5 Changes between 0.9.6e and 0.9.6f [XX xxx XXXX]
7 *) Use proper error handling instead of 'assertions' in buffer
8 overflow checks added in 0.9.6e. This prevents DoS (the
9 assertions could call abort()).
10 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
12 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14 *) Fix cipher selection routines: ciphers without encryption had no flags
15 for the cipher strength set and where therefore not handled correctly
16 by the selection routines (PR #130).
19 *) Fix EVP_dsa_sha macro.
23 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
24 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
25 that was added in OpenSSL 0.9.6d.
27 As the countermeasure turned out to be incompatible with some
28 broken SSL implementations, the new option is part of SSL_OP_ALL.
29 SSL_OP_ALL is usually employed when compatibility with weird SSL
30 implementations is desired (e.g. '-bugs' option to 's_client' and
31 's_server'), so the new option is automatically set in many
35 *) Changes in security patch:
37 Changes marked "(CHATS)" were sponsored by the Defense Advanced
38 Research Projects Agency (DARPA) and Air Force Research Laboratory,
39 Air Force Materiel Command, USAF, under agreement number
42 *) Add various sanity checks to asn1_get_length() to reject
43 the ASN1 length bytes if they exceed sizeof(long), will appear
44 negative or the content length exceeds the length of the
46 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
48 *) Assertions for various potential buffer overflows, not known to
52 *) Various temporary buffers to hold ASCII versions of integers were
53 too small for 64 bit platforms. (CAN-2002-0655)
54 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
56 *) Remote buffer overflow in SSL3 protocol - an attacker could
57 supply an oversized session ID to a client. (CAN-2002-0656)
60 *) Remote buffer overflow in SSL2 protocol - an attacker could
61 supply an oversized client master key. (CAN-2002-0656)
64 Changes between 0.9.6c and 0.9.6d [9 May 2002]
66 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
67 encoded as NULL) with id-dsa-with-sha1.
68 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
70 *) Check various X509_...() return values in apps/req.c.
71 [Nils Larsch <nla@trustcenter.de>]
73 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
74 an end-of-file condition would erronously be flagged, when the CRLF
75 was just at the end of a processed block. The bug was discovered when
76 processing data through a buffering memory BIO handing the data to a
77 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
78 <ptsekov@syntrex.com> and Nedelcho Stanev.
81 *) Implement a countermeasure against a vulnerability recently found
82 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
83 before application data chunks to avoid the use of known IVs
84 with data potentially chosen by the attacker.
87 *) Fix length checks in ssl3_get_client_hello().
90 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
91 to prevent ssl3_read_internal() from incorrectly assuming that
92 ssl3_read_bytes() found application data while handshake
93 processing was enabled when in fact s->s3->in_read_app_data was
94 merely automatically cleared during the initial handshake.
95 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
97 *) Fix object definitions for Private and Enterprise: they were not
98 recognized in their shortname (=lowercase) representation. Extend
99 obj_dat.pl to issue an error when using undefined keywords instead
100 of silently ignoring the problem (Svenning Sorensen
101 <sss@sss.dnsalias.net>).
104 *) Fix DH_generate_parameters() so that it works for 'non-standard'
105 generators, i.e. generators other than 2 and 5. (Previously, the
106 code did not properly initialise the 'add' and 'rem' values to
107 BN_generate_prime().)
109 In the new general case, we do not insist that 'generator' is
110 actually a primitive root: This requirement is rather pointless;
111 a generator of the order-q subgroup is just as good, if not
115 *) Map new X509 verification errors to alerts. Discovered and submitted by
116 Tom Wu <tom@arcot.com>.
119 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
120 returning non-zero before the data has been completely received
121 when using non-blocking I/O.
122 [Bodo Moeller; problem pointed out by John Hughes]
124 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
125 [Ben Laurie, Lutz Jaenicke]
127 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
128 Yoram Zahavi <YoramZ@gilian.com>).
131 *) Add information about CygWin 1.3 and on, and preserve proper
132 configuration for the versions before that.
133 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
135 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
136 check whether we deal with a copy of a session and do not delete from
137 the cache in this case. Problem reported by "Izhar Shoshani Levi"
138 <izhar@checkpoint.com>.
141 *) Do not store session data into the internal session cache, if it
142 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
143 flag is set). Proposed by Aslam <aslam@funk.com>.
146 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
150 *) [In 0.9.6c-engine release:]
151 Fix a crashbug and a logic bug in hwcrhk_load_pubkey()
152 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
154 *) Add the configuration target linux-s390x.
155 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
157 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
158 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
159 variable as an indication that a ClientHello message has been
160 received. As the flag value will be lost between multiple
161 invocations of ssl3_accept when using non-blocking I/O, the
162 function may not be aware that a handshake has actually taken
163 place, thus preventing a new session from being added to the
166 To avoid this problem, we now set s->new_session to 2 instead of
167 using a local variable.
168 [Lutz Jaenicke, Bodo Moeller]
170 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
171 if the SSL_R_LENGTH_MISMATCH error is detected.
172 [Geoff Thorpe, Bodo Moeller]
174 *) New 'shared_ldflag' column in Configure platform table.
177 *) Fix EVP_CIPHER_mode macro.
178 ["Dan S. Camper" <dan@bti.net>]
180 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
181 type, we must throw them away by setting rr->length to 0.
182 [D P Chang <dpc@qualys.com>]
184 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
186 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
187 <Dominikus.Scherkl@biodata.com>. (The previous implementation
188 worked incorrectly for those cases where range = 10..._2 and
189 3*range is two bits longer than range.)
192 *) Only add signing time to PKCS7 structures if it is not already
196 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
197 OBJ_ld_ce should be OBJ_id_ce.
198 Also some ip-pda OIDs in crypto/objects/objects.txt were
199 incorrect (cf. RFC 3039).
200 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
202 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
203 returns early because it has nothing to do.
204 [Andy Schneider <andy.schneider@bjss.co.uk>]
206 *) [In 0.9.6c-engine release:]
207 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
208 [Andy Schneider <andy.schneider@bjss.co.uk>]
210 *) [In 0.9.6c-engine release:]
211 Add support for Cryptographic Appliance's keyserver technology.
212 (Use engine 'keyclient')
213 [Cryptographic Appliances and Geoff Thorpe]
215 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
216 is called via tools/c89.sh because arguments have to be
217 rearranged (all '-L' options must appear before the first object
219 [Richard Shapiro <rshapiro@abinitio.com>]
221 *) [In 0.9.6c-engine release:]
222 Add support for Broadcom crypto accelerator cards, backported
224 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
226 *) [In 0.9.6c-engine release:]
227 Add support for SureWare crypto accelerator cards from
228 Baltimore Technologies. (Use engine 'sureware')
229 [Baltimore Technologies and Mark Cox]
231 *) [In 0.9.6c-engine release:]
232 Add support for crypto accelerator cards from Accelerated
233 Encryption Processing, www.aep.ie. (Use engine 'aep')
234 [AEP Inc. and Mark Cox]
236 *) Add a configuration entry for gcc on UnixWare.
237 [Gary Benson <gbenson@redhat.com>]
239 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
240 messages are stored in a single piece (fixed-length part and
241 variable-length part combined) and fix various bugs found on the way.
244 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
245 instead. BIO_gethostbyname() does not know what timeouts are
246 appropriate, so entries would stay in cache even when they have
248 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
250 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
251 faced with a pathologically small ClientHello fragment that does
252 not contain client_version: Instead of aborting with an error,
253 simply choose the highest available protocol version (i.e.,
254 TLS 1.0 unless it is disabled). In practice, ClientHello
255 messages are never sent like this, but this change gives us
256 strictly correct behaviour at least for TLS.
259 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
260 never resets s->method to s->ctx->method when called from within
261 one of the SSL handshake functions.
262 [Bodo Moeller; problem pointed out by Niko Baric]
264 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
265 (sent using the client's version number) if client_version is
266 smaller than the protocol version in use. Also change
267 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
268 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
269 the client will at least see that alert.
272 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
276 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
277 client receives HelloRequest while in a handshake.
278 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
280 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
281 should end in 'break', not 'goto end' which circuments various
282 cleanups done in state SSL_ST_OK. But session related stuff
283 must be disabled for SSL_ST_OK in the case that we just sent a
286 Also avoid some overhead by not calling ssl_init_wbio_buffer()
287 before just sending a HelloRequest.
288 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
290 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
291 reveal whether illegal block cipher padding was found or a MAC
292 verification error occured. (Neither SSLerr() codes nor alerts
293 are directly visible to potential attackers, but the information
294 may leak via logfiles.)
296 Similar changes are not required for the SSL 2.0 implementation
297 because the number of padding bytes is sent in clear for SSL 2.0,
298 and the extra bytes are just ignored. However ssl/s2_pkt.c
299 failed to verify that the purported number of padding bytes is in
303 *) Add OpenUNIX-8 support including shared libraries
304 (Boyd Lynn Gerber <gerberb@zenez.com>).
307 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
308 'wristwatch attack' using huge encoding parameters (cf.
309 James H. Manger's CRYPTO 2001 paper). Note that the
310 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
311 encoding parameters and hence was not vulnerable.
315 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
317 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
318 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
319 followed by modular reduction.
320 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
322 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
323 equivalent based on BN_pseudo_rand() instead of BN_rand().
326 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
327 This function was broken, as the check for a new client hello message
328 to handle SGC did not allow these large messages.
329 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
332 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
335 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
336 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
339 *) Rework the configuration and shared library support for Tru64 Unix.
340 The configuration part makes use of modern compiler features and
341 still retains old compiler behavior for those that run older versions
342 of the OS. The shared library support part includes a variant that
343 uses the RPATH feature, and is available through the special
344 configuration target "alpha-cc-rpath", which will never be selected
346 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
348 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
349 with the same message size as in ssl3_get_certificate_request().
350 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
351 messages might inadvertently be reject as too long.
352 [Petr Lampa <lampa@fee.vutbr.cz>]
354 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
357 *) Modified SSL library such that the verify_callback that has been set
358 specificly for an SSL object with SSL_set_verify() is actually being
359 used. Before the change, a verify_callback set with this function was
360 ignored and the verify_callback() set in the SSL_CTX at the time of
361 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
362 to allow the necessary settings.
365 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
366 explicitly to NULL, as at least on Solaris 8 this seems not always to be
367 done automatically (in contradiction to the requirements of the C
368 standard). This made problems when used from OpenSSH.
371 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
372 dh->length and always used
374 BN_rand_range(priv_key, dh->p).
376 BN_rand_range() is not necessary for Diffie-Hellman, and this
377 specific range makes Diffie-Hellman unnecessarily inefficient if
378 dh->length (recommended exponent length) is much smaller than the
379 length of dh->p. We could use BN_rand_range() if the order of
380 the subgroup was stored in the DH structure, but we only have
385 BN_rand(priv_key, l, ...)
387 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
393 RSA_eay_public_encrypt
394 RSA_eay_private_decrypt
395 RSA_eay_private_encrypt (signing)
396 RSA_eay_public_decrypt (signature verification)
398 (default implementations for RSA_public_encrypt,
399 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
400 always reject numbers >= n.
403 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
404 to synchronize access to 'locking_thread'. This is necessary on
405 systems where access to 'locking_thread' (an 'unsigned long'
406 variable) is not atomic.
409 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
410 *before* setting the 'crypto_lock_rand' flag. The previous code had
411 a race condition if 0 is a valid thread ID.
412 [Travis Vitek <vitek@roguewave.com>]
414 *) Add support for shared libraries under Irix.
415 [Albert Chin-A-Young <china@thewrittenword.com>]
417 *) Add configuration option to build on Linux on both big-endian and
419 [Ralf Baechle <ralf@uni-koblenz.de>]
421 *) Add the possibility to create shared libraries on HP-UX.
424 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
426 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
427 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
428 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
429 PRNG state recovery was possible based on the output of
430 one PRNG request appropriately sized to gain knowledge on
431 'md' followed by enough consecutive 1-byte PRNG requests
432 to traverse all of 'state'.
434 1. When updating 'md_local' (the current thread's copy of 'md')
435 during PRNG output generation, hash all of the previous
436 'md_local' value, not just the half used for PRNG output.
438 2. Make the number of bytes from 'state' included into the hash
439 independent from the number of PRNG bytes requested.
441 The first measure alone would be sufficient to avoid
442 Markku-Juhani's attack. (Actually it had never occurred
443 to me that the half of 'md_local' used for chaining was the
444 half from which PRNG output bytes were taken -- I had always
445 assumed that the secret half would be used.) The second
446 measure makes sure that additional data from 'state' is never
447 mixed into 'md_local' in small portions; this heuristically
448 further strengthens the PRNG.
451 *) Fix crypto/bn/asm/mips3.s.
454 *) When only the key is given to "enc", the IV is undefined. Print out
455 an error message in this case.
458 *) Handle special case when X509_NAME is empty in X509 printing routines.
461 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
462 positive and less than q.
465 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
466 used: it isn't thread safe and the add_lock_callback should handle
468 [Paul Rose <Paul.Rose@bridge.com>]
470 *) Verify that incoming data obeys the block size in
471 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
475 [Ulf Möller, Bodo Möller]
477 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
478 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
479 when fixing the server behaviour for backwards-compatible 'client
480 hello' messages. (Note that the attack is impractical against
481 SSL 3.0 and TLS 1.0 anyway because length and version checking
482 means that the probability of guessing a valid ciphertext is
483 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
486 Before 0.9.5, the countermeasure (hide the error by generating a
487 random 'decryption result') did not work properly because
488 ERR_clear_error() was missing, meaning that SSL_get_error() would
489 detect the supposedly ignored error.
491 Both problems are now fixed.
494 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
495 (previously it was 1024).
498 *) Fix for compatibility mode trust settings: ignore trust settings
499 unless some valid trust or reject settings are present.
502 *) Fix for blowfish EVP: its a variable length cipher.
505 *) Fix various bugs related to DSA S/MIME verification. Handle missing
506 parameters in DSA public key structures and return an error in the
507 DSA routines if parameters are absent.
510 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
511 in the current directory if neither $RANDFILE nor $HOME was set.
512 RAND_file_name() in 0.9.6a returned NULL in this case. This has
513 caused some confusion to Windows users who haven't defined $HOME.
514 Thus RAND_file_name() is changed again: e_os.h can define a
515 DEFAULT_HOME, which will be used if $HOME is not set.
516 For Windows, we use "C:"; on other platforms, we still require
517 environment variables.
519 *) Move 'if (!initialized) RAND_poll()' into regions protected by
520 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
521 having multiple threads call RAND_poll() concurrently.
524 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
525 combination of a flag and a thread ID variable.
526 Otherwise while one thread is in ssleay_rand_bytes (which sets the
527 flag), *other* threads can enter ssleay_add_bytes without obeying
528 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
529 that they do not hold after the first thread unsets add_do_not_lock).
532 *) Change bctest again: '-x' expressions are not available in all
536 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
538 *) Fix a couple of memory leaks in PKCS7_dataDecode()
539 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
541 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
542 the default extension for executables, if any. Also, make the perl
543 scripts that use symlink() to test if it really exists and use "cp"
544 if it doesn't. All this made OpenSSL compilable and installable in
548 *) Fix for asn1_GetSequence() for indefinite length constructed data.
549 If SEQUENCE is length is indefinite just set c->slen to the total
550 amount of data available.
551 [Steve Henson, reported by shige@FreeBSD.org]
552 [This change does not apply to 0.9.7.]
554 *) Change bctest to avoid here-documents inside command substitution
555 (workaround for FreeBSD /bin/sh bug).
556 For compatibility with Ultrix, avoid shell functions (introduced
557 in the bctest version that searches along $PATH).
560 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
561 with des_encrypt() defined on some operating systems, like Solaris
565 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
566 On the Importance of Eliminating Errors in Cryptographic
567 Computations, J. Cryptology 14 (2001) 2, 101-119,
568 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
571 *) MIPS assembler BIGNUM division bug fix.
574 *) Disabled incorrect Alpha assembler code.
577 *) Fix PKCS#7 decode routines so they correctly update the length
578 after reading an EOC for the EXPLICIT tag.
580 [This change does not apply to 0.9.7.]
582 *) Fix bug in PKCS#12 key generation routines. This was triggered
583 if a 3DES key was generated with a 0 initial byte. Include
584 PKCS12_BROKEN_KEYGEN compilation option to retain the old
585 (but broken) behaviour.
588 *) Enhance bctest to search for a working bc along $PATH and print
590 [Tim Rice <tim@multitalents.net> via Richard Levitte]
592 *) Fix memory leaks in err.c: free err_data string if necessary;
593 don't write to the wrong index in ERR_set_error_data.
596 *) Implement ssl23_peek (analogous to ssl23_read), which previously
600 *) Replace rdtsc with _emit statements for VC++ version 5.
601 [Jeremy Cooper <jeremy@baymoo.org>]
603 *) Make it possible to reuse SSLv2 sessions.
606 *) In copy_email() check for >= 0 as a return value for
607 X509_NAME_get_index_by_NID() since 0 is a valid index.
608 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
610 *) Avoid coredump with unsupported or invalid public keys by checking if
611 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
612 PKCS7_verify() fails with non detached data.
615 *) Don't use getenv in library functions when run as setuid/setgid.
616 New function OPENSSL_issetugid().
619 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
620 due to incorrect handling of multi-threading:
622 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
624 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
626 3. Count how many times MemCheck_off() has been called so that
627 nested use can be treated correctly. This also avoids
628 inband-signalling in the previous code (which relied on the
629 assumption that thread ID 0 is impossible).
632 *) Add "-rand" option also to s_client and s_server.
635 *) Fix CPU detection on Irix 6.x.
636 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
637 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
639 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
642 [This change does not apply to 0.9.7.]
644 *) Use the cached encoding of an X509_NAME structure rather than
645 copying it. This is apparently the reason for the libsafe "errors"
646 but the code is actually correct.
649 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
650 Bleichenbacher's DSA attack.
651 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
652 to be set and top=0 forces the highest bit to be set; top=-1 is new
653 and leaves the highest bit random.
654 [Ulf Moeller, Bodo Moeller]
656 *) In the NCONF_...-based implementations for CONF_... queries
657 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
658 a temporary CONF structure with the data component set to NULL
659 (which gives segmentation faults in lh_retrieve).
660 Instead, use NULL for the CONF pointer in CONF_get_string and
661 CONF_get_number (which may use environment variables) and directly
662 return NULL from CONF_get_section.
665 *) Fix potential buffer overrun for EBCDIC.
668 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
669 keyUsage if basicConstraints absent for a CA.
672 *) Make SMIME_write_PKCS7() write mail header values with a format that
673 is more generally accepted (no spaces before the semicolon), since
674 some programs can't parse those values properly otherwise. Also make
675 sure BIO's that break lines after each write do not create invalid
679 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
680 macros previously used would not encode an empty SEQUENCE OF
681 and break the signature.
683 [This change does not apply to 0.9.7.]
685 *) Zero the premaster secret after deriving the master secret in
689 *) Add some EVP_add_digest_alias registrations (as found in
690 OpenSSL_add_all_digests()) to SSL_library_init()
691 aka OpenSSL_add_ssl_algorithms(). This provides improved
692 compatibility with peers using X.509 certificates
693 with unconventional AlgorithmIdentifier OIDs.
696 *) Fix for Irix with NO_ASM.
697 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
699 *) ./config script fixes.
700 [Ulf Moeller, Richard Levitte]
702 *) Fix 'openssl passwd -1'.
705 *) Change PKCS12_key_gen_asc() so it can cope with non null
706 terminated strings whose length is passed in the passlen
707 parameter, for example from PEM callbacks. This was done
708 by adding an extra length parameter to asc2uni().
709 [Steve Henson, reported by <oddissey@samsung.co.kr>]
711 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
712 call failed, free the DSA structure.
715 *) Fix to uni2asc() to cope with zero length Unicode strings.
716 These are present in some PKCS#12 files.
719 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
720 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
721 when writing a 32767 byte record.
722 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
724 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
725 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
727 (RSA objects have a reference count access to which is protected
728 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
729 so they are meant to be shared between threads.)
730 [Bodo Moeller, Geoff Thorpe; original patch submitted by
731 "Reddie, Steven" <Steven.Reddie@ca.com>]
733 *) Fix a deadlock in CRYPTO_mem_leaks().
736 *) Use better test patterns in bntest.
739 *) rand_win.c fix for Borland C.
742 *) BN_rshift bugfix for n == 0.
745 *) Add a 'bctest' script that checks for some known 'bc' bugs
746 so that 'make test' does not abort just because 'bc' is broken.
749 *) Store verify_result within SSL_SESSION also for client side to
750 avoid potential security hole. (Re-used sessions on the client side
751 always resulted in verify_result==X509_V_OK, not using the original
752 result of the server certificate verification.)
755 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
756 SSL3_RT_APPLICATION_DATA, return 0.
757 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
761 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
762 releases, have been re-implemented by renaming the previous
763 implementations of ssl2_read and ssl3_read to ssl2_read_internal
764 and ssl3_read_internal, respectively, and adding 'peek' parameters
765 to them. The new ssl[23]_{read,peek} functions are calls to
766 ssl[23]_read_internal with the 'peek' flag set appropriately.
767 A 'peek' parameter has also been added to ssl3_read_bytes, which
768 does the actual work for ssl3_read_internal.
771 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
772 the method-specific "init()" handler. Also clean up ex_data after
773 calling the method-specific "finish()" handler. Previously, this was
774 happening the other way round.
777 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
778 The previous value, 12, was not always sufficient for BN_mod_exp().
781 *) Make sure that shared libraries get the internal name engine with
782 the full version number and not just 0. This should mark the
783 shared libraries as not backward compatible. Of course, this should
784 be changed again when we can guarantee backward binary compatibility.
787 *) Fix typo in get_cert_by_subject() in by_dir.c
788 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
790 *) Rework the system to generate shared libraries:
792 - Make note of the expected extension for the shared libraries and
793 if there is a need for symbolic links from for example libcrypto.so.0
794 to libcrypto.so.0.9.7. There is extended info in Configure for
797 - Make as few rebuilds of the shared libraries as possible.
799 - Still avoid linking the OpenSSL programs with the shared libraries.
801 - When installing, install the shared libraries separately from the
805 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
807 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
808 and not in SSL_clear because the latter is also used by the
809 accept/connect functions; previously, the settings made by
810 SSL_set_read_ahead would be lost during the handshake.
811 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
813 *) Correct util/mkdef.pl to be selective about disabled algorithms.
814 Previously, it would create entries for disableed algorithms no
818 *) Added several new manual pages for SSL_* function.
821 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
823 *) In ssl23_get_client_hello, generate an error message when faced
824 with an initial SSL 3.0/TLS record that is too small to contain the
825 first two bytes of the ClientHello message, i.e. client_version.
826 (Note that this is a pathologic case that probably has never happened
827 in real life.) The previous approach was to use the version number
828 from the record header as a substitute; but our protocol choice
829 should not depend on that one because it is not authenticated
830 by the Finished messages.
833 *) More robust randomness gathering functions for Windows.
834 [Jeffrey Altman <jaltman@columbia.edu>]
836 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
837 not set then we don't setup the error code for issuer check errors
838 to avoid possibly overwriting other errors which the callback does
839 handle. If an application does set the flag then we assume it knows
840 what it is doing and can handle the new informational codes
844 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
845 a general "ANY" type, as such it should be able to decode anything
846 including tagged types. However it didn't check the class so it would
847 wrongly interpret tagged types in the same way as their universal
848 counterpart and unknown types were just rejected. Changed so that the
849 tagged and unknown types are handled in the same way as a SEQUENCE:
850 that is the encoding is stored intact. There is also a new type
851 "V_ASN1_OTHER" which is used when the class is not universal, in this
852 case we have no idea what the actual type is so we just lump them all
856 *) On VMS, stdout may very well lead to a file that is written to
857 in a record-oriented fashion. That means that every write() will
858 write a separate record, which will be read separately by the
859 programs trying to read from it. This can be very confusing.
861 The solution is to put a BIO filter in the way that will buffer
862 text until a linefeed is reached, and then write everything a
863 line at a time, so every record written will be an actual line,
864 not chunks of lines and not (usually doesn't happen, but I've
865 seen it once) several lines in one record. BIO_f_linebuffer() is
868 Currently, it's a VMS-only method, because that's where it has
869 been tested well enough.
872 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
873 it can return incorrect results.
874 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
875 but it was in 0.9.6-beta[12].)
878 *) Disable the check for content being present when verifying detached
879 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
880 include zero length content when signing messages.
883 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
884 BIO_ctrl (for BIO pairs).
887 *) Add DSO method for VMS.
890 *) Bug fix: Montgomery multiplication could produce results with the
894 *) Add RPM specification openssl.spec and modify it to build three
895 packages. The default package contains applications, application
896 documentation and run-time libraries. The devel package contains
897 include files, static libraries and function documentation. The
898 doc package contains the contents of the doc directory. The original
899 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
902 *) Add a large number of documentation files for many SSL routines.
903 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
905 *) Add a configuration entry for Sony News 4.
906 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
908 *) Don't set the two most significant bits to one when generating a
909 random number < q in the DSA library.
912 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
913 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
914 the underlying transport is blocking) if a handshake took place.
915 (The default behaviour is needed by applications such as s_client
916 and s_server that use select() to determine when to use SSL_read;
917 but for applications that know in advance when to expect data, it
918 just makes things more complicated.)
921 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
925 *) Add a few more EBCDIC conditionals that make `req' and `x509'
926 work better on such systems.
927 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
929 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
930 Update PKCS12_parse() so it copies the friendlyName and the
931 keyid to the certificates aux info.
934 *) Fix bug in PKCS7_verify() which caused an infinite loop
935 if there was more than one signature.
936 [Sven Uszpelkat <su@celocom.de>]
938 *) Major change in util/mkdef.pl to include extra information
939 about each symbol, as well as presentig variables as well
940 as functions. This change means that there's n more need
941 to rebuild the .num files when some algorithms are excluded.
944 *) Allow the verify time to be set by an application,
945 rather than always using the current time.
948 *) Phase 2 verify code reorganisation. The certificate
949 verify code now looks up an issuer certificate by a
950 number of criteria: subject name, authority key id
951 and key usage. It also verifies self signed certificates
952 by the same criteria. The main comparison function is
953 X509_check_issued() which performs these checks.
955 Lot of changes were necessary in order to support this
956 without completely rewriting the lookup code.
958 Authority and subject key identifier are now cached.
960 The LHASH 'certs' is X509_STORE has now been replaced
961 by a STACK_OF(X509_OBJECT). This is mainly because an
962 LHASH can't store or retrieve multiple objects with
965 As a result various functions (which were all internal
966 use only) have changed to handle the new X509_STORE
967 structure. This will break anything that messed round
968 with X509_STORE internally.
970 The functions X509_STORE_add_cert() now checks for an
971 exact match, rather than just subject name.
973 The X509_STORE API doesn't directly support the retrieval
974 of multiple certificates matching a given criteria, however
975 this can be worked round by performing a lookup first
976 (which will fill the cache with candidate certificates)
977 and then examining the cache for matches. This is probably
978 the best we can do without throwing out X509_LOOKUP
979 entirely (maybe later...).
981 The X509_VERIFY_CTX structure has been enhanced considerably.
983 All certificate lookup operations now go via a get_issuer()
984 callback. Although this currently uses an X509_STORE it
985 can be replaced by custom lookups. This is a simple way
986 to bypass the X509_STORE hackery necessary to make this
987 work and makes it possible to use more efficient techniques
988 in future. A very simple version which uses a simple
989 STACK for its trusted certificate store is also provided
990 using X509_STORE_CTX_trusted_stack().
992 The verify_cb() and verify() callbacks now have equivalents
993 in the X509_STORE_CTX structure.
995 X509_STORE_CTX also has a 'flags' field which can be used
996 to customise the verify behaviour.
999 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
1000 excludes S/MIME capabilities.
1003 *) When a certificate request is read in keep a copy of the
1004 original encoding of the signed data and use it when outputing
1005 again. Signatures then use the original encoding rather than
1006 a decoded, encoded version which may cause problems if the
1007 request is improperly encoded.
1010 *) For consistency with other BIO_puts implementations, call
1011 buffer_write(b, ...) directly in buffer_puts instead of calling
1014 In BIO_puts, increment b->num_write as in BIO_write.
1015 [Peter.Sylvester@EdelWeb.fr]
1017 *) Fix BN_mul_word for the case where the word is 0. (We have to use
1018 BN_zero, we may not return a BIGNUM with an array consisting of
1022 *) Avoid calling abort() from within the library when problems are
1023 detected, except if preprocessor symbols have been defined
1024 (such as REF_CHECK, BN_DEBUG etc.).
1027 *) New openssl application 'rsautl'. This utility can be
1028 used for low level RSA operations. DER public key
1029 BIO/fp routines also added.
1032 *) New Configure entry and patches for compiling on QNX 4.
1033 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
1035 *) A demo state-machine implementation was sponsored by
1036 Nuron (http://www.nuron.com/) and is now available in
1037 demos/state_machine.
1040 *) New options added to the 'dgst' utility for signature
1041 generation and verification.
1044 *) Unrecognized PKCS#7 content types are now handled via a
1045 catch all ASN1_TYPE structure. This allows unsupported
1046 types to be stored as a "blob" and an application can
1047 encode and decode it manually.
1050 *) Fix various signed/unsigned issues to make a_strex.c
1052 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
1054 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
1055 length if passed a buffer. ASN1_INTEGER_to_BN failed
1056 if passed a NULL BN and its argument was negative.
1057 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
1059 *) Modification to PKCS#7 encoding routines to output definite
1060 length encoding. Since currently the whole structures are in
1061 memory there's not real point in using indefinite length
1062 constructed encoding. However if OpenSSL is compiled with
1063 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1066 *) Added BIO_vprintf() and BIO_vsnprintf().
1069 *) Added more prefixes to parse for in the the strings written
1070 through a logging bio, to cover all the levels that are available
1071 through syslog. The prefixes are now:
1073 PANIC, EMERG, EMR => LOG_EMERG
1074 ALERT, ALR => LOG_ALERT
1075 CRIT, CRI => LOG_CRIT
1076 ERROR, ERR => LOG_ERR
1077 WARNING, WARN, WAR => LOG_WARNING
1078 NOTICE, NOTE, NOT => LOG_NOTICE
1079 INFO, INF => LOG_INFO
1080 DEBUG, DBG => LOG_DEBUG
1082 and as before, if none of those prefixes are present at the
1083 beginning of the string, LOG_ERR is chosen.
1085 On Win32, the LOG_* levels are mapped according to this:
1087 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1088 LOG_WARNING => EVENTLOG_WARNING_TYPE
1089 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1093 *) Made it possible to reconfigure with just the configuration
1094 argument "reconf" or "reconfigure". The command line arguments
1095 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1096 and are retrieved from there when reconfiguring.
1100 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1102 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1105 *) The obj_dat.pl script was messing up the sorting of object
1106 names. The reason was that it compared the quoted version
1107 of strings as a result "OCSP" > "OCSP Signing" because
1108 " > SPACE. Changed script to store unquoted versions of
1109 names and add quotes on output. It was also omitting some
1110 names from the lookup table if they were given a default
1111 value (that is if SN is missing it is given the same
1112 value as LN and vice versa), these are now added on the
1113 grounds that if an object has a name we should be able to
1114 look it up. Finally added warning output when duplicate
1115 short or long names are found.
1118 *) Changes needed for Tandem NSK.
1119 [Scott Uroff <scott@xypro.com>]
1121 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1122 RSA_padding_check_SSLv23(), special padding was never detected
1123 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1124 version rollback attacks was not effective.
1126 In s23_clnt.c, don't use special rollback-attack detection padding
1127 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1128 client; similarly, in s23_srvr.c, don't do the rollback check if
1129 SSL 2.0 is the only protocol enabled in the server.
1132 *) Make it possible to get hexdumps of unprintable data with 'openssl
1133 asn1parse'. By implication, the functions ASN1_parse_dump() and
1134 BIO_dump_indent() are added.
1137 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1138 these print out strings and name structures based on various
1139 flags including RFC2253 support and proper handling of
1140 multibyte characters. Added options to the 'x509' utility
1141 to allow the various flags to be set.
1144 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1145 Also change the functions X509_cmp_current_time() and
1146 X509_gmtime_adj() work with an ASN1_TIME structure,
1147 this will enable certificates using GeneralizedTime in validity
1148 dates to be checked.
1151 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1152 negative public key encodings) on by default,
1153 NO_NEG_PUBKEY_BUG can be set to disable it.
1156 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1157 content octets. An i2c_ASN1_OBJECT is unnecessary because
1158 the encoding can be trivially obtained from the structure.
1161 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1162 not read locks (CRYPTO_r_[un]lock).
1165 *) A first attempt at creating official support for shared
1166 libraries through configuration. I've kept it so the
1167 default is static libraries only, and the OpenSSL programs
1168 are always statically linked for now, but there are
1169 preparations for dynamic linking in place.
1170 This has been tested on Linux and Tru64.
1173 *) Randomness polling function for Win9x, as described in:
1174 Peter Gutmann, Software Generation of Practically Strong
1178 *) Fix so PRNG is seeded in req if using an already existing
1182 *) New options to smime application. -inform and -outform
1183 allow alternative formats for the S/MIME message including
1184 PEM and DER. The -content option allows the content to be
1185 specified separately. This should allow things like Netscape
1186 form signing output easier to verify.
1189 *) Fix the ASN1 encoding of tags using the 'long form'.
1192 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1193 STRING types. These convert content octets to and from the
1194 underlying type. The actual tag and length octets are
1195 already assumed to have been read in and checked. These
1196 are needed because all other string types have virtually
1197 identical handling apart from the tag. By having versions
1198 of the ASN1 functions that just operate on content octets
1199 IMPLICIT tagging can be handled properly. It also allows
1200 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1201 and ASN1_INTEGER are identical apart from the tag.
1204 *) Change the handling of OID objects as follows:
1206 - New object identifiers are inserted in objects.txt, following
1207 the syntax given in objects.README.
1208 - objects.pl is used to process obj_mac.num and create a new
1210 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1213 This is currently kind of a hack, and the perl code in objects.pl
1214 isn't very elegant, but it works as I intended. The simplest way
1215 to check that it worked correctly is to look in obj_dat.h and
1216 check the array nid_objs and make sure the objects haven't moved
1217 around (this is important!). Additions are OK, as well as
1218 consistent name changes.
1221 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1224 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1225 The given file adds to whatever has already been seeded into the
1226 random pool through the RANDFILE configuration file option or
1227 environment variable, or the default random state file.
1230 *) mkstack.pl now sorts each macro group into lexical order.
1231 Previously the output order depended on the order the files
1232 appeared in the directory, resulting in needless rewriting
1236 *) Patches to make OpenSSL compile under Win32 again. Mostly
1237 work arounds for the VC++ problem that it treats func() as
1238 func(void). Also stripped out the parts of mkdef.pl that
1239 added extra typesafe functions: these no longer exist.
1242 *) Reorganisation of the stack code. The macros are now all
1243 collected in safestack.h . Each macro is defined in terms of
1244 a "stack macro" of the form SKM_<name>(type, a, b). The
1245 DEBUG_SAFESTACK is now handled in terms of function casts,
1246 this has the advantage of retaining type safety without the
1247 use of additional functions. If DEBUG_SAFESTACK is not defined
1248 then the non typesafe macros are used instead. Also modified the
1249 mkstack.pl script to handle the new form. Needs testing to see
1250 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1251 the default if no major problems. Similar behaviour for ASN1_SET_OF
1252 and PKCS12_STACK_OF.
1255 *) When some versions of IIS use the 'NET' form of private key the
1256 key derivation algorithm is different. Normally MD5(password) is
1257 used as a 128 bit RC4 key. In the modified case
1258 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1259 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1260 as the old Netscape_RSA functions except they have an additional
1261 'sgckey' parameter which uses the modified algorithm. Also added
1262 an -sgckey command line option to the rsa utility. Thanks to
1263 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1264 algorithm to openssl-dev.
1267 *) The evp_local.h macros were using 'c.##kname' which resulted in
1268 invalid expansion on some systems (SCO 5.0.5 for example).
1269 Corrected to 'c.kname'.
1270 [Phillip Porch <root@theporch.com>]
1272 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1273 a STACK of email addresses from a certificate or request, these look
1274 in the subject name and the subject alternative name extensions and
1275 omit any duplicate addresses.
1278 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1279 This makes DSA verification about 2 % faster.
1282 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1283 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1284 plus overhead for 1024 bit moduli).
1285 This makes exponentiations about 0.5 % faster for 1024 bit
1286 exponents (as measured by "openssl speed rsa2048").
1289 *) Rename memory handling macros to avoid conflicts with other
1291 Malloc => OPENSSL_malloc
1292 Malloc_locked => OPENSSL_malloc_locked
1293 Realloc => OPENSSL_realloc
1294 Free => OPENSSL_free
1297 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1298 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1301 *) CygWin32 support.
1302 [John Jarvie <jjarvie@newsguy.com>]
1304 *) The type-safe stack code has been rejigged. It is now only compiled
1305 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1306 by default all type-specific stack functions are "#define"d back to
1307 standard stack functions. This results in more streamlined output
1308 but retains the type-safety checking possibilities of the original
1312 *) The STACK code has been cleaned up, and certain type declarations
1313 that didn't make a lot of sense have been brought in line. This has
1314 also involved a cleanup of sorts in safestack.h to more correctly
1315 map type-safe stack functions onto their plain stack counterparts.
1316 This work has also resulted in a variety of "const"ifications of
1317 lots of the code, especially "_cmp" operations which should normally
1318 be prototyped with "const" parameters anyway.
1321 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1322 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1323 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1324 where all of 'md' is used each time the PRNG is used, but 'state'
1325 is used only indexed by a cyclic counter. As entropy may not be
1326 well distributed from the beginning, 'md' is important as a
1327 chaining variable. However, the output function chains only half
1328 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1329 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1330 in all of 'state' being rewritten, with the new values depending
1331 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1334 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1335 the handshake is continued after ssl_verify_cert_chain();
1336 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1337 can lead to 'unexplainable' connection aborts later.
1338 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1340 *) Major EVP API cipher revision.
1341 Add hooks for extra EVP features. This allows various cipher
1342 parameters to be set in the EVP interface. Support added for variable
1343 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1344 setting of RC2 and RC5 parameters.
1346 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1349 Remove lots of duplicated code from the EVP library. For example *every*
1350 cipher init() function handles the 'iv' in the same way according to the
1351 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1352 for CFB and OFB modes they zero ctx->num.
1354 New functionality allows removal of S/MIME code RC2 hack.
1356 Most of the routines have the same form and so can be declared in terms
1359 By shifting this to the top level EVP_CipherInit() it can be removed from
1360 all individual ciphers. If the cipher wants to handle IVs or keys
1361 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1364 Change lots of functions like EVP_EncryptUpdate() to now return a
1365 value: although software versions of the algorithms cannot fail
1366 any installed hardware versions can.
1369 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1370 this option is set, tolerate broken clients that send the negotiated
1371 protocol version number instead of the requested protocol version
1375 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1376 i.e. non-zero for export ciphersuites, zero otherwise.
1377 Previous versions had this flag inverted, inconsistent with
1378 rsa_tmp_cb (..._TMP_RSA_CB).
1379 [Bodo Moeller; problem reported by Amit Chopra]
1381 *) Add missing DSA library text string. Work around for some IIS
1382 key files with invalid SEQUENCE encoding.
1385 *) Add a document (doc/standards.txt) that list all kinds of standards
1386 and so on that are implemented in OpenSSL.
1389 *) Enhance c_rehash script. Old version would mishandle certificates
1390 with the same subject name hash and wouldn't handle CRLs at all.
1391 Added -fingerprint option to crl utility, to support new c_rehash
1395 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1398 *) Fix for SSL server purpose checking. Server checking was
1399 rejecting certificates which had extended key usage present
1400 but no ssl client purpose.
1401 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1403 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1404 is a little unclear about how a blank password is handled.
1405 Since the password in encoded as a BMPString with terminating
1406 double NULL a zero length password would end up as just the
1407 double NULL. However no password at all is different and is
1408 handled differently in the PKCS#12 key generation code. NS
1409 treats a blank password as zero length. MSIE treats it as no
1410 password on export: but it will try both on import. We now do
1411 the same: PKCS12_parse() tries zero length and no password if
1412 the password is set to "" or NULL (NULL is now a valid password:
1413 it wasn't before) as does the pkcs12 application.
1416 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1417 perror when PEM_read_bio_X509_REQ fails, the error message must
1418 be obtained from the error queue.
1421 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1422 it in ERR_remove_state if appropriate, and change ERR_get_state
1423 accordingly to avoid race conditions (this is necessary because
1424 thread_hash is no longer constant once set).
1427 *) Bugfix for linux-elf makefile.one.
1430 *) RSA_get_default_method() will now cause a default
1431 RSA_METHOD to be chosen if one doesn't exist already.
1432 Previously this was only set during a call to RSA_new()
1433 or RSA_new_method(NULL) meaning it was possible for
1434 RSA_get_default_method() to return NULL.
1437 *) Added native name translation to the existing DSO code
1438 that will convert (if the flag to do so is set) filenames
1439 that are sufficiently small and have no path information
1440 into a canonical native form. Eg. "blah" converted to
1441 "libblah.so" or "blah.dll" etc.
1444 *) New function ERR_error_string_n(e, buf, len) which is like
1445 ERR_error_string(e, buf), but writes at most 'len' bytes
1446 including the 0 terminator. For ERR_error_string_n, 'buf'
1448 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1450 *) CONF library reworked to become more general. A new CONF
1451 configuration file reader "class" is implemented as well as a
1452 new functions (NCONF_*, for "New CONF") to handle it. The now
1453 old CONF_* functions are still there, but are reimplemented to
1454 work in terms of the new functions. Also, a set of functions
1455 to handle the internal storage of the configuration data is
1456 provided to make it easier to write new configuration file
1457 reader "classes" (I can definitely see something reading a
1458 configuration file in XML format, for example), called _CONF_*,
1459 or "the configuration storage API"...
1461 The new configuration file reading functions are:
1463 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1464 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1466 NCONF_default, NCONF_WIN32
1468 NCONF_dump_fp, NCONF_dump_bio
1470 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1471 NCONF_new creates a new CONF object. This works in the same way
1472 as other interfaces in OpenSSL, like the BIO interface.
1473 NCONF_dump_* dump the internal storage of the configuration file,
1474 which is useful for debugging. All other functions take the same
1475 arguments as the old CONF_* functions wth the exception of the
1476 first that must be a `CONF *' instead of a `LHASH *'.
1478 To make it easer to use the new classes with the old CONF_* functions,
1479 the function CONF_set_default_method is provided.
1482 *) Add '-tls1' option to 'openssl ciphers', which was already
1483 mentioned in the documentation but had not been implemented.
1484 (This option is not yet really useful because even the additional
1485 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1488 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1489 OpenSSL-based applications) load shared libraries and bind to
1490 them in a portable way.
1491 [Geoff Thorpe, with contributions from Richard Levitte]
1493 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1495 *) Make sure _lrotl and _lrotr are only used with MSVC.
1497 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1498 (the default implementation of RAND_status).
1500 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1501 to '-clrext' (= clear extensions), as intended and documented.
1502 [Bodo Moeller; inconsistency pointed out by Michael Attili
1505 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1506 was larger than the MD block size.
1507 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1509 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1510 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1511 using the passed key: if the passed key was a private key the result
1512 of X509_print(), for example, would be to print out all the private key
1516 *) des_quad_cksum() byte order bug fix.
1517 [Ulf Möller, using the problem description in krb4-0.9.7, where
1518 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1520 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1522 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1524 *) For easily testing in shell scripts whether some command
1525 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1526 returns with exit code 0 iff no command of the given name is available.
1527 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1528 the output goes to stdout and nothing is printed to stderr.
1529 Additional arguments are always ignored.
1531 Since for each cipher there is a command of the same name,
1532 the 'no-cipher' compilation switches can be tested this way.
1534 ('openssl no-XXX' is not able to detect pseudo-commands such
1535 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1538 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1541 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1542 is set; it will be thrown away anyway because each handshake creates
1544 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1545 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1546 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1547 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1550 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1551 'Q' and 'R' lose their special meanings (quit/renegotiate).
1552 This is part of what -quiet does; unlike -quiet, -ign_eof
1553 does not suppress any output.
1556 *) Add compatibility options to the purpose and trust code. The
1557 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1558 accepts a certificate or CA, this was the previous behaviour,
1559 with all the associated security issues.
1561 X509_TRUST_COMPAT is the old trust behaviour: only and
1562 automatically trust self signed roots in certificate store. A
1563 new trust setting X509_TRUST_DEFAULT is used to specify that
1564 a purpose has no associated trust setting and it should instead
1565 use the value in the default purpose.
1568 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1569 and fix a memory leak.
1572 *) In util/mkerr.pl (which implements 'make errors'), preserve
1573 reason strings from the previous version of the .c file, as
1574 the default to have only downcase letters (and digits) in
1575 automatically generated reasons codes is not always appropriate.
1578 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1579 using strerror. Previously, ERR_reason_error_string() returned
1580 library names as reason strings for SYSerr; but SYSerr is a special
1581 case where small numbers are errno values, not library numbers.
1584 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1585 converts DSA parameters into DH parameters. (When creating parameters,
1586 DSA_generate_parameters is used.)
1589 *) Include 'length' (recommended exponent length) in C code generated
1590 by 'openssl dhparam -C'.
1593 *) The second argument to set_label in perlasm was already being used
1594 so couldn't be used as a "file scope" flag. Moved to third argument
1598 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1599 instead of RAND_bytes for encryption IVs and salts.
1602 *) Include RAND_status() into RAND_METHOD instead of implementing
1603 it only for md_rand.c Otherwise replacing the PRNG by calling
1604 RAND_set_rand_method would be impossible.
1607 *) Don't let DSA_generate_key() enter an infinite loop if the random
1608 number generation fails.
1611 *) New 'rand' application for creating pseudo-random output.
1614 *) Added configuration support for Linux/IA64
1615 [Rolf Haberrecker <rolf@suse.de>]
1617 *) Assembler module support for Mingw32.
1620 *) Shared library support for HPUX (in shlib/).
1621 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1623 *) Shared library support for Solaris gcc.
1624 [Lutz Behnke <behnke@trustcenter.de>]
1626 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1628 *) PKCS7_encrypt() was adding text MIME headers twice because they
1629 were added manually and by SMIME_crlf_copy().
1632 *) In bntest.c don't call BN_rand with zero bits argument.
1633 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1635 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1636 case was implemented. This caused BN_div_recp() to fail occasionally.
1639 *) Add an optional second argument to the set_label() in the perl
1640 assembly language builder. If this argument exists and is set
1641 to 1 it signals that the assembler should use a symbol whose
1642 scope is the entire file, not just the current function. This
1643 is needed with MASM which uses the format label:: for this scope.
1644 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1646 *) Change the ASN1 types so they are typedefs by default. Before
1647 almost all types were #define'd to ASN1_STRING which was causing
1648 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1652 *) Change names of new functions to the new get1/get0 naming
1653 convention: After 'get1', the caller owns a reference count
1654 and has to call ..._free; 'get0' returns a pointer to some
1655 data structure without incrementing reference counters.
1656 (Some of the existing 'get' functions increment a reference
1657 counter, some don't.)
1658 Similarly, 'set1' and 'add1' functions increase reference
1659 counters or duplicate objects.
1662 *) Allow for the possibility of temp RSA key generation failure:
1663 the code used to assume it always worked and crashed on failure.
1666 *) Fix potential buffer overrun problem in BIO_printf().
1667 [Ulf Möller, using public domain code by Patrick Powell; problem
1668 pointed out by David Sacerdote <das33@cornell.edu>]
1670 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1671 RAND_egd() and RAND_status(). In the command line application,
1672 the EGD socket can be specified like a seed file using RANDFILE
1676 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1677 Some CAs (e.g. Verisign) distribute certificates in this form.
1680 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1681 list to exclude them. This means that no special compilation option
1682 is needed to use anonymous DH: it just needs to be included in the
1686 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1687 EVP_MD_type. The old functionality is available in a new macro called
1688 EVP_MD_md(). Change code that uses it and update docs.
1691 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1692 where the 'void *' argument is replaced by a function pointer argument.
1693 Previously 'void *' was abused to point to functions, which works on
1694 many platforms, but is not correct. As these functions are usually
1695 called by macros defined in OpenSSL header files, most source code
1696 should work without changes.
1699 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1700 sections with information on -D... compiler switches used for
1701 compiling the library so that applications can see them. To enable
1702 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1703 must be defined. E.g.,
1704 #define OPENSSL_ALGORITHM_DEFINES
1705 #include <openssl/opensslconf.h>
1706 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1707 [Richard Levitte, Ulf and Bodo Möller]
1709 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1713 *) Change the 'other' type in certificate aux info to a STACK_OF
1714 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1715 the required ASN1 format: arbitrary types determined by an OID.
1718 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1719 argument to 'req'. This is not because the function is newer or
1720 better than others it just uses the work 'NEW' in the certificate
1721 request header lines. Some software needs this.
1724 *) Reorganise password command line arguments: now passwords can be
1725 obtained from various sources. Delete the PEM_cb function and make
1726 it the default behaviour: i.e. if the callback is NULL and the
1727 usrdata argument is not NULL interpret it as a null terminated pass
1728 phrase. If usrdata and the callback are NULL then the pass phrase
1729 is prompted for as usual.
1732 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1733 the support is automatically enabled. The resulting binaries will
1734 autodetect the card and use it if present.
1735 [Ben Laurie and Compaq Inc.]
1737 *) Work around for Netscape hang bug. This sends certificate request
1738 and server done in one record. Since this is perfectly legal in the
1739 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1740 the bugs/SSLv3 entry for more info.
1743 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1746 *) Add -rand argument to smime and pkcs12 applications and read/write
1750 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1753 *) Add command line password options to the remaining applications.
1756 *) Bug fix for BN_div_recp() for numerators with an even number of
1760 *) More tests in bntest.c, and changed test_bn output.
1763 *) ./config recognizes MacOS X now.
1766 *) Bug fix for BN_div() when the first words of num and divsor are
1767 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1770 *) Add support for various broken PKCS#8 formats, and command line
1771 options to produce them.
1774 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1775 get temporary BIGNUMs from a BN_CTX.
1778 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1782 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1783 include a #define from the old name to the new. The original intent
1784 was that statically linked binaries could for example just call
1785 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1786 link with digests. This never worked becayse SSLeay_add_all_digests()
1787 and SSLeay_add_all_ciphers() were in the same source file so calling
1788 one would link with the other. They are now in separate source files.
1791 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1794 *) Use a less unusual form of the Miller-Rabin primality test (it used
1795 a binary algorithm for exponentiation integrated into the Miller-Rabin
1796 loop, our standard modexp algorithms are faster).
1799 *) Support for the EBCDIC character set completed.
1800 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1802 *) Source code cleanups: use const where appropriate, eliminate casts,
1803 use void * instead of char * in lhash.
1806 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1807 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1808 this the server could overwrite ephemeral keys that the client
1812 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1813 using 50 iterations of the Rabin-Miller test.
1815 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1816 iterations of the Rabin-Miller test as required by the appendix
1817 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1818 As BN_is_prime_fasttest includes trial division, DSA parameter
1819 generation becomes much faster.
1821 This implies a change for the callback functions in DSA_is_prime
1822 and DSA_generate_parameters: The callback function is called once
1823 for each positive witness in the Rabin-Miller test, not just
1824 occasionally in the inner loop; and the parameters to the
1825 callback function now provide an iteration count for the outer
1826 loop rather than for the current invocation of the inner loop.
1827 DSA_generate_parameters additionally can call the callback
1828 function with an 'iteration count' of -1, meaning that a
1829 candidate has passed the trial division test (when q is generated
1830 from an application-provided seed, trial division is skipped).
1833 *) New function BN_is_prime_fasttest that optionally does trial
1834 division before starting the Rabin-Miller test and has
1835 an additional BN_CTX * argument (whereas BN_is_prime always
1836 has to allocate at least one BN_CTX).
1837 'callback(1, -1, cb_arg)' is called when a number has passed the
1838 trial division stage.
1841 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1845 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1848 *) New function BN_pseudo_rand().
1851 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1852 bignum version of BN_from_montgomery() with the working code from
1853 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1857 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1858 made it impossible to use the same SSL_SESSION data structure in
1859 SSL2 clients in multiple threads.
1862 *) The return value of RAND_load_file() no longer counts bytes obtained
1863 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1864 to seed the PRNG (previously an explicit byte count was required).
1865 [Ulf Möller, Bodo Möller]
1867 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1868 used (char *) instead of (void *) and had casts all over the place.
1871 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1874 *) Retain source code compatibility for BN_prime_checks macro:
1875 BN_is_prime(..., BN_prime_checks, ...) now uses
1876 BN_prime_checks_for_size to determine the appropriate number of
1877 Rabin-Miller iterations.
1880 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1881 DH_CHECK_P_NOT_SAFE_PRIME.
1882 (Check if this is true? OpenPGP calls them "strong".)
1885 *) Merge the functionality of "dh" and "gendh" programs into a new program
1886 "dhparam". The old programs are retained for now but will handle DH keys
1887 (instead of parameters) in future.
1890 *) Make the ciphers, s_server and s_client programs check the return values
1891 when a new cipher list is set.
1894 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1895 ciphers. Before when the 56bit ciphers were enabled the sorting was
1898 The syntax for the cipher sorting has been extended to support sorting by
1899 cipher-strength (using the strength_bits hard coded in the tables).
1900 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1902 Fix a bug in the cipher-command parser: when supplying a cipher command
1903 string with an "undefined" symbol (neither command nor alphanumeric
1904 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1905 an error is flagged.
1907 Due to the strength-sorting extension, the code of the
1908 ssl_create_cipher_list() function was completely rearranged. I hope that
1909 the readability was also increased :-)
1910 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1912 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1913 for the first serial number and places 2 in the serial number file. This
1914 avoids problems when the root CA is created with serial number zero and
1915 the first user certificate has the same issuer name and serial number
1919 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1920 the new code. Add documentation for this stuff.
1923 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1924 X509_*() to X509at_*() on the grounds that they don't handle X509
1925 structures and behave in an analagous way to the X509v3 functions:
1926 they shouldn't be called directly but wrapper functions should be used
1929 So we also now have some wrapper functions that call the X509at functions
1930 when passed certificate requests. (TO DO: similar things can be done with
1931 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1932 things. Some of these need some d2i or i2d and print functionality
1933 because they handle more complex structures.)
1936 *) Add missing #ifndefs that caused missing symbols when building libssl
1937 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1938 NO_RSA in ssl/s2*.c.
1939 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1941 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1942 has a return value which indicates the quality of the random data
1943 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1944 error queue. New function RAND_pseudo_bytes() generates output that is
1945 guaranteed to be unique but not unpredictable. RAND_add is like
1946 RAND_seed, but takes an extra argument for an entropy estimate
1947 (RAND_seed always assumes full entropy).
1950 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1951 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1952 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1953 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1954 false-positive rate of at most 2^-80 for random input.
1957 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1960 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1961 in the 0.9.5 release), this returns the chain
1962 from an X509_CTX structure with a dup of the stack and all
1963 the X509 reference counts upped: so the stack will exist
1964 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1967 Also make SSL_SESSION_print() print out the verify return
1971 *) Add manpage for the pkcs12 command. Also change the default
1972 behaviour so MAC iteration counts are used unless the new
1973 -nomaciter option is used. This improves file security and
1974 only older versions of MSIE (4.0 for example) need it.
1977 *) Honor the no-xxx Configure options when creating .DEF files.
1980 *) Add PKCS#10 attributes to field table: challengePassword,
1981 unstructuredName and unstructuredAddress. These are taken from
1982 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1983 international characters are used.
1985 More changes to X509_ATTRIBUTE code: allow the setting of types
1986 based on strings. Remove the 'loc' parameter when adding
1987 attributes because these will be a SET OF encoding which is sorted
1991 *) Initial changes to the 'req' utility to allow request generation
1992 automation. This will allow an application to just generate a template
1993 file containing all the field values and have req construct the
1996 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1997 used all over the place including certificate requests and PKCS#7
1998 structures. They are currently handled manually where necessary with
1999 some primitive wrappers for PKCS#7. The new functions behave in a
2000 manner analogous to the X509 extension functions: they allow
2001 attributes to be looked up by NID and added.
2003 Later something similar to the X509V3 code would be desirable to
2004 automatically handle the encoding, decoding and printing of the
2005 more complex types. The string types like challengePassword can
2006 be handled by the string table functions.
2008 Also modified the multi byte string table handling. Now there is
2009 a 'global mask' which masks out certain types. The table itself
2010 can use the flag STABLE_NO_MASK to ignore the mask setting: this
2011 is useful when for example there is only one permissible type
2012 (as in countryName) and using the mask might result in no valid
2016 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
2017 SSL_get_peer_finished to allow applications to obtain the latest
2018 Finished messages sent to the peer or expected from the peer,
2019 respectively. (SSL_get_peer_finished is usually the Finished message
2020 actually received from the peer, otherwise the protocol will be aborted.)
2022 As the Finished message are message digests of the complete handshake
2023 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
2024 be used for external authentication procedures when the authentication
2025 provided by SSL/TLS is not desired or is not enough.
2028 *) Enhanced support for Alpha Linux is added. Now ./config checks if
2029 the host supports BWX extension and if Compaq C is present on the
2030 $PATH. Just exploiting of the BWX extension results in 20-30%
2031 performance kick for some algorithms, e.g. DES and RC4 to mention
2032 a couple. Compaq C in turn generates ~20% faster code for MD5 and
2036 *) Add support for MS "fast SGC". This is arguably a violation of the
2037 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
2038 weak crypto and after checking the certificate is SGC a second one
2039 with strong crypto. MS SGC stops the first handshake after receiving
2040 the server certificate message and sends a second client hello. Since
2041 a server will typically do all the time consuming operations before
2042 expecting any further messages from the client (server key exchange
2043 is the most expensive) there is little difference between the two.
2045 To get OpenSSL to support MS SGC we have to permit a second client
2046 hello message after we have sent server done. In addition we have to
2047 reset the MAC if we do get this second client hello.
2050 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
2051 if a DER encoded private key is RSA or DSA traditional format. Changed
2052 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
2053 format DER encoded private key. Newer code should use PKCS#8 format which
2054 has the key type encoded in the ASN1 structure. Added DER private key
2055 support to pkcs8 application.
2058 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
2059 ciphersuites has been selected (as required by the SSL 3/TLS 1
2060 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
2061 is set, we interpret this as a request to violate the specification
2062 (the worst that can happen is a handshake failure, and 'correct'
2063 behaviour would result in a handshake failure anyway).
2066 *) In SSL_CTX_add_session, take into account that there might be multiple
2067 SSL_SESSION structures with the same session ID (e.g. when two threads
2068 concurrently obtain them from an external cache).
2069 The internal cache can handle only one SSL_SESSION with a given ID,
2070 so if there's a conflict, we now throw out the old one to achieve
2074 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2075 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2076 some routines that use cipher OIDs: some ciphers do not have OIDs
2077 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2081 *) Simplify the trust setting structure and code. Now we just have
2082 two sequences of OIDs for trusted and rejected settings. These will
2083 typically have values the same as the extended key usage extension
2084 and any application specific purposes.
2086 The trust checking code now has a default behaviour: it will just
2087 check for an object with the same NID as the passed id. Functions can
2088 be provided to override either the default behaviour or the behaviour
2089 for a given id. SSL client, server and email already have functions
2090 in place for compatibility: they check the NID and also return "trusted"
2091 if the certificate is self signed.
2094 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2095 traditional format into an EVP_PKEY structure.
2098 *) Add a password callback function PEM_cb() which either prompts for
2099 a password if usr_data is NULL or otherwise assumes it is a null
2100 terminated password. Allow passwords to be passed on command line
2101 environment or config files in a few more utilities.
2104 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2105 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2106 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2107 Update documentation.
2110 *) Support for ASN1 "NULL" type. This could be handled before by using
2111 ASN1_TYPE but there wasn't any function that would try to read a NULL
2112 and produce an error if it couldn't. For compatibility we also have
2113 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2114 don't allocate anything because they don't need to.
2117 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2119 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2121 *) Rebuild of the memory allocation routines used by OpenSSL code and
2122 possibly others as well. The purpose is to make an interface that
2123 provide hooks so anyone can build a separate set of allocation and
2124 deallocation routines to be used by OpenSSL, for example memory
2125 pool implementations, or something else, which was previously hard
2126 since Malloc(), Realloc() and Free() were defined as macros having
2127 the values malloc, realloc and free, respectively (except for Win32
2128 compilations). The same is provided for memory debugging code.
2129 OpenSSL already comes with functionality to find memory leaks, but
2130 this gives people a chance to debug other memory problems.
2132 With these changes, a new set of functions and macros have appeared:
2134 CRYPTO_set_mem_debug_functions() [F]
2135 CRYPTO_get_mem_debug_functions() [F]
2136 CRYPTO_dbg_set_options() [F]
2137 CRYPTO_dbg_get_options() [F]
2138 CRYPTO_malloc_debug_init() [M]
2140 The memory debug functions are NULL by default, unless the library
2141 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2142 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2143 gives the standard debugging functions that come with OpenSSL) or
2144 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2145 provided by the library user) must be used. When the standard
2146 debugging functions are used, CRYPTO_dbg_set_options can be used to
2147 request additional information:
2148 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2149 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2151 Also, things like CRYPTO_set_mem_functions will always give the
2152 expected result (the new set of functions is used for allocation
2153 and deallocation) at all times, regardless of platform and compiler
2156 To finish it up, some functions that were never use in any other
2157 way than through macros have a new API and new semantic:
2160 CRYPTO_dbg_realloc()
2163 All macros of value have retained their old syntax.
2164 [Richard Levitte and Bodo Moeller]
2166 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2167 ordering of SMIMECapabilities wasn't in "strength order" and there
2168 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2172 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2173 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2174 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2176 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2177 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2178 functionality to handle multipart/signed properly) and a utility
2179 called 'smime' to call all this stuff. This is based on code I
2180 originally wrote for Celo who have kindly allowed it to be
2181 included in OpenSSL.
2184 *) Add variants des_set_key_checked and des_set_key_unchecked of
2185 des_set_key (aka des_key_sched). Global variable des_check_key
2186 decides which of these is called by des_set_key; this way
2187 des_check_key behaves as it always did, but applications and
2188 the library itself, which was buggy for des_check_key == 1,
2189 have a cleaner way to pick the version they need.
2192 *) New function PKCS12_newpass() which changes the password of a
2196 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2197 dynamic mix. In both cases the ids can be used as an index into the
2198 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2199 functions so they accept a list of the field values and the
2200 application doesn't need to directly manipulate the X509_TRUST
2204 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2208 *) Modify the way the V3 extension code looks up extensions. This now
2209 works in a similar way to the object code: we have some "standard"
2210 extensions in a static table which is searched with OBJ_bsearch()
2211 and the application can add dynamic ones if needed. The file
2212 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2213 updated whenever a new extension is added to the core code and kept
2214 in ext_nid order. There is a simple program 'tabtest.c' which checks
2215 this. New extensions are not added too often so this file can readily
2216 be maintained manually.
2218 There are two big advantages in doing things this way. The extensions
2219 can be looked up immediately and no longer need to be "added" using
2220 X509V3_add_standard_extensions(): this function now does nothing.
2221 [Side note: I get *lots* of email saying the extension code doesn't
2222 work because people forget to call this function]
2223 Also no dynamic allocation is done unless new extensions are added:
2224 so if we don't add custom extensions there is no need to call
2225 X509V3_EXT_cleanup().
2228 *) Modify enc utility's salting as follows: make salting the default. Add a
2229 magic header, so unsalted files fail gracefully instead of just decrypting
2230 to garbage. This is because not salting is a big security hole, so people
2231 should be discouraged from doing it.
2234 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2235 digest to be passed on the command line but it only used this
2236 parameter when signing a certificate. Modified so all relevant
2237 operations are affected by the digest parameter including the
2238 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2239 DSA key was used because it didn't fix the digest.
2242 *) Initial certificate chain verify code. Currently tests the untrusted
2243 certificates for consistency with the verify purpose (which is set
2244 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2246 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2247 this is because it will reject chains with invalid extensions whereas
2248 every previous version of OpenSSL and SSLeay made no checks at all.
2250 Trust code: checks the root CA for the relevant trust settings. Trust
2251 settings have an initial value consistent with the verify purpose: e.g.
2252 if the verify purpose is for SSL client use it expects the CA to be
2253 trusted for SSL client use. However the default value can be changed to
2254 permit custom trust settings: one example of this would be to only trust
2255 certificates from a specific "secure" set of CAs.
2257 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2258 which should be used for version portability: especially since the
2259 verify structure is likely to change more often now.
2261 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2262 to set them. If not set then assume SSL clients will verify SSL servers
2265 Two new options to the verify program: -untrusted allows a set of
2266 untrusted certificates to be passed in and -purpose which sets the
2267 intended purpose of the certificate. If a purpose is set then the
2268 new chain verify code is used to check extension consistency.
2271 *) Support for the authority information access extension.
2274 *) Modify RSA and DSA PEM read routines to transparently handle
2275 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2276 public keys in a format compatible with certificate
2277 SubjectPublicKeyInfo structures. Unfortunately there were already
2278 functions called *_PublicKey_* which used various odd formats so
2279 these are retained for compatibility: however the DSA variants were
2280 never in a public release so they have been deleted. Changed dsa/rsa
2281 utilities to handle the new format: note no releases ever handled public
2282 keys so we should be OK.
2284 The primary motivation for this change is to avoid the same fiasco
2285 that dogs private keys: there are several incompatible private key
2286 formats some of which are standard and some OpenSSL specific and
2287 require various evil hacks to allow partial transparent handling and
2288 even then it doesn't work with DER formats. Given the option anything
2289 other than PKCS#8 should be dumped: but the other formats have to
2290 stay in the name of compatibility.
2292 With public keys and the benefit of hindsight one standard format
2293 is used which works with EVP_PKEY, RSA or DSA structures: though
2294 it clearly returns an error if you try to read the wrong kind of key.
2296 Added a -pubkey option to the 'x509' utility to output the public key.
2297 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2298 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2299 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2300 that do the same as the EVP_PKEY_assign_*() except they up the
2301 reference count of the added key (they don't "swallow" the
2305 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2306 CRLs would fail if the file contained no certificates or no CRLs:
2307 added a new function to read in both types and return the number
2308 read: this means that if none are read it will be an error. The
2309 DER versions of the certificate and CRL reader would always fail
2310 because it isn't possible to mix certificates and CRLs in DER format
2311 without choking one or the other routine. Changed this to just read
2312 a certificate: this is the best we can do. Also modified the code
2313 in apps/verify.c to take notice of return codes: it was previously
2314 attempting to read in certificates from NULL pointers and ignoring
2315 any errors: this is one reason why the cert and CRL reader seemed
2316 to work. It doesn't check return codes from the default certificate
2317 routines: these may well fail if the certificates aren't installed.
2320 *) Code to support otherName option in GeneralName.
2323 *) First update to verify code. Change the verify utility
2324 so it warns if it is passed a self signed certificate:
2325 for consistency with the normal behaviour. X509_verify
2326 has been modified to it will now verify a self signed
2327 certificate if *exactly* the same certificate appears
2328 in the store: it was previously impossible to trust a
2329 single self signed certificate. This means that:
2330 openssl verify ss.pem
2331 now gives a warning about a self signed certificate but
2332 openssl verify -CAfile ss.pem ss.pem
2336 *) For servers, store verify_result in SSL_SESSION data structure
2337 (and add it to external session representation).
2338 This is needed when client certificate verifications fails,
2339 but an application-provided verification callback (set by
2340 SSL_CTX_set_cert_verify_callback) allows accepting the session
2341 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2342 but returns 1): When the session is reused, we have to set
2343 ssl->verify_result to the appropriate error code to avoid
2345 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2347 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2348 case in PKCS7_dataInit() where the signed PKCS7 structure
2349 didn't contain any existing data because it was being created.
2350 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2352 *) Add a salt to the key derivation routines in enc.c. This
2353 forms the first 8 bytes of the encrypted file. Also add a
2354 -S option to allow a salt to be input on the command line.
2357 *) New function X509_cmp(). Oddly enough there wasn't a function
2358 to compare two certificates. We do this by working out the SHA1
2359 hash and comparing that. X509_cmp() will be needed by the trust
2363 *) SSL_get1_session() is like SSL_get_session(), but increments
2364 the reference count in the SSL_SESSION returned.
2365 [Geoff Thorpe <geoff@eu.c2.net>]
2367 *) Fix for 'req': it was adding a null to request attributes.
2368 Also change the X509_LOOKUP and X509_INFO code to handle
2369 certificate auxiliary information.
2372 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2376 *) Add the possibility to add extra information to the memory leak
2377 detecting output, to form tracebacks, showing from where each
2378 allocation was originated: CRYPTO_push_info("constant string") adds
2379 the string plus current file name and line number to a per-thread
2380 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2381 is like calling CYRPTO_pop_info() until the stack is empty.
2382 Also updated memory leak detection code to be multi-thread-safe.
2385 *) Add options -text and -noout to pkcs7 utility and delete the
2386 encryption options which never did anything. Update docs.
2389 *) Add options to some of the utilities to allow the pass phrase
2390 to be included on either the command line (not recommended on
2391 OSes like Unix) or read from the environment. Update the
2392 manpages and fix a few bugs.
2395 *) Add a few manpages for some of the openssl commands.
2398 *) Fix the -revoke option in ca. It was freeing up memory twice,
2399 leaking and not finding already revoked certificates.
2402 *) Extensive changes to support certificate auxiliary information.
2403 This involves the use of X509_CERT_AUX structure and X509_AUX
2404 functions. An X509_AUX function such as PEM_read_X509_AUX()
2405 can still read in a certificate file in the usual way but it
2406 will also read in any additional "auxiliary information". By
2407 doing things this way a fair degree of compatibility can be
2408 retained: existing certificates can have this information added
2409 using the new 'x509' options.
2411 Current auxiliary information includes an "alias" and some trust
2412 settings. The trust settings will ultimately be used in enhanced
2413 certificate chain verification routines: currently a certificate
2414 can only be trusted if it is self signed and then it is trusted
2418 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2419 The problem was that one of the replacement routines had not been working
2420 since SSLeay releases. For now the offending routine has been replaced
2421 with non-optimised assembler. Even so, this now gives around 95%
2422 performance improvement for 1024 bit RSA signs.
2425 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2426 handling. Most clients have the effective key size in bits equal to
2427 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2428 A few however don't do this and instead use the size of the decrypted key
2429 to determine the RC2 key length and the AlgorithmIdentifier to determine
2430 the effective key length. In this case the effective key length can still
2431 be 40 bits but the key length can be 168 bits for example. This is fixed
2432 by manually forcing an RC2 key into the EVP_PKEY structure because the
2433 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2434 the key length and effective key length are equal.
2437 *) Add a bunch of functions that should simplify the creation of
2438 X509_NAME structures. Now you should be able to do:
2439 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2440 and have it automatically work out the correct field type and fill in
2441 the structures. The more adventurous can try:
2442 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2443 and it will (hopefully) work out the correct multibyte encoding.
2446 *) Change the 'req' utility to use the new field handling and multibyte
2447 copy routines. Before the DN field creation was handled in an ad hoc
2448 way in req, ca, and x509 which was rather broken and didn't support
2449 BMPStrings or UTF8Strings. Since some software doesn't implement
2450 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2451 using the dirstring_type option. See the new comment in the default
2452 openssl.cnf for more info.
2455 *) Make crypto/rand/md_rand.c more robust:
2456 - Assure unique random numbers after fork().
2457 - Make sure that concurrent threads access the global counter and
2458 md serializably so that we never lose entropy in them
2459 or use exactly the same state in multiple threads.
2460 Access to the large state is not always serializable because
2461 the additional locking could be a performance killer, and
2462 md should be large enough anyway.
2465 *) New file apps/app_rand.c with commonly needed functionality
2466 for handling the random seed file.
2468 Use the random seed file in some applications that previously did not:
2470 dsaparam -genkey (which also ignored its '-rand' option),
2473 x509 (when signing).
2474 Except on systems with /dev/urandom, it is crucial to have a random
2475 seed file at least for key creation, DSA signing, and for DH exchanges;
2476 for RSA signatures we could do without one.
2478 gendh and gendsa (unlike genrsa) used to read only the first byte
2479 of each file listed in the '-rand' option. The function as previously
2480 found in genrsa is now in app_rand.c and is used by all programs
2481 that support '-rand'.
2484 *) In RAND_write_file, use mode 0600 for creating files;
2485 don't just chmod when it may be too late.
2488 *) Report an error from X509_STORE_load_locations
2489 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2492 *) New function ASN1_mbstring_copy() this copies a string in either
2493 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2494 into an ASN1_STRING type. A mask of permissible types is passed
2495 and it chooses the "minimal" type to use or an error if not type
2499 *) Add function equivalents to the various macros in asn1.h. The old
2500 macros are retained with an M_ prefix. Code inside the library can
2501 use the M_ macros. External code (including the openssl utility)
2502 should *NOT* in order to be "shared library friendly".
2505 *) Add various functions that can check a certificate's extensions
2506 to see if it usable for various purposes such as SSL client,
2507 server or S/MIME and CAs of these types. This is currently
2508 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2509 verification. Also added a -purpose flag to x509 utility to
2510 print out all the purposes.
2513 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2517 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2518 for, obtain and decode and extension and obtain its critical flag.
2519 This allows all the necessary extension code to be handled in a
2520 single function call.
2523 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2524 platforms. See crypto/rc4/rc4_enc.c for further details.
2527 *) New -noout option to asn1parse. This causes no output to be produced
2528 its main use is when combined with -strparse and -out to extract data
2529 from a file (which may not be in ASN.1 format).
2532 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2533 when producing the local key id.
2534 [Richard Levitte <levitte@stacken.kth.se>]
2536 *) New option -dhparam in s_server. This allows a DH parameter file to be
2537 stated explicitly. If it is not stated then it tries the first server
2538 certificate file. The previous behaviour hard coded the filename
2542 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2543 a public key to be input or output. For example:
2544 openssl rsa -in key.pem -pubout -out pubkey.pem
2545 Also added necessary DSA public key functions to handle this.
2548 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2549 in the message. This was handled by allowing
2550 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2551 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2553 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2554 to the end of the strings whereas this didn't. This would cause problems
2555 if strings read with d2i_ASN1_bytes() were later modified.
2556 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2558 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2559 data and it contains EOF it will end up returning an error. This is
2560 caused by input 46 bytes long. The cause is due to the way base64
2561 BIOs find the start of base64 encoded data. They do this by trying a
2562 trial decode on each line until they find one that works. When they
2563 do a flag is set and it starts again knowing it can pass all the
2564 data directly through the decoder. Unfortunately it doesn't reset
2565 the context it uses. This means that if EOF is reached an attempt
2566 is made to pass two EOFs through the context and this causes the
2567 resulting error. This can also cause other problems as well. As is
2568 usual with these problems it takes *ages* to find and the fix is
2569 trivial: move one line.
2570 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2572 *) Ugly workaround to get s_client and s_server working under Windows. The
2573 old code wouldn't work because it needed to select() on sockets and the
2574 tty (for keypresses and to see if data could be written). Win32 only
2575 supports select() on sockets so we select() with a 1s timeout on the
2576 sockets and then see if any characters are waiting to be read, if none
2577 are present then we retry, we also assume we can always write data to
2578 the tty. This isn't nice because the code then blocks until we've
2579 received a complete line of data and it is effectively polling the
2580 keyboard at 1s intervals: however it's quite a bit better than not
2581 working at all :-) A dedicated Windows application might handle this
2582 with an event loop for example.
2585 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2586 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2587 will be called when RSA_sign() and RSA_verify() are used. This is useful
2588 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2589 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2590 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2591 This necessitated the support of an extra signature type NID_md5_sha1
2592 for SSL signatures and modifications to the SSL library to use it instead
2593 of calling RSA_public_decrypt() and RSA_private_encrypt().
2596 *) Add new -verify -CAfile and -CApath options to the crl program, these
2597 will lookup a CRL issuers certificate and verify the signature in a
2598 similar way to the verify program. Tidy up the crl program so it
2599 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2600 less strict. It will now permit CRL extensions even if it is not
2601 a V2 CRL: this will allow it to tolerate some broken CRLs.
2604 *) Initialize all non-automatic variables each time one of the openssl
2605 sub-programs is started (this is necessary as they may be started
2606 multiple times from the "OpenSSL>" prompt).
2607 [Lennart Bang, Bodo Moeller]
2609 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2610 removing all other RSA functionality (this is what NO_RSA does). This
2611 is so (for example) those in the US can disable those operations covered
2612 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2616 *) Non-copying interface to BIO pairs.
2617 (still largely untested)
2620 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2621 ASCII string. This was handled independently in various places before.
2624 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2625 UTF8 strings a character at a time.
2628 *) Use client_version from client hello to select the protocol
2629 (s23_srvr.c) and for RSA client key exchange verification
2630 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2633 *) Add various utility functions to handle SPKACs, these were previously
2634 handled by poking round in the structure internals. Added new function
2635 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2636 print, verify and generate SPKACs. Based on an original idea from
2637 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2640 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2643 *) Allow the config file extension section to be overwritten on the
2644 command line. Based on an original idea from Massimiliano Pala
2645 <madwolf@comune.modena.it>. The new option is called -extensions
2646 and can be applied to ca, req and x509. Also -reqexts to override
2647 the request extensions in req and -crlexts to override the crl extensions
2651 *) Add new feature to the SPKAC handling in ca. Now you can include
2652 the same field multiple times by preceding it by "XXXX." for example:
2655 this is the same syntax as used in the req config file.
2658 *) Allow certificate extensions to be added to certificate requests. These
2659 are specified in a 'req_extensions' option of the req section of the
2660 config file. They can be printed out with the -text option to req but
2661 are otherwise ignored at present.
2664 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2665 data read consists of only the final block it would not decrypted because
2666 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2667 A misplaced 'break' also meant the decrypted final block might not be
2668 copied until the next read.
2671 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2672 a few extra parameters to the DH structure: these will be useful if
2673 for example we want the value of 'q' or implement X9.42 DH.
2676 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2677 provides hooks that allow the default DSA functions or functions on a
2678 "per key" basis to be replaced. This allows hardware acceleration and
2679 hardware key storage to be handled without major modification to the
2680 library. Also added low level modexp hooks and CRYPTO_EX structure and
2681 associated functions.
2684 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2685 as "read only": it can't be written to and the buffer it points to will
2686 not be freed. Reading from a read only BIO is much more efficient than
2687 a normal memory BIO. This was added because there are several times when
2688 an area of memory needs to be read from a BIO. The previous method was
2689 to create a memory BIO and write the data to it, this results in two
2690 copies of the data and an O(n^2) reading algorithm. There is a new
2691 function BIO_new_mem_buf() which creates a read only memory BIO from
2692 an area of memory. Also modified the PKCS#7 routines to use read only
2696 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2697 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2698 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2699 but a retry condition occured while trying to read the rest.
2702 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2703 NID_pkcs7_encrypted by default: this was wrong since this should almost
2704 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2705 the encrypted data type: this is a more sensible place to put it and it
2706 allows the PKCS#12 code to be tidied up that duplicated this
2710 *) Changed obj_dat.pl script so it takes its input and output files on
2711 the command line. This should avoid shell escape redirection problems
2715 *) Initial support for certificate extension requests, these are included
2716 in things like Xenroll certificate requests. Included functions to allow
2717 extensions to be obtained and added.
2720 *) -crlf option to s_client and s_server for sending newlines as
2721 CRLF (as required by many protocols).
2724 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2726 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2727 [Ralf S. Engelschall]
2729 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2730 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2732 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2736 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2737 DH parameters/keys (q is lost during that conversion, but the resulting
2738 DH parameters contain its length).
2740 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2741 much faster than DH_generate_parameters (which creates parameters
2742 where p = 2*q + 1), and also the smaller q makes DH computations
2743 much more efficient (160-bit exponentiation instead of 1024-bit
2744 exponentiation); so this provides a convenient way to support DHE
2745 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2746 utter importance to use
2747 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2749 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2750 when such DH parameters are used, because otherwise small subgroup
2751 attacks may become possible!
2754 *) Avoid memory leak in i2d_DHparams.
2757 *) Allow the -k option to be used more than once in the enc program:
2758 this allows the same encrypted message to be read by multiple recipients.
2761 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2762 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2763 it will always use the numerical form of the OID, even if it has a short
2767 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2768 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2769 otherwise bn_mod_exp was called. In the case of hardware keys for example
2770 no private key components need be present and it might store extra data
2771 in the RSA structure, which cannot be accessed from bn_mod_exp.
2772 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2773 private key operations.
2776 *) Added support for SPARC Linux.
2779 *) pem_password_cb function type incompatibly changed from
2780 typedef int pem_password_cb(char *buf, int size, int rwflag);
2782 ....(char *buf, int size, int rwflag, void *userdata);
2783 so that applications can pass data to their callbacks:
2784 The PEM[_ASN1]_{read,write}... functions and macros now take an
2785 additional void * argument, which is just handed through whenever
2786 the password callback is called.
2787 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2789 New function SSL_CTX_set_default_passwd_cb_userdata.
2791 Compatibility note: As many C implementations push function arguments
2792 onto the stack in reverse order, the new library version is likely to
2793 interoperate with programs that have been compiled with the old
2794 pem_password_cb definition (PEM_whatever takes some data that
2795 happens to be on the stack as its last argument, and the callback
2796 just ignores this garbage); but there is no guarantee whatsoever that
2799 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2800 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2801 problems not only on Windows, but also on some Unix platforms.
2802 To avoid problematic command lines, these definitions are now in an
2803 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2804 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2807 *) MIPS III/IV assembler module is reimplemented.
2810 *) More DES library cleanups: remove references to srand/rand and
2811 delete an unused file.
2814 *) Add support for the the free Netwide assembler (NASM) under Win32,
2815 since not many people have MASM (ml) and it can be hard to obtain.
2816 This is currently experimental but it seems to work OK and pass all
2817 the tests. Check out INSTALL.W32 for info.
2820 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2821 without temporary keys kept an extra copy of the server key,
2822 and connections with temporary keys did not free everything in case
2826 *) New function RSA_check_key and new openssl rsa option -check
2827 for verifying the consistency of RSA keys.
2828 [Ulf Moeller, Bodo Moeller]
2830 *) Various changes to make Win32 compile work:
2831 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2832 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2833 comparison" warnings.
2834 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2837 *) Add a debugging option to PKCS#5 v2 key generation function: when
2838 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2839 derived keys are printed to stderr.
2842 *) Copy the flags in ASN1_STRING_dup().
2843 [Roman E. Pavlov <pre@mo.msk.ru>]
2845 *) The x509 application mishandled signing requests containing DSA
2846 keys when the signing key was also DSA and the parameters didn't match.
2848 It was supposed to omit the parameters when they matched the signing key:
2849 the verifying software was then supposed to automatically use the CA's
2850 parameters if they were absent from the end user certificate.
2852 Omitting parameters is no longer recommended. The test was also
2853 the wrong way round! This was probably due to unusual behaviour in
2854 EVP_cmp_parameters() which returns 1 if the parameters match.
2855 This meant that parameters were omitted when they *didn't* match and
2856 the certificate was useless. Certificates signed with 'ca' didn't have
2858 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2860 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2861 The interface is as follows:
2862 Applications can use
2863 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2864 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2865 "off" is now the default.
2866 The library internally uses
2867 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2868 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2869 to disable memory-checking temporarily.
2871 Some inconsistent states that previously were possible (and were
2872 even the default) are now avoided.
2874 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2875 with each memory chunk allocated; this is occasionally more helpful
2876 than just having a counter.
2878 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2880 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2884 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2885 which largely parallels "options", but is for changing API behaviour,
2886 whereas "options" are about protocol behaviour.
2887 Initial "mode" flags are:
2889 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2890 a single record has been written.
2891 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2892 retries use the same buffer location.
2893 (But all of the contents must be
2897 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
2900 *) Fix problems with no-hmac etc.
2901 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2903 *) New functions RSA_get_default_method(), RSA_set_method() and
2904 RSA_get_method(). These allows replacement of RSA_METHODs without having
2905 to mess around with the internals of an RSA structure.
2908 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2909 Also really enable memory leak checks in openssl.c and in some
2911 [Chad C. Mulligan, Bodo Moeller]
2913 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2914 up the length of negative integers. This has now been simplified to just
2915 store the length when it is first determined and use it later, rather
2916 than trying to keep track of where data is copied and updating it to
2918 [Steve Henson, reported by Brien Wheeler
2919 <bwheeler@authentica-security.com>]
2921 *) Add a new function PKCS7_signatureVerify. This allows the verification
2922 of a PKCS#7 signature but with the signing certificate passed to the
2923 function itself. This contrasts with PKCS7_dataVerify which assumes the
2924 certificate is present in the PKCS#7 structure. This isn't always the
2925 case: certificates can be omitted from a PKCS#7 structure and be
2926 distributed by "out of band" means (such as a certificate database).
2929 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2930 function prototypes in pem.h, also change util/mkdef.pl to add the
2931 necessary function names.
2934 *) mk1mf.pl (used by Windows builds) did not properly read the
2935 options set by Configure in the top level Makefile, and Configure
2936 was not even able to write more than one option correctly.
2937 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2940 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2941 file to be loaded from a BIO or FILE pointer. The BIO version will
2942 for example allow memory BIOs to contain config info.
2945 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2946 Whoever hopes to achieve shared-library compatibility across versions
2947 must use this, not the compile-time macro.
2948 (Exercise 0.9.4: Which is the minimum library version required by
2950 Note: All this applies only to multi-threaded programs, others don't
2954 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2955 through a BIO pair triggered the default case, i.e.
2956 SSLerr(...,SSL_R_UNKNOWN_STATE).
2959 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2960 can use the SSL library even if none of the specific BIOs is
2964 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2965 for the encoded length.
2966 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2968 *) Add initial documentation of the X509V3 functions.
2971 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2972 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2973 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2974 secure PKCS#8 private key format with a high iteration count.
2977 *) Fix determination of Perl interpreter: A perl or perl5
2978 _directory_ in $PATH was also accepted as the interpreter.
2979 [Ralf S. Engelschall]
2981 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2982 wrong with it but it was very old and did things like calling
2983 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2987 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2988 to use the new extension code.
2991 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2992 with macros. This should make it easier to change their form, add extra
2993 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2997 *) Add to configuration table a new entry that can specify an alternative
2998 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2999 according to Mark Crispin <MRC@Panda.COM>.
3003 *) DES CBC did not update the IV. Weird.
3006 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
3007 Changing the behaviour of the former might break existing programs --
3008 where IV updating is needed, des_ncbc_encrypt can be used.
3011 *) When bntest is run from "make test" it drives bc to check its
3012 calculations, as well as internally checking them. If an internal check
3013 fails, it needs to cause bc to give a non-zero result or make test carries
3014 on without noticing the failure. Fixed.
3017 *) DES library cleanups.
3020 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
3021 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
3022 ciphers. NOTE: although the key derivation function has been verified
3023 against some published test vectors it has not been extensively tested
3024 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
3028 *) Instead of "mkdir -p", which is not fully portable, use new
3029 Perl script "util/mkdir-p.pl".
3032 *) Rewrite the way password based encryption (PBE) is handled. It used to
3033 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
3034 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
3035 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
3036 the 'parameter' field of the AlgorithmIdentifier is passed to the
3037 underlying key generation function so it must do its own ASN1 parsing.
3038 This has also changed the EVP_PBE_CipherInit() function which now has a
3039 'parameter' argument instead of literal salt and iteration count values
3040 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
3043 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
3044 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
3045 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
3046 KEY" because this clashed with PKCS#8 unencrypted string. Since this
3047 value was just used as a "magic string" and not used directly its
3048 value doesn't matter.
3051 *) Introduce some semblance of const correctness to BN. Shame C doesn't
3055 *) "linux-sparc64" configuration (ultrapenguin).
3056 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
3057 "linux-sparc" configuration.
3058 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
3060 *) config now generates no-xxx options for missing ciphers.
3063 *) Support the EBCDIC character set (work in progress).
3064 File ebcdic.c not yet included because it has a different license.
3065 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3067 *) Support BS2000/OSD-POSIX.
3068 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3070 *) Make callbacks for key generation use void * instead of char *.
3073 *) Make S/MIME samples compile (not yet tested).
3076 *) Additional typesafe stacks.
3079 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3083 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3085 *) New configuration variant "sco5-gcc".
3087 *) Updated some demos.
3088 [Sean O Riordain, Wade Scholine]
3090 *) Add missing BIO_free at exit of pkcs12 application.
3093 *) Fix memory leak in conf.c.
3096 *) Updates for Win32 to assembler version of MD5.
3099 *) Set #! path to perl in apps/der_chop to where we found it
3100 instead of using a fixed path.
3103 *) SHA library changes for irix64-mips4-cc.
3106 *) Improvements for VMS support.
3110 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3112 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3113 This also avoids the problems with SC4.2 and unpatched SC5.
3114 [Andy Polyakov <appro@fy.chalmers.se>]
3116 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3117 These are required because of the typesafe stack would otherwise break
3118 existing code. If old code used a structure member which used to be STACK
3119 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3120 sk_num or sk_value it would produce an error because the num, data members
3121 are not present in STACK_OF. Now it just produces a warning. sk_set
3122 replaces the old method of assigning a value to sk_value
3123 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3124 that does this will no longer work (and should use sk_set instead) but
3125 this could be regarded as a "questionable" behaviour anyway.
3128 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3129 correctly handle encrypted S/MIME data.
3132 *) Change type of various DES function arguments from des_cblock
3133 (which means, in function argument declarations, pointer to char)
3134 to des_cblock * (meaning pointer to array with 8 char elements),
3135 which allows the compiler to do more typechecking; it was like
3136 that back in SSLeay, but with lots of ugly casts.
3138 Introduce new type const_des_cblock.
3141 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3142 problems: find RecipientInfo structure that matches recipient certificate
3143 and initialise the ASN1 structures properly based on passed cipher.
3146 *) Belatedly make the BN tests actually check the results.
3149 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3150 to and from BNs: it was completely broken. New compilation option
3151 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3152 key elements as negative integers.
3155 *) Reorganize and speed up MD5.
3156 [Andy Polyakov <appro@fy.chalmers.se>]
3159 [Richard Levitte <richard@levitte.org>]
3161 *) New option -out to asn1parse to allow the parsed structure to be
3162 output to a file. This is most useful when combined with the -strparse
3163 option to examine the output of things like OCTET STRINGS.
3166 *) Make SSL library a little more fool-proof by not requiring any longer
3167 that SSL_set_{accept,connect}_state be called before
3168 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3169 in many applications because usually everything *appeared* to work as
3170 intended anyway -- now it really works as intended).
3173 *) Move openssl.cnf out of lib/.
3176 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3177 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3178 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3179 [Ralf S. Engelschall]
3181 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3182 handle PKCS#7 enveloped data properly.
3183 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3185 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3186 copying pointers. The cert_st handling is changed by this in
3187 various ways (and thus what used to be known as ctx->default_cert
3188 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3189 any longer when s->cert does not give us what we need).
3190 ssl_cert_instantiate becomes obsolete by this change.
3191 As soon as we've got the new code right (possibly it already is?),
3192 we have solved a couple of bugs of the earlier code where s->cert
3193 was used as if it could not have been shared with other SSL structures.
3195 Note that using the SSL API in certain dirty ways now will result
3196 in different behaviour than observed with earlier library versions:
3197 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3198 does not influence s as it used to.
3200 In order to clean up things more thoroughly, inside SSL_SESSION
3201 we don't use CERT any longer, but a new structure SESS_CERT
3202 that holds per-session data (if available); currently, this is
3203 the peer's certificate chain and, for clients, the server's certificate
3204 and temporary key. CERT holds only those values that can have
3205 meaningful defaults in an SSL_CTX.
3208 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3209 from the internal representation. Various PKCS#7 fixes: remove some
3210 evil casts and set the enc_dig_alg field properly based on the signing
3214 *) Allow PKCS#12 password to be set from the command line or the
3215 environment. Let 'ca' get its config file name from the environment
3216 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3220 *) Allow certificate policies extension to use an IA5STRING for the
3221 organization field. This is contrary to the PKIX definition but
3222 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3226 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3227 without disallowing inline assembler and the like for non-pedantic builds.
3230 *) Support Borland C++ builder.
3231 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3236 *) SHA-1 cleanups and performance enhancements.
3237 [Andy Polyakov <appro@fy.chalmers.se>]
3239 *) Sparc v8plus assembler for the bignum library.
3240 [Andy Polyakov <appro@fy.chalmers.se>]
3242 *) Accept any -xxx and +xxx compiler options in Configure.
3245 *) Update HPUX configuration.
3248 *) Add missing sk_<type>_unshift() function to safestack.h
3249 [Ralf S. Engelschall]
3251 *) New function SSL_CTX_use_certificate_chain_file that sets the
3252 "extra_cert"s in addition to the certificate. (This makes sense
3253 only for "PEM" format files, as chains as a whole are not
3257 *) Support verify_depth from the SSL API.
3258 x509_vfy.c had what can be considered an off-by-one-error:
3259 Its depth (which was not part of the external interface)
3260 was actually counting the number of certificates in a chain;
3261 now it really counts the depth.
3264 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3265 instead of X509err, which often resulted in confusing error
3266 messages since the error codes are not globally unique
3267 (e.g. an alleged error in ssl3_accept when a certificate
3268 didn't match the private key).
3270 *) New function SSL_CTX_set_session_id_context that allows to set a default
3271 value (so that you don't need SSL_set_session_id_context for each
3272 connection using the SSL_CTX).
3275 *) OAEP decoding bug fix.
3278 *) Support INSTALL_PREFIX for package builders, as proposed by
3282 *) New Configure options "threads" and "no-threads". For systems
3283 where the proper compiler options are known (currently Solaris
3284 and Linux), "threads" is the default.
3287 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3290 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3291 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3292 such as /usr/local/bin.
3295 *) "make linux-shared" to build shared libraries.
3296 [Niels Poppe <niels@netbox.org>]
3298 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3301 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3302 extension adding in x509 utility.
3305 *) Remove NOPROTO sections and error code comments.
3308 *) Partial rewrite of the DEF file generator to now parse the ANSI
3312 *) New Configure options --prefix=DIR and --openssldir=DIR.
3315 *) Complete rewrite of the error code script(s). It is all now handled
3316 by one script at the top level which handles error code gathering,
3317 header rewriting and C source file generation. It should be much better
3318 than the old method: it now uses a modified version of Ulf's parser to
3319 read the ANSI prototypes in all header files (thus the old K&R definitions
3320 aren't needed for error creation any more) and do a better job of
3321 translating function codes into names. The old 'ASN1 error code imbedded
3322 in a comment' is no longer necessary and it doesn't use .err files which
3323 have now been deleted. Also the error code call doesn't have to appear all
3324 on one line (which resulted in some large lines...).
3327 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3330 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3331 0 (which usually indicates a closed connection), but continue reading.
3334 *) Fix some race conditions.
3337 *) Add support for CRL distribution points extension. Add Certificate
3338 Policies and CRL distribution points documentation.
3341 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3344 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3345 8 of keying material. Merlin has also confirmed interop with this fix
3346 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3347 [Merlin Hughes <merlin@baltimore.ie>]
3349 *) Fix lots of warnings.
3350 [Richard Levitte <levitte@stacken.kth.se>]
3352 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3353 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3354 [Richard Levitte <levitte@stacken.kth.se>]
3356 *) Fix problems with sizeof(long) == 8.
3357 [Andy Polyakov <appro@fy.chalmers.se>]
3359 *) Change functions to ANSI C.
3362 *) Fix typos in error codes.
3363 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3365 *) Remove defunct assembler files from Configure.
3368 *) SPARC v8 assembler BIGNUM implementation.
3369 [Andy Polyakov <appro@fy.chalmers.se>]
3371 *) Support for Certificate Policies extension: both print and set.
3372 Various additions to support the r2i method this uses.
3375 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3376 return a const string when you are expecting an allocated buffer.
3379 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3380 types DirectoryString and DisplayText.
3383 *) Add code to allow r2i extensions to access the configuration database,
3384 add an LHASH database driver and add several ctx helper functions.
3387 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3388 fail when they extended the size of a BIGNUM.
3391 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3392 support typesafe stack.
3395 *) Fix typo in SSL_[gs]et_options().
3396 [Nils Frostberg <nils@medcom.se>]
3398 *) Delete various functions and files that belonged to the (now obsolete)
3399 old X509V3 handling code.
3402 *) New Configure option "rsaref".
3405 *) Don't auto-generate pem.h.
3408 *) Introduce type-safe ASN.1 SETs.
3411 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3412 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3414 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3415 that links with OpenSSL (well at least cause lots of warnings), but fear
3416 not: the conversion is trivial, and it eliminates loads of evil casts. A
3417 few STACKed things have been converted already. Feel free to convert more.
3418 In the fullness of time, I'll do away with the STACK type altogether.
3421 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3422 specified in <certfile> by updating the entry in the index.txt file.
3423 This way one no longer has to edit the index.txt file manually for
3424 revoking a certificate. The -revoke option does the gory details now.
3425 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3427 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3428 `-text' option at all and this way the `-noout -text' combination was
3429 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3430 [Ralf S. Engelschall]
3432 *) Make sure a corresponding plain text error message exists for the
3433 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3434 verify callback function determined that a certificate was revoked.
3435 [Ralf S. Engelschall]
3437 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3438 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3439 all available cipers including rc5, which was forgotten until now.
3440 In order to let the testing shell script know which algorithms
3441 are available, a new (up to now undocumented) command
3442 "openssl list-cipher-commands" is used.
3445 *) Bugfix: s_client occasionally would sleep in select() when
3446 it should have checked SSL_pending() first.
3449 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3450 the raw DSA values prior to ASN.1 encoding.
3453 *) Tweaks to Configure
3454 [Niels Poppe <niels@netbox.org>]
3456 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3460 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3463 *) New config option to avoid instructions that are illegal on the 80386.
3464 The default code is faster, but requires at least a 486.
3467 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3468 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3469 same as SSL2_VERSION anyway.
3472 *) New "-showcerts" option for s_client.
3475 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3476 application. Various cleanups and fixes.
3479 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3480 modify error routines to work internally. Add error codes and PBE init
3481 to library startup routines.
3484 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3485 packing functions to asn1 and evp. Changed function names and error
3486 codes along the way.
3489 *) PKCS12 integration: and so it begins... First of several patches to
3490 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3491 objects to objects.h
3494 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3495 and display support for Thawte strong extranet extension.
3498 *) Add LinuxPPC support.
3499 [Jeff Dubrule <igor@pobox.org>]
3501 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3502 bn_div_words in alpha.s.
3503 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3505 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3506 OAEP isn't supported when OpenSSL is built with RSAref.
3507 [Ulf Moeller <ulf@fitug.de>]
3509 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3510 so they no longer are missing under -DNOPROTO.
3511 [Soren S. Jorvang <soren@t.dk>]
3514 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3516 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3517 doesn't work when the session is reused. Coming soon!
3520 *) Fix a security hole, that allows sessions to be reused in the wrong
3521 context thus bypassing client cert protection! All software that uses
3522 client certs and session caches in multiple contexts NEEDS PATCHING to
3523 allow session reuse! A fuller solution is in the works.
3524 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3526 *) Some more source tree cleanups (removed obsolete files
3527 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3528 permission on "config" script to be executable) and a fix for the INSTALL
3530 [Ulf Moeller <ulf@fitug.de>]
3532 *) Remove some legacy and erroneous uses of malloc, free instead of
3534 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3536 *) Make rsa_oaep_test return non-zero on error.
3537 [Ulf Moeller <ulf@fitug.de>]
3539 *) Add support for native Solaris shared libraries. Configure
3540 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3541 if someone would make that last step automatic.
3542 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3544 *) ctx_size was not built with the right compiler during "make links". Fixed.
3547 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3548 except NULL ciphers". This means the default cipher list will no longer
3549 enable NULL ciphers. They need to be specifically enabled e.g. with
3550 the string "DEFAULT:eNULL".
3553 *) Fix to RSA private encryption routines: if p < q then it would
3554 occasionally produce an invalid result. This will only happen with
3555 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3558 *) Be less restrictive and allow also `perl util/perlpath.pl
3559 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3560 because this way one can also use an interpreter named `perl5' (which is
3561 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3562 installed as `perl').
3563 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3565 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3566 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3568 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3569 advapi32.lib to Win32 build and change the pem test comparision
3570 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3571 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3572 and crypto/des/ede_cbcm_enc.c.
3575 *) DES quad checksum was broken on big-endian architectures. Fixed.
3578 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3579 Win32 test batch file so it (might) work again. The Win32 test batch file
3580 is horrible: I feel ill....
3583 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3584 in e_os.h. Audit of header files to check ANSI and non ANSI
3585 sections: 10 functions were absent from non ANSI section and not exported
3586 from Windows DLLs. Fixed up libeay.num for new functions.
3589 *) Make `openssl version' output lines consistent.
3590 [Ralf S. Engelschall]
3592 *) Fix Win32 symbol export lists for BIO functions: Added
3593 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3594 to ms/libeay{16,32}.def.
3595 [Ralf S. Engelschall]
3597 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3598 fine under Unix and passes some trivial tests I've now added. But the
3599 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3600 added to make sure no one expects that this stuff really works in the
3601 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3602 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3604 [Ralf S. Engelschall]
3606 *) Fix the generation of two part addresses in perl.
3607 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3609 *) Add config entry for Linux on MIPS.
3610 [John Tobey <jtobey@channel1.com>]
3612 *) Make links whenever Configure is run, unless we are on Windoze.
3615 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3616 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3620 *) Add a useful kludge to allow package maintainers to specify compiler and
3621 other platforms details on the command line without having to patch the
3622 Configure script everytime: One now can use ``perl Configure
3623 <id>:<details>'', i.e. platform ids are allowed to have details appended
3624 to them (seperated by colons). This is treated as there would be a static
3625 pre-configured entry in Configure's %table under key <id> with value
3626 <details> and ``perl Configure <id>'' is called. So, when you want to
3627 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3628 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3629 now, which overrides the FreeBSD-elf entry on-the-fly.
3630 [Ralf S. Engelschall]
3632 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3635 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3636 on the `perl Configure ...' command line. This way one can compile
3637 OpenSSL libraries with Position Independent Code (PIC) which is needed
3638 for linking it into DSOs.
3639 [Ralf S. Engelschall]
3641 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3645 *) Cleaned up the LICENSE document: The official contact for any license
3646 questions now is the OpenSSL core team under openssl-core@openssl.org.
3647 And add a paragraph about the dual-license situation to make sure people
3648 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3649 to the OpenSSL toolkit.
3650 [Ralf S. Engelschall]
3652 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3653 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3654 Additonally cleaned up the `make links' target: Remove unnecessary
3655 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3656 to speed processing and no longer clutter the display with confusing
3657 stuff. Instead only the actually done links are displayed.
3658 [Ralf S. Engelschall]
3660 *) Permit null encryption ciphersuites, used for authentication only. It used
3661 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3662 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3666 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3667 signed attributes when verifying signatures (this would break them),
3668 the detached data encoding was wrong and public keys obtained using
3669 X509_get_pubkey() weren't freed.
3672 *) Add text documentation for the BUFFER functions. Also added a work around
3673 to a Win95 console bug. This was triggered by the password read stuff: the
3674 last character typed gets carried over to the next fread(). If you were
3675 generating a new cert request using 'req' for example then the last
3676 character of the passphrase would be CR which would then enter the first
3680 *) Added the new `Includes OpenSSL Cryptography Software' button as
3681 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3682 button and can be used by applications based on OpenSSL to show the
3683 relationship to the OpenSSL project.
3684 [Ralf S. Engelschall]
3686 *) Remove confusing variables in function signatures in files
3687 ssl/ssl_lib.c and ssl/ssl.h.
3688 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3690 *) Don't install bss_file.c under PREFIX/include/
3691 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3693 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3694 functions that return function pointers and has support for NT specific
3695 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3696 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3697 unsigned to signed types: this was killing the Win32 compile.
3700 *) Add new certificate file to stack functions,
3701 SSL_add_dir_cert_subjects_to_stack() and
3702 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3703 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3704 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3705 This means that Apache-SSL and similar packages don't have to mess around
3706 to add as many CAs as they want to the preferred list.
3709 *) Experiment with doxygen documentation. Currently only partially applied to
3711 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3712 openssl.doxy as the configuration file.
3715 *) Get rid of remaining C++-style comments which strict C compilers hate.
3716 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3718 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3719 compiled in by default: it has problems with large keys.
3722 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3723 DH private keys and/or callback functions which directly correspond to
3724 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3725 is needed for applications which have to configure certificates on a
3726 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3728 For the RSA certificate situation is makes no difference, but
3729 for the DSA certificate situation this fixes the "no shared cipher"
3730 problem where the OpenSSL cipher selection procedure failed because the
3731 temporary keys were not overtaken from the context and the API provided
3732 no way to reconfigure them.
3733 The new functions now let applications reconfigure the stuff and they
3734 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3735 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3736 non-public-API function ssl_cert_instantiate() is used as a helper
3737 function and also to reduce code redundancy inside ssl_rsa.c.
3738 [Ralf S. Engelschall]
3740 *) Move s_server -dcert and -dkey options out of the undocumented feature
3741 area because they are useful for the DSA situation and should be
3742 recognized by the users.
3743 [Ralf S. Engelschall]
3745 *) Fix the cipher decision scheme for export ciphers: the export bits are
3746 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3747 SSL_EXP_MASK. So, the original variable has to be used instead of the
3748 already masked variable.
3749 [Richard Levitte <levitte@stacken.kth.se>]
3751 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3752 [Richard Levitte <levitte@stacken.kth.se>]
3754 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3755 from `int' to `unsigned int' because it's a length and initialized by
3756 EVP_DigestFinal() which expects an `unsigned int *'.
3757 [Richard Levitte <levitte@stacken.kth.se>]
3759 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3760 script. Instead use the usual Shell->Perl transition trick.
3761 [Ralf S. Engelschall]
3763 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3764 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3765 -noout -modulus' as it's already the case for `openssl rsa -noout
3766 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3767 currently the public key is printed (a decision which was already done by
3768 `openssl dsa -modulus' in the past) which serves a similar purpose.
3769 Additionally the NO_RSA no longer completely removes the whole -modulus
3770 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3772 [Ralf S. Engelschall]
3774 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3775 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3776 [Arne Ansper <arne@ats.cyber.ee>]
3778 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3779 to be added. Now both 'req' and 'ca' can use new objects defined in the
3783 *) Add cool BIO that does syslog (or event log on NT).
3784 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3786 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3787 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3788 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3789 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3792 *) Add preliminary config info for new extension code.
3795 *) Make RSA_NO_PADDING really use no padding.
3796 [Ulf Moeller <ulf@fitug.de>]
3798 *) Generate errors when private/public key check is done.
3801 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3802 for some CRL extensions and new objects added.
3805 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3806 key usage extension and fuller support for authority key id.
3809 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3810 padding method for RSA, which is recommended for new applications in PKCS
3811 #1 v2.0 (RFC 2437, October 1998).
3812 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3813 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3814 against Bleichbacher's attack on RSA.
3815 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3818 *) Updates to the new SSL compression code
3819 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3821 *) Fix so that the version number in the master secret, when passed
3822 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3823 (because the server will not accept higher), that the version number
3824 is 0x03,0x01, not 0x03,0x00
3825 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3827 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3828 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3829 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3832 *) Support for RAW extensions where an arbitrary extension can be
3833 created by including its DER encoding. See apps/openssl.cnf for
3837 *) Make sure latest Perl versions don't interpret some generated C array
3838 code as Perl array code in the crypto/err/err_genc.pl script.
3839 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3841 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3842 not many people have the assembler. Various Win32 compilation fixes and
3843 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3847 *) Modify configure script 'Configure' to automatically create crypto/date.h
3848 file under Win32 and also build pem.h from pem.org. New script
3849 util/mkfiles.pl to create the MINFO file on environments that can't do a
3850 'make files': perl util/mkfiles.pl >MINFO should work.
3853 *) Major rework of DES function declarations, in the pursuit of correctness
3854 and purity. As a result, many evil casts evaporated, and some weirdness,
3855 too. You may find this causes warnings in your code. Zapping your evil
3856 casts will probably fix them. Mostly.
3859 *) Fix for a typo in asn1.h. Bug fix to object creation script
3860 obj_dat.pl. It considered a zero in an object definition to mean
3861 "end of object": none of the objects in objects.h have any zeros
3862 so it wasn't spotted.
3863 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3865 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3866 Masking (CBCM). In the absence of test vectors, the best I have been able
3867 to do is check that the decrypt undoes the encrypt, so far. Send me test
3868 vectors if you have them.
3871 *) Correct calculation of key length for export ciphers (too much space was
3872 allocated for null ciphers). This has not been tested!
3875 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3876 message is now correct (it understands "crypto" and "ssl" on its
3877 command line). There is also now an "update" option. This will update
3878 the util/ssleay.num and util/libeay.num files with any new functions.
3880 perl util/mkdef.pl crypto ssl update
3881 it will update them.
3884 *) Overhauled the Perl interface (perl/*):
3885 - ported BN stuff to OpenSSL's different BN library
3886 - made the perl/ source tree CVS-aware
3887 - renamed the package from SSLeay to OpenSSL (the files still contain
3888 their history because I've copied them in the repository)
3889 - removed obsolete files (the test scripts will be replaced
3890 by better Test::Harness variants in the future)
3891 [Ralf S. Engelschall]
3893 *) First cut for a very conservative source tree cleanup:
3894 1. merge various obsolete readme texts into doc/ssleay.txt
3895 where we collect the old documents and readme texts.
3896 2. remove the first part of files where I'm already sure that we no
3897 longer need them because of three reasons: either they are just temporary
3898 files which were left by Eric or they are preserved original files where
3899 I've verified that the diff is also available in the CVS via "cvs diff
3900 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3901 the crypto/md/ stuff).
3902 [Ralf S. Engelschall]
3904 *) More extension code. Incomplete support for subject and issuer alt
3905 name, issuer and authority key id. Change the i2v function parameters
3906 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3907 what that's for :-) Fix to ASN1 macro which messed up
3908 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3911 *) Preliminary support for ENUMERATED type. This is largely copied from the
3915 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3916 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3918 *) Make sure `make rehash' target really finds the `openssl' program.
3919 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3921 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3922 like to hear about it if this slows down other processors.
3925 *) Add CygWin32 platform information to Configure script.
3926 [Alan Batie <batie@aahz.jf.intel.com>]
3928 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3929 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3931 *) New program nseq to manipulate netscape certificate sequences
3934 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3938 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3939 but the BN code had some problems that would cause failures when
3940 doing certificate verification and some other functions.
3941 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3943 *) Add ASN1 and PEM code to support netscape certificate sequences.
3946 *) Add ASN1 and PEM code to support netscape certificate sequences.
3949 *) Add several PKIX and private extended key usage OIDs.
3952 *) Modify the 'ca' program to handle the new extension code. Modify
3953 openssl.cnf for new extension format, add comments.
3956 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3957 and add a sample to openssl.cnf so req -x509 now adds appropriate
3961 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3962 error code, add initial support to X509_print() and x509 application.
3965 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3966 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3967 stuff is currently isolated and isn't even compiled yet.
3970 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3971 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3972 Removed the versions check from X509 routines when loading extensions:
3973 this allows certain broken certificates that don't set the version
3974 properly to be processed.
3977 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3978 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3979 can still be regenerated with "make depend".
3982 *) Spelling mistake in C version of CAST-128.
3983 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3985 *) Changes to the error generation code. The perl script err-code.pl
3986 now reads in the old error codes and retains the old numbers, only
3987 adding new ones if necessary. It also only changes the .err files if new
3988 codes are added. The makefiles have been modified to only insert errors
3989 when needed (to avoid needlessly modifying header files). This is done
3990 by only inserting errors if the .err file is newer than the auto generated
3991 C file. To rebuild all the error codes from scratch (the old behaviour)
3992 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3993 or delete all the .err files.
3996 *) CAST-128 was incorrectly implemented for short keys. The C version has
3997 been fixed, but is untested. The assembler versions are also fixed, but
3998 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3999 to regenerate it if needed.
4000 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
4001 Hagino <itojun@kame.net>]
4003 *) File was opened incorrectly in randfile.c.
4004 [Ulf Möller <ulf@fitug.de>]
4006 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
4007 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
4008 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
4009 al: it's just almost always a UTCTime. Note this patch adds new error
4010 codes so do a "make errors" if there are problems.
4013 *) Correct Linux 1 recognition in config.
4014 [Ulf Möller <ulf@fitug.de>]
4016 *) Remove pointless MD5 hash when using DSA keys in ca.
4017 [Anonymous <nobody@replay.com>]
4019 *) Generate an error if given an empty string as a cert directory. Also
4020 generate an error if handed NULL (previously returned 0 to indicate an
4021 error, but didn't set one).
4022 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
4024 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
4027 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
4028 parameters. This was causing a warning which killed off the Win32 compile.
4031 *) Remove C++ style comments from crypto/bn/bn_local.h.
4032 [Neil Costigan <neil.costigan@celocom.com>]
4034 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
4035 based on a text string, looking up short and long names and finally
4036 "dot" format. The "dot" format stuff didn't work. Added new function
4037 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
4038 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
4039 OID is not part of the table.
4042 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
4043 X509_LOOKUP_by_alias().
4046 *) Sort openssl functions by name.
4049 *) Get the gendsa program working (hopefully) and add it to app list. Remove
4050 encryption from sample DSA keys (in case anyone is interested the password
4054 *) Make _all_ *_free functions accept a NULL pointer.
4055 [Frans Heymans <fheymans@isaserver.be>]
4057 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
4059 [Anonymous <nobody@replay.com>]
4061 *) s_server should send the CAfile as acceptable CAs, not its own cert.
4062 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4064 *) Don't blow it for numeric -newkey arguments to apps/req.
4065 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4067 *) Temp key "for export" tests were wrong in s3_srvr.c.
4068 [Anonymous <nobody@replay.com>]
4070 *) Add prototype for temp key callback functions
4071 SSL_CTX_set_tmp_{rsa,dh}_callback().
4074 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4075 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4078 *) X509_name_add_entry() freed the wrong thing after an error.
4079 [Arne Ansper <arne@ats.cyber.ee>]
4081 *) rsa_eay.c would attempt to free a NULL context.
4082 [Arne Ansper <arne@ats.cyber.ee>]
4084 *) BIO_s_socket() had a broken should_retry() on Windoze.
4085 [Arne Ansper <arne@ats.cyber.ee>]
4087 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4088 [Arne Ansper <arne@ats.cyber.ee>]
4090 *) Make sure the already existing X509_STORE->depth variable is initialized
4091 in X509_STORE_new(), but document the fact that this variable is still
4092 unused in the certificate verification process.
4093 [Ralf S. Engelschall]
4095 *) Fix the various library and apps files to free up pkeys obtained from
4096 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4099 *) Fix reference counting in X509_PUBKEY_get(). This makes
4100 demos/maurice/example2.c work, amongst others, probably.
4101 [Steve Henson and Ben Laurie]
4103 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4104 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4105 are no longer created. This way we have a single and consistent command
4106 line interface `openssl <command>', similar to `cvs <command>'.
4107 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4109 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4110 BIT STRING wrapper always have zero unused bits.
4113 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4116 *) Make the top-level INSTALL documentation easier to understand.
4119 *) Makefiles updated to exit if an error occurs in a sub-directory
4120 make (including if user presses ^C) [Paul Sutton]
4122 *) Make Montgomery context stuff explicit in RSA data structure.
4125 *) Fix build order of pem and err to allow for generated pem.h.
4128 *) Fix renumbering bug in X509_NAME_delete_entry().
4131 *) Enhanced the err-ins.pl script so it makes the error library number
4132 global and can add a library name. This is needed for external ASN1 and
4133 other error libraries.
4136 *) Fixed sk_insert which never worked properly.
4139 *) Fix ASN1 macros so they can handle indefinite length construted
4140 EXPLICIT tags. Some non standard certificates use these: they can now
4144 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4145 into a single doc/ssleay.txt bundle. This way the information is still
4146 preserved but no longer messes up this directory. Now it's new room for
4147 the new set of documenation files.
4148 [Ralf S. Engelschall]
4150 *) SETs were incorrectly DER encoded. This was a major pain, because they
4151 shared code with SEQUENCEs, which aren't coded the same. This means that
4152 almost everything to do with SETs or SEQUENCEs has either changed name or
4153 number of arguments.
4154 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4156 *) Fix test data to work with the above.
4159 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4160 was already fixed by Eric for 0.9.1 it seems.
4161 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4163 *) Autodetect FreeBSD3.
4166 *) Fix various bugs in Configure. This affects the following platforms:
4170 unixware-2.0-pentium
4174 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4175 before they are needed.
4178 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4182 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4184 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4185 changed SSLeay to OpenSSL in version strings.
4186 [Ralf S. Engelschall]
4188 *) Some fixups to the top-level documents.
4191 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4192 because the symlink to include/ was missing.
4193 [Ralf S. Engelschall]
4195 *) Incorporated the popular no-RSA/DSA-only patches
4196 which allow to compile a RSA-free SSLeay.
4197 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4199 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4200 when "ssleay" is still not found.
4201 [Ralf S. Engelschall]
4203 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4204 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4206 *) Updated the README file.
4207 [Ralf S. Engelschall]
4209 *) Added various .cvsignore files in the CVS repository subdirs
4210 to make a "cvs update" really silent.
4211 [Ralf S. Engelschall]
4213 *) Recompiled the error-definition header files and added
4214 missing symbols to the Win32 linker tables.
4215 [Ralf S. Engelschall]
4217 *) Cleaned up the top-level documents;
4218 o new files: CHANGES and LICENSE
4219 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4220 o merged COPYRIGHT into LICENSE
4221 o removed obsolete TODO file
4222 o renamed MICROSOFT to INSTALL.W32
4223 [Ralf S. Engelschall]
4225 *) Removed dummy files from the 0.9.1b source tree:
4226 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4227 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4228 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4229 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4230 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4231 [Ralf S. Engelschall]
4233 *) Added various platform portability fixes.
4236 *) The Genesis of the OpenSSL rpject:
4237 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4238 Young and Tim J. Hudson created while they were working for C2Net until
4240 [The OpenSSL Project]
4243 Changes between 0.9.0b and 0.9.1b [not released]
4245 *) Updated a few CA certificates under certs/
4248 *) Changed some BIGNUM api stuff.
4251 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4252 DGUX x86, Linux Alpha, etc.
4255 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4256 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4260 *) Add -strparse option to asn1pars program which parses nested
4262 [Dr Stephen Henson <shenson@bigfoot.com>]
4264 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4267 *) DSA fix for "ca" program.
4270 *) Added "-genkey" option to "dsaparam" program.
4273 *) Added RIPE MD160 (rmd160) message digest.
4276 *) Added -a (all) option to "ssleay version" command.
4279 *) Added PLATFORM define which is the id given to Configure.
4282 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4285 *) Extended the ASN.1 parser routines.
4288 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4291 *) Added a BN_CTX to the BN library.
4294 *) Fixed the weak key values in DES library
4297 *) Changed API in EVP library for cipher aliases.
4300 *) Added support for RC2/64bit cipher.
4303 *) Converted the lhash library to the crypto/mem.c functions.
4306 *) Added more recognized ASN.1 object ids.
4309 *) Added more RSA padding checks for SSL/TLS.
4312 *) Added BIO proxy/filter functionality.
4315 *) Added extra_certs to SSL_CTX which can be used
4316 send extra CA certificates to the client in the CA cert chain sending
4317 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4320 *) Now Fortezza is denied in the authentication phase because
4321 this is key exchange mechanism is not supported by SSLeay at all.
4324 *) Additional PKCS1 checks.
4327 *) Support the string "TLSv1" for all TLS v1 ciphers.
4330 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4331 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4334 *) Fixed a few memory leaks.
4337 *) Fixed various code and comment typos.
4340 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4341 bytes sent in the client random.
4342 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob