
2214 *) Make _all_ *_free functions accept a NULL pointer.
2215 [Frans Heymans <fheymans@isaserver.be>]
2217 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
2219 [Anonymous <nobody@replay.com>]
2221 *) s_server should send the CAfile as acceptable CAs, not its own cert.
2222 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
2224 *) Don't blow it for numeric -newkey arguments to apps/req.
2225 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
2227 *) Temp key "for export" tests were wrong in s3_srvr.c.
2228 [Anonymous <nobody@replay.com>]
2230 *) Add prototype for temp key callback functions
2231 SSL_CTX_set_tmp_{rsa,dh}_callback().
2234 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
2235 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
2238 *) X509_name_add_entry() freed the wrong thing after an error.
2239 [Arne Ansper <arne@ats.cyber.ee>]
2241 *) rsa_eay.c would attempt to free a NULL context.
2242 [Arne Ansper <arne@ats.cyber.ee>]
2244 *) BIO_s_socket() had a broken should_retry() on Windoze.
2245 [Arne Ansper <arne@ats.cyber.ee>]
2247 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
2248 [Arne Ansper <arne@ats.cyber.ee>]
2250 *) Make sure the already existing X509_STORE->depth variable is initialized
2251 in X509_STORE_new(), but document the fact that this variable is still
2252 unused in the certificate verification process.
2253 [Ralf S. Engelschall]
2255 *) Fix the various library and apps files to free up pkeys obtained from
2256 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
2259 *) Fix reference counting in X509_PUBKEY_get(). This makes
2260 demos/maurice/example2.c work, amongst others, probably.
2261 [Steve Henson and Ben Laurie]
2263 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
2264 `openssl' and second, the shortcut symlinks for the `openssl <command>'
2265 are no longer created. This way we have a single and consistent command
2266 line interface `openssl <command>', similar to `cvs <command>'.
2267 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
2269 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
2270 BIT STRING wrapper always have zero unused bits.
2273 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
2276 *) Make the top-level INSTALL documentation easier to understand.
2279 *) Makefiles updated to exit if an error occurs in a sub-directory
2280 make (including if user presses ^C) [Paul Sutton]
2282 *) Make Montgomery context stuff explicit in RSA data structure.
2285 *) Fix build order of pem and err to allow for generated pem.h.
2288 *) Fix renumbering bug in X509_NAME_delete_entry().
2291 *) Enhanced the err-ins.pl script so it makes the error library number
2292 global and can add a library name. This is needed for external ASN1 and
2293 other error libraries.
2296 *) Fixed sk_insert which never worked properly.
2299 *) Fix ASN1 macros so they can handle indefinite length construted
2300 EXPLICIT tags. Some non standard certificates use these: they can now
2304 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
2305 into a single doc/ssleay.txt bundle. This way the information is still
2306 preserved but no longer messes up this directory. Now it's new room for
2307 the new set of documenation files.
2308 [Ralf S. Engelschall]
2310 *) SETs were incorrectly DER encoded. This was a major pain, because they
2311 shared code with SEQUENCEs, which aren't coded the same. This means that
2312 almost everything to do with SETs or SEQUENCEs has either changed name or
2313 number of arguments.
2314 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
2316 *) Fix test data to work with the above.
2319 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
2320 was already fixed by Eric for 0.9.1 it seems.
2321 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
2323 *) Autodetect FreeBSD3.
2326 *) Fix various bugs in Configure. This affects the following platforms:
2330 unixware-2.0-pentium
2334 *) Eliminate generated files from CVS. Reorder tests to regenerate files
2335 before they are needed.
2338 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
2342 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
2344 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
2345 changed SSLeay to OpenSSL in version strings.
2346 [Ralf S. Engelschall]
2348 *) Some fixups to the top-level documents.
2351 *) Fixed the nasty bug where rsaref.h was not found under compile-time
2352 because the symlink to include/ was missing.
2353 [Ralf S. Engelschall]
2355 *) Incorporated the popular no-RSA/DSA-only patches
2356 which allow to compile a RSA-free SSLeay.
2357 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
2359 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
2360 when "ssleay" is still not found.
2361 [Ralf S. Engelschall]
2363 *) Added more platforms to Configure: Cray T3E, HPUX 11,
2364 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
2366 *) Updated the README file.
2367 [Ralf S. Engelschall]
2369 *) Added various .cvsignore files in the CVS repository subdirs
2370 to make a "cvs update" really silent.
2371 [Ralf S. Engelschall]
2373 *) Recompiled the error-definition header files and added
2374 missing symbols to the Win32 linker tables.
2375 [Ralf S. Engelschall]
2377 *) Cleaned up the top-level documents;
2378 o new files: CHANGES and LICENSE
2379 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
2380 o merged COPYRIGHT into LICENSE
2381 o removed obsolete TODO file
2382 o renamed MICROSOFT to INSTALL.W32
2383 [Ralf S. Engelschall]
2385 *) Removed dummy files from the 0.9.1b source tree:
2386 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
2387 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
2388 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
2389 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
2390 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
2391 [Ralf S. Engelschall]
2393 *) Added various platform portability fixes.
2396 *) The Genesis of the OpenSSL rpject:
2397 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
2398 Young and Tim J. Hudson created while they were working for C2Net until
2400 [The OpenSSL Project]
2403 Changes between 0.9.0b and 0.9.1b [not released]
2405 *) Updated a few CA certificates under certs/
2408 *) Changed some BIGNUM api stuff.
2411 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
2412 DGUX x86, Linux Alpha, etc.
2415 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
2416 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
2420 *) Add -strparse option to asn1pars program which parses nested
2422 [Dr Stephen Henson <shenson@bigfoot.com>]
2424 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
2427 *) DSA fix for "ca" program.
2430 *) Added "-genkey" option to "dsaparam" program.
2433 *) Added RIPE MD160 (rmd160) message digest.
2436 *) Added -a (all) option to "ssleay version" command.
2439 *) Added PLATFORM define which is the id given to Configure.
2442 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
2445 *) Extended the ASN.1 parser routines.
2448 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
2451 *) Added a BN_CTX to the BN library.
2454 *) Fixed the weak key values in DES library
2457 *) Changed API in EVP library for cipher aliases.
2460 *) Added support for RC2/64bit cipher.
2463 *) Converted the lhash library to the crypto/mem.c functions.
2466 *) Added more recognized ASN.1 object ids.
2469 *) Added more RSA padding checks for SSL/TLS.
2472 *) Added BIO proxy/filter functionality.
2475 *) Added extra_certs to SSL_CTX which can be used
2476 send extra CA certificates to the client in the CA cert chain sending
2477 process. It can be configured with SSL_CTX_add_extra_chain_cert().
2480 *) Now Fortezza is denied in the authentication phase because
2481 this is key exchange mechanism is not supported by SSLeay at all.
2484 *) Additional PKCS1 checks.
2487 *) Support the string "TLSv1" for all TLS v1 ciphers.
2490 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
2491 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
2494 *) Fixed a few memory leaks.
2497 *) Fixed various code and comment typos.
2500 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
2501 bytes sent in the client random.
2502 [Edward Bishop <ebishop@spyglass.com>]