5 Changes between 0.9.6c and 0.9.6d [XX xxx XXXX]
7 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8 an end-of-file condition would erronously be flagged, when the CRLF
9 was just at the end of a processed block. The bug was discovered when
10 processing data through a buffering memory BIO handing the data to a
11 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
12 <ptsekov@syntrex.com> and Nedelcho Stanev.
15 *) Implement a countermeasure against a vulnerability recently found
16 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
17 before application data chunks to avoid the use of known IVs
18 with data potentially chosen by the attacker.
21 *) Fix length checks in ssl3_get_client_hello().
24 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
25 to prevent ssl3_read_internal() from incorrectly assuming that
26 ssl3_read_bytes() found application data while handshake
27 processing was enabled when in fact s->s3->in_read_app_data was
28 merely automatically cleared during the initial handshake.
29 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
31 *) Fix object definitions for Private and Enterprise: they were not
32 recognized in their shortname (=lowercase) representation. Extend
33 obj_dat.pl to issue an error when using undefined keywords instead
34 of silently ignoring the problem (Svenning Sorensen
35 <sss@sss.dnsalias.net>).
38 *) Fix DH_generate_parameters() so that it works for 'non-standard'
39 generators, i.e. generators other than 2 and 5. (Previously, the
40 code did not properly initialise the 'add' and 'rem' values to
43 In the new general case, we do not insist that 'generator' is
44 actually a primitive root: This requirement is rather pointless;
45 a generator of the order-q subgroup is just as good, if not
49 *) Map new X509 verification errors to alerts. Discovered and submitted by
50 Tom Wu <tom@arcot.com>.
53 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
54 returning non-zero before the data has been completely received
55 when using non-blocking I/O.
56 [Bodo Moeller; problem pointed out by John Hughes]
58 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
59 [Ben Laurie, Lutz Jaenicke]
61 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
62 Yoram Zahavi <YoramZ@gilian.com>).
65 *) Add information about CygWin 1.3 and on, and preserve proper
66 configuration for the versions before that.
67 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
69 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
70 check whether we deal with a copy of a session and do not delete from
71 the cache in this case. Problem reported by "Izhar Shoshani Levi"
72 <izhar@checkpoint.com>.
75 *) Do not store session data into the internal session cache, if it
76 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
77 flag is set). Proposed by Aslam <aslam@funk.com>.
80 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
84 *) [In 0.9.6c-engine release:]
85 Fix a crashbug and a logic bug in hwcrhk_load_pubkey()
86 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
88 *) Add the configuration target linux-s390x.
89 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
91 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
92 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
93 variable as an indication that a ClientHello message has been
94 received. As the flag value will be lost between multiple
95 invocations of ssl3_accept when using non-blocking I/O, the
96 function may not be aware that a handshake has actually taken
97 place, thus preventing a new session from being added to the
100 To avoid this problem, we now set s->new_session to 2 instead of
101 using a local variable.
102 [Lutz Jaenicke, Bodo Moeller]
104 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
105 if the SSL_R_LENGTH_MISMATCH error is detected.
106 [Geoff Thorpe, Bodo Moeller]
108 *) New 'shared_ldflag' column in Configure platform table.
111 *) Fix EVP_CIPHER_mode macro.
112 ["Dan S. Camper" <dan@bti.net>]
114 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
115 type, we must throw them away by setting rr->length to 0.
116 [D P Chang <dpc@qualys.com>]
118 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
120 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
121 <Dominikus.Scherkl@biodata.com>. (The previous implementation
122 worked incorrectly for those cases where range = 10..._2 and
123 3*range is two bits longer than range.)
126 *) Only add signing time to PKCS7 structures if it is not already
130 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
131 OBJ_ld_ce should be OBJ_id_ce.
132 Also some ip-pda OIDs in crypto/objects/objects.txt were
133 incorrect (cf. RFC 3039).
134 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
136 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
137 returns early because it has nothing to do.
138 [Andy Schneider <andy.schneider@bjss.co.uk>]
140 *) [In 0.9.6c-engine release:]
141 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
142 [Andy Schneider <andy.schneider@bjss.co.uk>]
144 *) [In 0.9.6c-engine release:]
145 Add support for Cryptographic Appliance's keyserver technology.
146 (Use engine 'keyclient')
147 [Cryptographic Appliances and Geoff Thorpe]
149 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
150 is called via tools/c89.sh because arguments have to be
151 rearranged (all '-L' options must appear before the first object
153 [Richard Shapiro <rshapiro@abinitio.com>]
155 *) [In 0.9.6c-engine release:]
156 Add support for Broadcom crypto accelerator cards, backported
158 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
160 *) [In 0.9.6c-engine release:]
161 Add support for SureWare crypto accelerator cards from
162 Baltimore Technologies. (Use engine 'sureware')
163 [Baltimore Technologies and Mark Cox]
165 *) [In 0.9.6c-engine release:]
166 Add support for crypto accelerator cards from Accelerated
167 Encryption Processing, www.aep.ie. (Use engine 'aep')
168 [AEP Inc. and Mark Cox]
170 *) Add a configuration entry for gcc on UnixWare.
171 [Gary Benson <gbenson@redhat.com>]
173 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
174 messages are stored in a single piece (fixed-length part and
175 variable-length part combined) and fix various bugs found on the way.
178 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
179 instead. BIO_gethostbyname() does not know what timeouts are
180 appropriate, so entries would stay in cache even when they have
182 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
184 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
185 faced with a pathologically small ClientHello fragment that does
186 not contain client_version: Instead of aborting with an error,
187 simply choose the highest available protocol version (i.e.,
188 TLS 1.0 unless it is disabled). In practice, ClientHello
189 messages are never sent like this, but this change gives us
190 strictly correct behaviour at least for TLS.
193 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
194 never resets s->method to s->ctx->method when called from within
195 one of the SSL handshake functions.
196 [Bodo Moeller; problem pointed out by Niko Baric]
198 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
199 (sent using the client's version number) if client_version is
200 smaller than the protocol version in use. Also change
201 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
202 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
203 the client will at least see that alert.
206 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
210 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
211 client receives HelloRequest while in a handshake.
212 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
214 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
215 should end in 'break', not 'goto end' which circuments various
216 cleanups done in state SSL_ST_OK. But session related stuff
217 must be disabled for SSL_ST_OK in the case that we just sent a
220 Also avoid some overhead by not calling ssl_init_wbio_buffer()
221 before just sending a HelloRequest.
222 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
224 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
225 reveal whether illegal block cipher padding was found or a MAC
226 verification error occured. (Neither SSLerr() codes nor alerts
227 are directly visible to potential attackers, but the information
228 may leak via logfiles.)
230 Similar changes are not required for the SSL 2.0 implementation
231 because the number of padding bytes is sent in clear for SSL 2.0,
232 and the extra bytes are just ignored. However ssl/s2_pkt.c
233 failed to verify that the purported number of padding bytes is in
237 *) Add OpenUNIX-8 support including shared libraries
238 (Boyd Lynn Gerber <gerberb@zenez.com>).
241 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
242 'wristwatch attack' using huge encoding parameters (cf.
243 James H. Manger's CRYPTO 2001 paper). Note that the
244 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
245 encoding parameters and hence was not vulnerable.
249 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
251 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
252 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
253 followed by modular reduction.
254 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
256 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
257 equivalent based on BN_pseudo_rand() instead of BN_rand().
260 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
261 This function was broken, as the check for a new client hello message
262 to handle SGC did not allow these large messages.
263 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
266 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
269 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
270 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
273 *) Rework the configuration and shared library support for Tru64 Unix.
274 The configuration part makes use of modern compiler features and
275 still retains old compiler behavior for those that run older versions
276 of the OS. The shared library support part includes a variant that
277 uses the RPATH feature, and is available through the special
278 configuration target "alpha-cc-rpath", which will never be selected
280 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
282 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
283 with the same message size as in ssl3_get_certificate_request().
284 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
285 messages might inadvertently be reject as too long.
286 [Petr Lampa <lampa@fee.vutbr.cz>]
288 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
291 *) Modified SSL library such that the verify_callback that has been set
292 specificly for an SSL object with SSL_set_verify() is actually being
293 used. Before the change, a verify_callback set with this function was
294 ignored and the verify_callback() set in the SSL_CTX at the time of
295 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
296 to allow the necessary settings.
299 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
300 explicitly to NULL, as at least on Solaris 8 this seems not always to be
301 done automatically (in contradiction to the requirements of the C
302 standard). This made problems when used from OpenSSH.
305 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
306 dh->length and always used
308 BN_rand_range(priv_key, dh->p).
310 BN_rand_range() is not necessary for Diffie-Hellman, and this
311 specific range makes Diffie-Hellman unnecessarily inefficient if
312 dh->length (recommended exponent length) is much smaller than the
313 length of dh->p. We could use BN_rand_range() if the order of
314 the subgroup was stored in the DH structure, but we only have
319 BN_rand(priv_key, l, ...)
321 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
327 RSA_eay_public_encrypt
328 RSA_eay_private_decrypt
329 RSA_eay_private_encrypt (signing)
330 RSA_eay_public_decrypt (signature verification)
332 (default implementations for RSA_public_encrypt,
333 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
334 always reject numbers >= n.
337 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
338 to synchronize access to 'locking_thread'. This is necessary on
339 systems where access to 'locking_thread' (an 'unsigned long'
340 variable) is not atomic.
343 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
344 *before* setting the 'crypto_lock_rand' flag. The previous code had
345 a race condition if 0 is a valid thread ID.
346 [Travis Vitek <vitek@roguewave.com>]
348 *) Add support for shared libraries under Irix.
349 [Albert Chin-A-Young <china@thewrittenword.com>]
351 *) Add configuration option to build on Linux on both big-endian and
353 [Ralf Baechle <ralf@uni-koblenz.de>]
355 *) Add the possibility to create shared libraries on HP-UX.
358 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
360 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
361 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
362 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
363 PRNG state recovery was possible based on the output of
364 one PRNG request appropriately sized to gain knowledge on
365 'md' followed by enough consecutive 1-byte PRNG requests
366 to traverse all of 'state'.
368 1. When updating 'md_local' (the current thread's copy of 'md')
369 during PRNG output generation, hash all of the previous
370 'md_local' value, not just the half used for PRNG output.
372 2. Make the number of bytes from 'state' included into the hash
373 independent from the number of PRNG bytes requested.
375 The first measure alone would be sufficient to avoid
376 Markku-Juhani's attack. (Actually it had never occurred
377 to me that the half of 'md_local' used for chaining was the
378 half from which PRNG output bytes were taken -- I had always
379 assumed that the secret half would be used.) The second
380 measure makes sure that additional data from 'state' is never
381 mixed into 'md_local' in small portions; this heuristically
382 further strengthens the PRNG.
385 *) Fix crypto/bn/asm/mips3.s.
388 *) When only the key is given to "enc", the IV is undefined. Print out
389 an error message in this case.
392 *) Handle special case when X509_NAME is empty in X509 printing routines.
395 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
396 positive and less than q.
399 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
400 used: it isn't thread safe and the add_lock_callback should handle
402 [Paul Rose <Paul.Rose@bridge.com>]
404 *) Verify that incoming data obeys the block size in
405 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
409 [Ulf Möller, Bodo Möller]
411 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
412 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
413 when fixing the server behaviour for backwards-compatible 'client
414 hello' messages. (Note that the attack is impractical against
415 SSL 3.0 and TLS 1.0 anyway because length and version checking
416 means that the probability of guessing a valid ciphertext is
417 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
420 Before 0.9.5, the countermeasure (hide the error by generating a
421 random 'decryption result') did not work properly because
422 ERR_clear_error() was missing, meaning that SSL_get_error() would
423 detect the supposedly ignored error.
425 Both problems are now fixed.
428 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
429 (previously it was 1024).
432 *) Fix for compatibility mode trust settings: ignore trust settings
433 unless some valid trust or reject settings are present.
436 *) Fix for blowfish EVP: its a variable length cipher.
439 *) Fix various bugs related to DSA S/MIME verification. Handle missing
440 parameters in DSA public key structures and return an error in the
441 DSA routines if parameters are absent.
444 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
445 in the current directory if neither $RANDFILE nor $HOME was set.
446 RAND_file_name() in 0.9.6a returned NULL in this case. This has
447 caused some confusion to Windows users who haven't defined $HOME.
448 Thus RAND_file_name() is changed again: e_os.h can define a
449 DEFAULT_HOME, which will be used if $HOME is not set.
450 For Windows, we use "C:"; on other platforms, we still require
451 environment variables.
453 *) Move 'if (!initialized) RAND_poll()' into regions protected by
454 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
455 having multiple threads call RAND_poll() concurrently.
458 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
459 combination of a flag and a thread ID variable.
460 Otherwise while one thread is in ssleay_rand_bytes (which sets the
461 flag), *other* threads can enter ssleay_add_bytes without obeying
462 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
463 that they do not hold after the first thread unsets add_do_not_lock).
466 *) Change bctest again: '-x' expressions are not available in all
470 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
472 *) Fix a couple of memory leaks in PKCS7_dataDecode()
473 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
475 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
476 the default extension for executables, if any. Also, make the perl
477 scripts that use symlink() to test if it really exists and use "cp"
478 if it doesn't. All this made OpenSSL compilable and installable in
482 *) Fix for asn1_GetSequence() for indefinite length constructed data.
483 If SEQUENCE is length is indefinite just set c->slen to the total
484 amount of data available.
485 [Steve Henson, reported by shige@FreeBSD.org]
486 [This change does not apply to 0.9.7.]
488 *) Change bctest to avoid here-documents inside command substitution
489 (workaround for FreeBSD /bin/sh bug).
490 For compatibility with Ultrix, avoid shell functions (introduced
491 in the bctest version that searches along $PATH).
494 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
495 with des_encrypt() defined on some operating systems, like Solaris
499 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
500 On the Importance of Eliminating Errors in Cryptographic
501 Computations, J. Cryptology 14 (2001) 2, 101-119,
502 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
505 *) MIPS assembler BIGNUM division bug fix.
508 *) Disabled incorrect Alpha assembler code.
511 *) Fix PKCS#7 decode routines so they correctly update the length
512 after reading an EOC for the EXPLICIT tag.
514 [This change does not apply to 0.9.7.]
516 *) Fix bug in PKCS#12 key generation routines. This was triggered
517 if a 3DES key was generated with a 0 initial byte. Include
518 PKCS12_BROKEN_KEYGEN compilation option to retain the old
519 (but broken) behaviour.
522 *) Enhance bctest to search for a working bc along $PATH and print
524 [Tim Rice <tim@multitalents.net> via Richard Levitte]
526 *) Fix memory leaks in err.c: free err_data string if necessary;
527 don't write to the wrong index in ERR_set_error_data.
530 *) Implement ssl23_peek (analogous to ssl23_read), which previously
534 *) Replace rdtsc with _emit statements for VC++ version 5.
535 [Jeremy Cooper <jeremy@baymoo.org>]
537 *) Make it possible to reuse SSLv2 sessions.
540 *) In copy_email() check for >= 0 as a return value for
541 X509_NAME_get_index_by_NID() since 0 is a valid index.
542 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
544 *) Avoid coredump with unsupported or invalid public keys by checking if
545 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
546 PKCS7_verify() fails with non detached data.
549 *) Don't use getenv in library functions when run as setuid/setgid.
550 New function OPENSSL_issetugid().
553 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
554 due to incorrect handling of multi-threading:
556 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
558 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
560 3. Count how many times MemCheck_off() has been called so that
561 nested use can be treated correctly. This also avoids
562 inband-signalling in the previous code (which relied on the
563 assumption that thread ID 0 is impossible).
566 *) Add "-rand" option also to s_client and s_server.
569 *) Fix CPU detection on Irix 6.x.
570 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
571 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
573 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
576 [This change does not apply to 0.9.7.]
578 *) Use the cached encoding of an X509_NAME structure rather than
579 copying it. This is apparently the reason for the libsafe "errors"
580 but the code is actually correct.
583 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
584 Bleichenbacher's DSA attack.
585 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
586 to be set and top=0 forces the highest bit to be set; top=-1 is new
587 and leaves the highest bit random.
588 [Ulf Moeller, Bodo Moeller]
590 *) In the NCONF_...-based implementations for CONF_... queries
591 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
592 a temporary CONF structure with the data component set to NULL
593 (which gives segmentation faults in lh_retrieve).
594 Instead, use NULL for the CONF pointer in CONF_get_string and
595 CONF_get_number (which may use environment variables) and directly
596 return NULL from CONF_get_section.
599 *) Fix potential buffer overrun for EBCDIC.
602 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
603 keyUsage if basicConstraints absent for a CA.
606 *) Make SMIME_write_PKCS7() write mail header values with a format that
607 is more generally accepted (no spaces before the semicolon), since
608 some programs can't parse those values properly otherwise. Also make
609 sure BIO's that break lines after each write do not create invalid
613 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
614 macros previously used would not encode an empty SEQUENCE OF
615 and break the signature.
617 [This change does not apply to 0.9.7.]
619 *) Zero the premaster secret after deriving the master secret in
623 *) Add some EVP_add_digest_alias registrations (as found in
624 OpenSSL_add_all_digests()) to SSL_library_init()
625 aka OpenSSL_add_ssl_algorithms(). This provides improved
626 compatibility with peers using X.509 certificates
627 with unconventional AlgorithmIdentifier OIDs.
630 *) Fix for Irix with NO_ASM.
631 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
633 *) ./config script fixes.
634 [Ulf Moeller, Richard Levitte]
636 *) Fix 'openssl passwd -1'.
639 *) Change PKCS12_key_gen_asc() so it can cope with non null
640 terminated strings whose length is passed in the passlen
641 parameter, for example from PEM callbacks. This was done
642 by adding an extra length parameter to asc2uni().
643 [Steve Henson, reported by <oddissey@samsung.co.kr>]
645 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
646 call failed, free the DSA structure.
649 *) Fix to uni2asc() to cope with zero length Unicode strings.
650 These are present in some PKCS#12 files.
653 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
654 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
655 when writing a 32767 byte record.
656 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
658 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
659 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
661 (RSA objects have a reference count access to which is protected
662 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
663 so they are meant to be shared between threads.)
664 [Bodo Moeller, Geoff Thorpe; original patch submitted by
665 "Reddie, Steven" <Steven.Reddie@ca.com>]
667 *) Fix a deadlock in CRYPTO_mem_leaks().
670 *) Use better test patterns in bntest.
673 *) rand_win.c fix for Borland C.
676 *) BN_rshift bugfix for n == 0.
679 *) Add a 'bctest' script that checks for some known 'bc' bugs
680 so that 'make test' does not abort just because 'bc' is broken.
683 *) Store verify_result within SSL_SESSION also for client side to
684 avoid potential security hole. (Re-used sessions on the client side
685 always resulted in verify_result==X509_V_OK, not using the original
686 result of the server certificate verification.)
689 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
690 SSL3_RT_APPLICATION_DATA, return 0.
691 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
695 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
696 releases, have been re-implemented by renaming the previous
697 implementations of ssl2_read and ssl3_read to ssl2_read_internal
698 and ssl3_read_internal, respectively, and adding 'peek' parameters
699 to them. The new ssl[23]_{read,peek} functions are calls to
700 ssl[23]_read_internal with the 'peek' flag set appropriately.
701 A 'peek' parameter has also been added to ssl3_read_bytes, which
702 does the actual work for ssl3_read_internal.
705 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
706 the method-specific "init()" handler. Also clean up ex_data after
707 calling the method-specific "finish()" handler. Previously, this was
708 happening the other way round.
711 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
712 The previous value, 12, was not always sufficient for BN_mod_exp().
715 *) Make sure that shared libraries get the internal name engine with
716 the full version number and not just 0. This should mark the
717 shared libraries as not backward compatible. Of course, this should
718 be changed again when we can guarantee backward binary compatibility.
721 *) Fix typo in get_cert_by_subject() in by_dir.c
722 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
724 *) Rework the system to generate shared libraries:
726 - Make note of the expected extension for the shared libraries and
727 if there is a need for symbolic links from for example libcrypto.so.0
728 to libcrypto.so.0.9.7. There is extended info in Configure for
731 - Make as few rebuilds of the shared libraries as possible.
733 - Still avoid linking the OpenSSL programs with the shared libraries.
735 - When installing, install the shared libraries separately from the
739 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
741 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
742 and not in SSL_clear because the latter is also used by the
743 accept/connect functions; previously, the settings made by
744 SSL_set_read_ahead would be lost during the handshake.
745 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
747 *) Correct util/mkdef.pl to be selective about disabled algorithms.
748 Previously, it would create entries for disableed algorithms no
752 *) Added several new manual pages for SSL_* function.
755 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
757 *) In ssl23_get_client_hello, generate an error message when faced
758 with an initial SSL 3.0/TLS record that is too small to contain the
759 first two bytes of the ClientHello message, i.e. client_version.
760 (Note that this is a pathologic case that probably has never happened
761 in real life.) The previous approach was to use the version number
762 from the record header as a substitute; but our protocol choice
763 should not depend on that one because it is not authenticated
764 by the Finished messages.
767 *) More robust randomness gathering functions for Windows.
768 [Jeffrey Altman <jaltman@columbia.edu>]
770 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
771 not set then we don't setup the error code for issuer check errors
772 to avoid possibly overwriting other errors which the callback does
773 handle. If an application does set the flag then we assume it knows
774 what it is doing and can handle the new informational codes
778 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
779 a general "ANY" type, as such it should be able to decode anything
780 including tagged types. However it didn't check the class so it would
781 wrongly interpret tagged types in the same way as their universal
782 counterpart and unknown types were just rejected. Changed so that the
783 tagged and unknown types are handled in the same way as a SEQUENCE:
784 that is the encoding is stored intact. There is also a new type
785 "V_ASN1_OTHER" which is used when the class is not universal, in this
786 case we have no idea what the actual type is so we just lump them all
790 *) On VMS, stdout may very well lead to a file that is written to
791 in a record-oriented fashion. That means that every write() will
792 write a separate record, which will be read separately by the
793 programs trying to read from it. This can be very confusing.
795 The solution is to put a BIO filter in the way that will buffer
796 text until a linefeed is reached, and then write everything a
797 line at a time, so every record written will be an actual line,
798 not chunks of lines and not (usually doesn't happen, but I've
799 seen it once) several lines in one record. BIO_f_linebuffer() is
802 Currently, it's a VMS-only method, because that's where it has
803 been tested well enough.
806 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
807 it can return incorrect results.
808 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
809 but it was in 0.9.6-beta[12].)
812 *) Disable the check for content being present when verifying detached
813 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
814 include zero length content when signing messages.
817 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
818 BIO_ctrl (for BIO pairs).
821 *) Add DSO method for VMS.
824 *) Bug fix: Montgomery multiplication could produce results with the
828 *) Add RPM specification openssl.spec and modify it to build three
829 packages. The default package contains applications, application
830 documentation and run-time libraries. The devel package contains
831 include files, static libraries and function documentation. The
832 doc package contains the contents of the doc directory. The original
833 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
836 *) Add a large number of documentation files for many SSL routines.
837 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
839 *) Add a configuration entry for Sony News 4.
840 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
842 *) Don't set the two most significant bits to one when generating a
843 random number < q in the DSA library.
846 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
847 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
848 the underlying transport is blocking) if a handshake took place.
849 (The default behaviour is needed by applications such as s_client
850 and s_server that use select() to determine when to use SSL_read;
851 but for applications that know in advance when to expect data, it
852 just makes things more complicated.)
855 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
859 *) Add a few more EBCDIC conditionals that make `req' and `x509'
860 work better on such systems.
861 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
863 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
864 Update PKCS12_parse() so it copies the friendlyName and the
865 keyid to the certificates aux info.
868 *) Fix bug in PKCS7_verify() which caused an infinite loop
869 if there was more than one signature.
870 [Sven Uszpelkat <su@celocom.de>]
872 *) Major change in util/mkdef.pl to include extra information
873 about each symbol, as well as presentig variables as well
874 as functions. This change means that there's n more need
875 to rebuild the .num files when some algorithms are excluded.
878 *) Allow the verify time to be set by an application,
879 rather than always using the current time.
882 *) Phase 2 verify code reorganisation. The certificate
883 verify code now looks up an issuer certificate by a
884 number of criteria: subject name, authority key id
885 and key usage. It also verifies self signed certificates
886 by the same criteria. The main comparison function is
887 X509_check_issued() which performs these checks.
889 Lot of changes were necessary in order to support this
890 without completely rewriting the lookup code.
892 Authority and subject key identifier are now cached.
894 The LHASH 'certs' is X509_STORE has now been replaced
895 by a STACK_OF(X509_OBJECT). This is mainly because an
896 LHASH can't store or retrieve multiple objects with
899 As a result various functions (which were all internal
900 use only) have changed to handle the new X509_STORE
901 structure. This will break anything that messed round
902 with X509_STORE internally.
904 The functions X509_STORE_add_cert() now checks for an
905 exact match, rather than just subject name.
907 The X509_STORE API doesn't directly support the retrieval
908 of multiple certificates matching a given criteria, however
909 this can be worked round by performing a lookup first
910 (which will fill the cache with candidate certificates)
911 and then examining the cache for matches. This is probably
912 the best we can do without throwing out X509_LOOKUP
913 entirely (maybe later...).
915 The X509_VERIFY_CTX structure has been enhanced considerably.
917 All certificate lookup operations now go via a get_issuer()
918 callback. Although this currently uses an X509_STORE it
919 can be replaced by custom lookups. This is a simple way
920 to bypass the X509_STORE hackery necessary to make this
921 work and makes it possible to use more efficient techniques
922 in future. A very simple version which uses a simple
923 STACK for its trusted certificate store is also provided
924 using X509_STORE_CTX_trusted_stack().
926 The verify_cb() and verify() callbacks now have equivalents
927 in the X509_STORE_CTX structure.
929 X509_STORE_CTX also has a 'flags' field which can be used
930 to customise the verify behaviour.
933 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
934 excludes S/MIME capabilities.
937 *) When a certificate request is read in keep a copy of the
938 original encoding of the signed data and use it when outputing
939 again. Signatures then use the original encoding rather than
940 a decoded, encoded version which may cause problems if the
941 request is improperly encoded.
944 *) For consistency with other BIO_puts implementations, call
945 buffer_write(b, ...) directly in buffer_puts instead of calling
948 In BIO_puts, increment b->num_write as in BIO_write.
949 [Peter.Sylvester@EdelWeb.fr]
951 *) Fix BN_mul_word for the case where the word is 0. (We have to use
952 BN_zero, we may not return a BIGNUM with an array consisting of
956 *) Avoid calling abort() from within the library when problems are
957 detected, except if preprocessor symbols have been defined
958 (such as REF_CHECK, BN_DEBUG etc.).
961 *) New openssl application 'rsautl'. This utility can be
962 used for low level RSA operations. DER public key
963 BIO/fp routines also added.
966 *) New Configure entry and patches for compiling on QNX 4.
967 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
969 *) A demo state-machine implementation was sponsored by
970 Nuron (http://www.nuron.com/) and is now available in
974 *) New options added to the 'dgst' utility for signature
975 generation and verification.
978 *) Unrecognized PKCS#7 content types are now handled via a
979 catch all ASN1_TYPE structure. This allows unsupported
980 types to be stored as a "blob" and an application can
981 encode and decode it manually.
984 *) Fix various signed/unsigned issues to make a_strex.c
986 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
988 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
989 length if passed a buffer. ASN1_INTEGER_to_BN failed
990 if passed a NULL BN and its argument was negative.
991 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
993 *) Modification to PKCS#7 encoding routines to output definite
994 length encoding. Since currently the whole structures are in
995 memory there's not real point in using indefinite length
996 constructed encoding. However if OpenSSL is compiled with
997 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
1000 *) Added BIO_vprintf() and BIO_vsnprintf().
1003 *) Added more prefixes to parse for in the the strings written
1004 through a logging bio, to cover all the levels that are available
1005 through syslog. The prefixes are now:
1007 PANIC, EMERG, EMR => LOG_EMERG
1008 ALERT, ALR => LOG_ALERT
1009 CRIT, CRI => LOG_CRIT
1010 ERROR, ERR => LOG_ERR
1011 WARNING, WARN, WAR => LOG_WARNING
1012 NOTICE, NOTE, NOT => LOG_NOTICE
1013 INFO, INF => LOG_INFO
1014 DEBUG, DBG => LOG_DEBUG
1016 and as before, if none of those prefixes are present at the
1017 beginning of the string, LOG_ERR is chosen.
1019 On Win32, the LOG_* levels are mapped according to this:
1021 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
1022 LOG_WARNING => EVENTLOG_WARNING_TYPE
1023 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
1027 *) Made it possible to reconfigure with just the configuration
1028 argument "reconf" or "reconfigure". The command line arguments
1029 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
1030 and are retrieved from there when reconfiguring.
1034 [Assar Westerlund <assar@sics.se>, Richard Levitte]
1036 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
1039 *) The obj_dat.pl script was messing up the sorting of object
1040 names. The reason was that it compared the quoted version
1041 of strings as a result "OCSP" > "OCSP Signing" because
1042 " > SPACE. Changed script to store unquoted versions of
1043 names and add quotes on output. It was also omitting some
1044 names from the lookup table if they were given a default
1045 value (that is if SN is missing it is given the same
1046 value as LN and vice versa), these are now added on the
1047 grounds that if an object has a name we should be able to
1048 look it up. Finally added warning output when duplicate
1049 short or long names are found.
1052 *) Changes needed for Tandem NSK.
1053 [Scott Uroff <scott@xypro.com>]
1055 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
1056 RSA_padding_check_SSLv23(), special padding was never detected
1057 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
1058 version rollback attacks was not effective.
1060 In s23_clnt.c, don't use special rollback-attack detection padding
1061 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
1062 client; similarly, in s23_srvr.c, don't do the rollback check if
1063 SSL 2.0 is the only protocol enabled in the server.
1066 *) Make it possible to get hexdumps of unprintable data with 'openssl
1067 asn1parse'. By implication, the functions ASN1_parse_dump() and
1068 BIO_dump_indent() are added.
1071 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
1072 these print out strings and name structures based on various
1073 flags including RFC2253 support and proper handling of
1074 multibyte characters. Added options to the 'x509' utility
1075 to allow the various flags to be set.
1078 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
1079 Also change the functions X509_cmp_current_time() and
1080 X509_gmtime_adj() work with an ASN1_TIME structure,
1081 this will enable certificates using GeneralizedTime in validity
1082 dates to be checked.
1085 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1086 negative public key encodings) on by default,
1087 NO_NEG_PUBKEY_BUG can be set to disable it.
1090 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1091 content octets. An i2c_ASN1_OBJECT is unnecessary because
1092 the encoding can be trivially obtained from the structure.
1095 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1096 not read locks (CRYPTO_r_[un]lock).
1099 *) A first attempt at creating official support for shared
1100 libraries through configuration. I've kept it so the
1101 default is static libraries only, and the OpenSSL programs
1102 are always statically linked for now, but there are
1103 preparations for dynamic linking in place.
1104 This has been tested on Linux and Tru64.
1107 *) Randomness polling function for Win9x, as described in:
1108 Peter Gutmann, Software Generation of Practically Strong
1112 *) Fix so PRNG is seeded in req if using an already existing
1116 *) New options to smime application. -inform and -outform
1117 allow alternative formats for the S/MIME message including
1118 PEM and DER. The -content option allows the content to be
1119 specified separately. This should allow things like Netscape
1120 form signing output easier to verify.
1123 *) Fix the ASN1 encoding of tags using the 'long form'.
1126 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1127 STRING types. These convert content octets to and from the
1128 underlying type. The actual tag and length octets are
1129 already assumed to have been read in and checked. These
1130 are needed because all other string types have virtually
1131 identical handling apart from the tag. By having versions
1132 of the ASN1 functions that just operate on content octets
1133 IMPLICIT tagging can be handled properly. It also allows
1134 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1135 and ASN1_INTEGER are identical apart from the tag.
1138 *) Change the handling of OID objects as follows:
1140 - New object identifiers are inserted in objects.txt, following
1141 the syntax given in objects.README.
1142 - objects.pl is used to process obj_mac.num and create a new
1144 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1147 This is currently kind of a hack, and the perl code in objects.pl
1148 isn't very elegant, but it works as I intended. The simplest way
1149 to check that it worked correctly is to look in obj_dat.h and
1150 check the array nid_objs and make sure the objects haven't moved
1151 around (this is important!). Additions are OK, as well as
1152 consistent name changes.
1155 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1158 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1159 The given file adds to whatever has already been seeded into the
1160 random pool through the RANDFILE configuration file option or
1161 environment variable, or the default random state file.
1164 *) mkstack.pl now sorts each macro group into lexical order.
1165 Previously the output order depended on the order the files
1166 appeared in the directory, resulting in needless rewriting
1170 *) Patches to make OpenSSL compile under Win32 again. Mostly
1171 work arounds for the VC++ problem that it treats func() as
1172 func(void). Also stripped out the parts of mkdef.pl that
1173 added extra typesafe functions: these no longer exist.
1176 *) Reorganisation of the stack code. The macros are now all
1177 collected in safestack.h . Each macro is defined in terms of
1178 a "stack macro" of the form SKM_<name>(type, a, b). The
1179 DEBUG_SAFESTACK is now handled in terms of function casts,
1180 this has the advantage of retaining type safety without the
1181 use of additional functions. If DEBUG_SAFESTACK is not defined
1182 then the non typesafe macros are used instead. Also modified the
1183 mkstack.pl script to handle the new form. Needs testing to see
1184 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1185 the default if no major problems. Similar behaviour for ASN1_SET_OF
1186 and PKCS12_STACK_OF.
1189 *) When some versions of IIS use the 'NET' form of private key the
1190 key derivation algorithm is different. Normally MD5(password) is
1191 used as a 128 bit RC4 key. In the modified case
1192 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1193 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1194 as the old Netscape_RSA functions except they have an additional
1195 'sgckey' parameter which uses the modified algorithm. Also added
1196 an -sgckey command line option to the rsa utility. Thanks to
1197 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1198 algorithm to openssl-dev.
1201 *) The evp_local.h macros were using 'c.##kname' which resulted in
1202 invalid expansion on some systems (SCO 5.0.5 for example).
1203 Corrected to 'c.kname'.
1204 [Phillip Porch <root@theporch.com>]
1206 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1207 a STACK of email addresses from a certificate or request, these look
1208 in the subject name and the subject alternative name extensions and
1209 omit any duplicate addresses.
1212 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1213 This makes DSA verification about 2 % faster.
1216 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1217 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1218 plus overhead for 1024 bit moduli).
1219 This makes exponentiations about 0.5 % faster for 1024 bit
1220 exponents (as measured by "openssl speed rsa2048").
1223 *) Rename memory handling macros to avoid conflicts with other
1225 Malloc => OPENSSL_malloc
1226 Malloc_locked => OPENSSL_malloc_locked
1227 Realloc => OPENSSL_realloc
1228 Free => OPENSSL_free
1231 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1232 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1235 *) CygWin32 support.
1236 [John Jarvie <jjarvie@newsguy.com>]
1238 *) The type-safe stack code has been rejigged. It is now only compiled
1239 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1240 by default all type-specific stack functions are "#define"d back to
1241 standard stack functions. This results in more streamlined output
1242 but retains the type-safety checking possibilities of the original
1246 *) The STACK code has been cleaned up, and certain type declarations
1247 that didn't make a lot of sense have been brought in line. This has
1248 also involved a cleanup of sorts in safestack.h to more correctly
1249 map type-safe stack functions onto their plain stack counterparts.
1250 This work has also resulted in a variety of "const"ifications of
1251 lots of the code, especially "_cmp" operations which should normally
1252 be prototyped with "const" parameters anyway.
1255 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1256 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1257 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1258 where all of 'md' is used each time the PRNG is used, but 'state'
1259 is used only indexed by a cyclic counter. As entropy may not be
1260 well distributed from the beginning, 'md' is important as a
1261 chaining variable. However, the output function chains only half
1262 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1263 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1264 in all of 'state' being rewritten, with the new values depending
1265 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1268 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1269 the handshake is continued after ssl_verify_cert_chain();
1270 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1271 can lead to 'unexplainable' connection aborts later.
1272 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1274 *) Major EVP API cipher revision.
1275 Add hooks for extra EVP features. This allows various cipher
1276 parameters to be set in the EVP interface. Support added for variable
1277 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1278 setting of RC2 and RC5 parameters.
1280 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1283 Remove lots of duplicated code from the EVP library. For example *every*
1284 cipher init() function handles the 'iv' in the same way according to the
1285 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1286 for CFB and OFB modes they zero ctx->num.
1288 New functionality allows removal of S/MIME code RC2 hack.
1290 Most of the routines have the same form and so can be declared in terms
1293 By shifting this to the top level EVP_CipherInit() it can be removed from
1294 all individual ciphers. If the cipher wants to handle IVs or keys
1295 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1298 Change lots of functions like EVP_EncryptUpdate() to now return a
1299 value: although software versions of the algorithms cannot fail
1300 any installed hardware versions can.
1303 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1304 this option is set, tolerate broken clients that send the negotiated
1305 protocol version number instead of the requested protocol version
1309 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1310 i.e. non-zero for export ciphersuites, zero otherwise.
1311 Previous versions had this flag inverted, inconsistent with
1312 rsa_tmp_cb (..._TMP_RSA_CB).
1313 [Bodo Moeller; problem reported by Amit Chopra]
1315 *) Add missing DSA library text string. Work around for some IIS
1316 key files with invalid SEQUENCE encoding.
1319 *) Add a document (doc/standards.txt) that list all kinds of standards
1320 and so on that are implemented in OpenSSL.
1323 *) Enhance c_rehash script. Old version would mishandle certificates
1324 with the same subject name hash and wouldn't handle CRLs at all.
1325 Added -fingerprint option to crl utility, to support new c_rehash
1329 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1332 *) Fix for SSL server purpose checking. Server checking was
1333 rejecting certificates which had extended key usage present
1334 but no ssl client purpose.
1335 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1337 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1338 is a little unclear about how a blank password is handled.
1339 Since the password in encoded as a BMPString with terminating
1340 double NULL a zero length password would end up as just the
1341 double NULL. However no password at all is different and is
1342 handled differently in the PKCS#12 key generation code. NS
1343 treats a blank password as zero length. MSIE treats it as no
1344 password on export: but it will try both on import. We now do
1345 the same: PKCS12_parse() tries zero length and no password if
1346 the password is set to "" or NULL (NULL is now a valid password:
1347 it wasn't before) as does the pkcs12 application.
1350 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1351 perror when PEM_read_bio_X509_REQ fails, the error message must
1352 be obtained from the error queue.
1355 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1356 it in ERR_remove_state if appropriate, and change ERR_get_state
1357 accordingly to avoid race conditions (this is necessary because
1358 thread_hash is no longer constant once set).
1361 *) Bugfix for linux-elf makefile.one.
1364 *) RSA_get_default_method() will now cause a default
1365 RSA_METHOD to be chosen if one doesn't exist already.
1366 Previously this was only set during a call to RSA_new()
1367 or RSA_new_method(NULL) meaning it was possible for
1368 RSA_get_default_method() to return NULL.
1371 *) Added native name translation to the existing DSO code
1372 that will convert (if the flag to do so is set) filenames
1373 that are sufficiently small and have no path information
1374 into a canonical native form. Eg. "blah" converted to
1375 "libblah.so" or "blah.dll" etc.
1378 *) New function ERR_error_string_n(e, buf, len) which is like
1379 ERR_error_string(e, buf), but writes at most 'len' bytes
1380 including the 0 terminator. For ERR_error_string_n, 'buf'
1382 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1384 *) CONF library reworked to become more general. A new CONF
1385 configuration file reader "class" is implemented as well as a
1386 new functions (NCONF_*, for "New CONF") to handle it. The now
1387 old CONF_* functions are still there, but are reimplemented to
1388 work in terms of the new functions. Also, a set of functions
1389 to handle the internal storage of the configuration data is
1390 provided to make it easier to write new configuration file
1391 reader "classes" (I can definitely see something reading a
1392 configuration file in XML format, for example), called _CONF_*,
1393 or "the configuration storage API"...
1395 The new configuration file reading functions are:
1397 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1398 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1400 NCONF_default, NCONF_WIN32
1402 NCONF_dump_fp, NCONF_dump_bio
1404 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1405 NCONF_new creates a new CONF object. This works in the same way
1406 as other interfaces in OpenSSL, like the BIO interface.
1407 NCONF_dump_* dump the internal storage of the configuration file,
1408 which is useful for debugging. All other functions take the same
1409 arguments as the old CONF_* functions wth the exception of the
1410 first that must be a `CONF *' instead of a `LHASH *'.
1412 To make it easer to use the new classes with the old CONF_* functions,
1413 the function CONF_set_default_method is provided.
1416 *) Add '-tls1' option to 'openssl ciphers', which was already
1417 mentioned in the documentation but had not been implemented.
1418 (This option is not yet really useful because even the additional
1419 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1422 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1423 OpenSSL-based applications) load shared libraries and bind to
1424 them in a portable way.
1425 [Geoff Thorpe, with contributions from Richard Levitte]
1427 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1429 *) Make sure _lrotl and _lrotr are only used with MSVC.
1431 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1432 (the default implementation of RAND_status).
1434 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1435 to '-clrext' (= clear extensions), as intended and documented.
1436 [Bodo Moeller; inconsistency pointed out by Michael Attili
1439 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1440 was larger than the MD block size.
1441 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1443 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1444 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1445 using the passed key: if the passed key was a private key the result
1446 of X509_print(), for example, would be to print out all the private key
1450 *) des_quad_cksum() byte order bug fix.
1451 [Ulf Möller, using the problem description in krb4-0.9.7, where
1452 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1454 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1456 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1458 *) For easily testing in shell scripts whether some command
1459 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1460 returns with exit code 0 iff no command of the given name is available.
1461 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1462 the output goes to stdout and nothing is printed to stderr.
1463 Additional arguments are always ignored.
1465 Since for each cipher there is a command of the same name,
1466 the 'no-cipher' compilation switches can be tested this way.
1468 ('openssl no-XXX' is not able to detect pseudo-commands such
1469 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1472 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1475 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1476 is set; it will be thrown away anyway because each handshake creates
1478 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1479 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1480 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1481 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1484 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1485 'Q' and 'R' lose their special meanings (quit/renegotiate).
1486 This is part of what -quiet does; unlike -quiet, -ign_eof
1487 does not suppress any output.
1490 *) Add compatibility options to the purpose and trust code. The
1491 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1492 accepts a certificate or CA, this was the previous behaviour,
1493 with all the associated security issues.
1495 X509_TRUST_COMPAT is the old trust behaviour: only and
1496 automatically trust self signed roots in certificate store. A
1497 new trust setting X509_TRUST_DEFAULT is used to specify that
1498 a purpose has no associated trust setting and it should instead
1499 use the value in the default purpose.
1502 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1503 and fix a memory leak.
1506 *) In util/mkerr.pl (which implements 'make errors'), preserve
1507 reason strings from the previous version of the .c file, as
1508 the default to have only downcase letters (and digits) in
1509 automatically generated reasons codes is not always appropriate.
1512 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1513 using strerror. Previously, ERR_reason_error_string() returned
1514 library names as reason strings for SYSerr; but SYSerr is a special
1515 case where small numbers are errno values, not library numbers.
1518 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1519 converts DSA parameters into DH parameters. (When creating parameters,
1520 DSA_generate_parameters is used.)
1523 *) Include 'length' (recommended exponent length) in C code generated
1524 by 'openssl dhparam -C'.
1527 *) The second argument to set_label in perlasm was already being used
1528 so couldn't be used as a "file scope" flag. Moved to third argument
1532 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1533 instead of RAND_bytes for encryption IVs and salts.
1536 *) Include RAND_status() into RAND_METHOD instead of implementing
1537 it only for md_rand.c Otherwise replacing the PRNG by calling
1538 RAND_set_rand_method would be impossible.
1541 *) Don't let DSA_generate_key() enter an infinite loop if the random
1542 number generation fails.
1545 *) New 'rand' application for creating pseudo-random output.
1548 *) Added configuration support for Linux/IA64
1549 [Rolf Haberrecker <rolf@suse.de>]
1551 *) Assembler module support for Mingw32.
1554 *) Shared library support for HPUX (in shlib/).
1555 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1557 *) Shared library support for Solaris gcc.
1558 [Lutz Behnke <behnke@trustcenter.de>]
1560 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1562 *) PKCS7_encrypt() was adding text MIME headers twice because they
1563 were added manually and by SMIME_crlf_copy().
1566 *) In bntest.c don't call BN_rand with zero bits argument.
1567 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1569 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1570 case was implemented. This caused BN_div_recp() to fail occasionally.
1573 *) Add an optional second argument to the set_label() in the perl
1574 assembly language builder. If this argument exists and is set
1575 to 1 it signals that the assembler should use a symbol whose
1576 scope is the entire file, not just the current function. This
1577 is needed with MASM which uses the format label:: for this scope.
1578 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1580 *) Change the ASN1 types so they are typedefs by default. Before
1581 almost all types were #define'd to ASN1_STRING which was causing
1582 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1586 *) Change names of new functions to the new get1/get0 naming
1587 convention: After 'get1', the caller owns a reference count
1588 and has to call ..._free; 'get0' returns a pointer to some
1589 data structure without incrementing reference counters.
1590 (Some of the existing 'get' functions increment a reference
1591 counter, some don't.)
1592 Similarly, 'set1' and 'add1' functions increase reference
1593 counters or duplicate objects.
1596 *) Allow for the possibility of temp RSA key generation failure:
1597 the code used to assume it always worked and crashed on failure.
1600 *) Fix potential buffer overrun problem in BIO_printf().
1601 [Ulf Möller, using public domain code by Patrick Powell; problem
1602 pointed out by David Sacerdote <das33@cornell.edu>]
1604 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1605 RAND_egd() and RAND_status(). In the command line application,
1606 the EGD socket can be specified like a seed file using RANDFILE
1610 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1611 Some CAs (e.g. Verisign) distribute certificates in this form.
1614 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1615 list to exclude them. This means that no special compilation option
1616 is needed to use anonymous DH: it just needs to be included in the
1620 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1621 EVP_MD_type. The old functionality is available in a new macro called
1622 EVP_MD_md(). Change code that uses it and update docs.
1625 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1626 where the 'void *' argument is replaced by a function pointer argument.
1627 Previously 'void *' was abused to point to functions, which works on
1628 many platforms, but is not correct. As these functions are usually
1629 called by macros defined in OpenSSL header files, most source code
1630 should work without changes.
1633 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1634 sections with information on -D... compiler switches used for
1635 compiling the library so that applications can see them. To enable
1636 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1637 must be defined. E.g.,
1638 #define OPENSSL_ALGORITHM_DEFINES
1639 #include <openssl/opensslconf.h>
1640 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1641 [Richard Levitte, Ulf and Bodo Möller]
1643 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1647 *) Change the 'other' type in certificate aux info to a STACK_OF
1648 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1649 the required ASN1 format: arbitrary types determined by an OID.
1652 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1653 argument to 'req'. This is not because the function is newer or
1654 better than others it just uses the work 'NEW' in the certificate
1655 request header lines. Some software needs this.
1658 *) Reorganise password command line arguments: now passwords can be
1659 obtained from various sources. Delete the PEM_cb function and make
1660 it the default behaviour: i.e. if the callback is NULL and the
1661 usrdata argument is not NULL interpret it as a null terminated pass
1662 phrase. If usrdata and the callback are NULL then the pass phrase
1663 is prompted for as usual.
1666 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1667 the support is automatically enabled. The resulting binaries will
1668 autodetect the card and use it if present.
1669 [Ben Laurie and Compaq Inc.]
1671 *) Work around for Netscape hang bug. This sends certificate request
1672 and server done in one record. Since this is perfectly legal in the
1673 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1674 the bugs/SSLv3 entry for more info.
1677 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1680 *) Add -rand argument to smime and pkcs12 applications and read/write
1684 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1687 *) Add command line password options to the remaining applications.
1690 *) Bug fix for BN_div_recp() for numerators with an even number of
1694 *) More tests in bntest.c, and changed test_bn output.
1697 *) ./config recognizes MacOS X now.
1700 *) Bug fix for BN_div() when the first words of num and divsor are
1701 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1704 *) Add support for various broken PKCS#8 formats, and command line
1705 options to produce them.
1708 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1709 get temporary BIGNUMs from a BN_CTX.
1712 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1716 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1717 include a #define from the old name to the new. The original intent
1718 was that statically linked binaries could for example just call
1719 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1720 link with digests. This never worked becayse SSLeay_add_all_digests()
1721 and SSLeay_add_all_ciphers() were in the same source file so calling
1722 one would link with the other. They are now in separate source files.
1725 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1728 *) Use a less unusual form of the Miller-Rabin primality test (it used
1729 a binary algorithm for exponentiation integrated into the Miller-Rabin
1730 loop, our standard modexp algorithms are faster).
1733 *) Support for the EBCDIC character set completed.
1734 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1736 *) Source code cleanups: use const where appropriate, eliminate casts,
1737 use void * instead of char * in lhash.
1740 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1741 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1742 this the server could overwrite ephemeral keys that the client
1746 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1747 using 50 iterations of the Rabin-Miller test.
1749 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1750 iterations of the Rabin-Miller test as required by the appendix
1751 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1752 As BN_is_prime_fasttest includes trial division, DSA parameter
1753 generation becomes much faster.
1755 This implies a change for the callback functions in DSA_is_prime
1756 and DSA_generate_parameters: The callback function is called once
1757 for each positive witness in the Rabin-Miller test, not just
1758 occasionally in the inner loop; and the parameters to the
1759 callback function now provide an iteration count for the outer
1760 loop rather than for the current invocation of the inner loop.
1761 DSA_generate_parameters additionally can call the callback
1762 function with an 'iteration count' of -1, meaning that a
1763 candidate has passed the trial division test (when q is generated
1764 from an application-provided seed, trial division is skipped).
1767 *) New function BN_is_prime_fasttest that optionally does trial
1768 division before starting the Rabin-Miller test and has
1769 an additional BN_CTX * argument (whereas BN_is_prime always
1770 has to allocate at least one BN_CTX).
1771 'callback(1, -1, cb_arg)' is called when a number has passed the
1772 trial division stage.
1775 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1779 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1782 *) New function BN_pseudo_rand().
1785 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1786 bignum version of BN_from_montgomery() with the working code from
1787 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1791 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1792 made it impossible to use the same SSL_SESSION data structure in
1793 SSL2 clients in multiple threads.
1796 *) The return value of RAND_load_file() no longer counts bytes obtained
1797 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1798 to seed the PRNG (previously an explicit byte count was required).
1799 [Ulf Möller, Bodo Möller]
1801 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1802 used (char *) instead of (void *) and had casts all over the place.
1805 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1808 *) Retain source code compatibility for BN_prime_checks macro:
1809 BN_is_prime(..., BN_prime_checks, ...) now uses
1810 BN_prime_checks_for_size to determine the appropriate number of
1811 Rabin-Miller iterations.
1814 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1815 DH_CHECK_P_NOT_SAFE_PRIME.
1816 (Check if this is true? OpenPGP calls them "strong".)
1819 *) Merge the functionality of "dh" and "gendh" programs into a new program
1820 "dhparam". The old programs are retained for now but will handle DH keys
1821 (instead of parameters) in future.
1824 *) Make the ciphers, s_server and s_client programs check the return values
1825 when a new cipher list is set.
1828 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1829 ciphers. Before when the 56bit ciphers were enabled the sorting was
1832 The syntax for the cipher sorting has been extended to support sorting by
1833 cipher-strength (using the strength_bits hard coded in the tables).
1834 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1836 Fix a bug in the cipher-command parser: when supplying a cipher command
1837 string with an "undefined" symbol (neither command nor alphanumeric
1838 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1839 an error is flagged.
1841 Due to the strength-sorting extension, the code of the
1842 ssl_create_cipher_list() function was completely rearranged. I hope that
1843 the readability was also increased :-)
1844 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1846 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1847 for the first serial number and places 2 in the serial number file. This
1848 avoids problems when the root CA is created with serial number zero and
1849 the first user certificate has the same issuer name and serial number
1853 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1854 the new code. Add documentation for this stuff.
1857 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1858 X509_*() to X509at_*() on the grounds that they don't handle X509
1859 structures and behave in an analagous way to the X509v3 functions:
1860 they shouldn't be called directly but wrapper functions should be used
1863 So we also now have some wrapper functions that call the X509at functions
1864 when passed certificate requests. (TO DO: similar things can be done with
1865 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1866 things. Some of these need some d2i or i2d and print functionality
1867 because they handle more complex structures.)
1870 *) Add missing #ifndefs that caused missing symbols when building libssl
1871 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1872 NO_RSA in ssl/s2*.c.
1873 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1875 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1876 has a return value which indicates the quality of the random data
1877 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1878 error queue. New function RAND_pseudo_bytes() generates output that is
1879 guaranteed to be unique but not unpredictable. RAND_add is like
1880 RAND_seed, but takes an extra argument for an entropy estimate
1881 (RAND_seed always assumes full entropy).
1884 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1885 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1886 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1887 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1888 false-positive rate of at most 2^-80 for random input.
1891 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1894 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1895 in the 0.9.5 release), this returns the chain
1896 from an X509_CTX structure with a dup of the stack and all
1897 the X509 reference counts upped: so the stack will exist
1898 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1901 Also make SSL_SESSION_print() print out the verify return
1905 *) Add manpage for the pkcs12 command. Also change the default
1906 behaviour so MAC iteration counts are used unless the new
1907 -nomaciter option is used. This improves file security and
1908 only older versions of MSIE (4.0 for example) need it.
1911 *) Honor the no-xxx Configure options when creating .DEF files.
1914 *) Add PKCS#10 attributes to field table: challengePassword,
1915 unstructuredName and unstructuredAddress. These are taken from
1916 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1917 international characters are used.
1919 More changes to X509_ATTRIBUTE code: allow the setting of types
1920 based on strings. Remove the 'loc' parameter when adding
1921 attributes because these will be a SET OF encoding which is sorted
1925 *) Initial changes to the 'req' utility to allow request generation
1926 automation. This will allow an application to just generate a template
1927 file containing all the field values and have req construct the
1930 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1931 used all over the place including certificate requests and PKCS#7
1932 structures. They are currently handled manually where necessary with
1933 some primitive wrappers for PKCS#7. The new functions behave in a
1934 manner analogous to the X509 extension functions: they allow
1935 attributes to be looked up by NID and added.
1937 Later something similar to the X509V3 code would be desirable to
1938 automatically handle the encoding, decoding and printing of the
1939 more complex types. The string types like challengePassword can
1940 be handled by the string table functions.
1942 Also modified the multi byte string table handling. Now there is
1943 a 'global mask' which masks out certain types. The table itself
1944 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1945 is useful when for example there is only one permissible type
1946 (as in countryName) and using the mask might result in no valid
1950 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1951 SSL_get_peer_finished to allow applications to obtain the latest
1952 Finished messages sent to the peer or expected from the peer,
1953 respectively. (SSL_get_peer_finished is usually the Finished message
1954 actually received from the peer, otherwise the protocol will be aborted.)
1956 As the Finished message are message digests of the complete handshake
1957 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1958 be used for external authentication procedures when the authentication
1959 provided by SSL/TLS is not desired or is not enough.
1962 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1963 the host supports BWX extension and if Compaq C is present on the
1964 $PATH. Just exploiting of the BWX extension results in 20-30%
1965 performance kick for some algorithms, e.g. DES and RC4 to mention
1966 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1970 *) Add support for MS "fast SGC". This is arguably a violation of the
1971 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1972 weak crypto and after checking the certificate is SGC a second one
1973 with strong crypto. MS SGC stops the first handshake after receiving
1974 the server certificate message and sends a second client hello. Since
1975 a server will typically do all the time consuming operations before
1976 expecting any further messages from the client (server key exchange
1977 is the most expensive) there is little difference between the two.
1979 To get OpenSSL to support MS SGC we have to permit a second client
1980 hello message after we have sent server done. In addition we have to
1981 reset the MAC if we do get this second client hello.
1984 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1985 if a DER encoded private key is RSA or DSA traditional format. Changed
1986 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1987 format DER encoded private key. Newer code should use PKCS#8 format which
1988 has the key type encoded in the ASN1 structure. Added DER private key
1989 support to pkcs8 application.
1992 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1993 ciphersuites has been selected (as required by the SSL 3/TLS 1
1994 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1995 is set, we interpret this as a request to violate the specification
1996 (the worst that can happen is a handshake failure, and 'correct'
1997 behaviour would result in a handshake failure anyway).
2000 *) In SSL_CTX_add_session, take into account that there might be multiple
2001 SSL_SESSION structures with the same session ID (e.g. when two threads
2002 concurrently obtain them from an external cache).
2003 The internal cache can handle only one SSL_SESSION with a given ID,
2004 so if there's a conflict, we now throw out the old one to achieve
2008 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
2009 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
2010 some routines that use cipher OIDs: some ciphers do not have OIDs
2011 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
2015 *) Simplify the trust setting structure and code. Now we just have
2016 two sequences of OIDs for trusted and rejected settings. These will
2017 typically have values the same as the extended key usage extension
2018 and any application specific purposes.
2020 The trust checking code now has a default behaviour: it will just
2021 check for an object with the same NID as the passed id. Functions can
2022 be provided to override either the default behaviour or the behaviour
2023 for a given id. SSL client, server and email already have functions
2024 in place for compatibility: they check the NID and also return "trusted"
2025 if the certificate is self signed.
2028 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
2029 traditional format into an EVP_PKEY structure.
2032 *) Add a password callback function PEM_cb() which either prompts for
2033 a password if usr_data is NULL or otherwise assumes it is a null
2034 terminated password. Allow passwords to be passed on command line
2035 environment or config files in a few more utilities.
2038 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
2039 keys. Add some short names for PKCS#8 PBE algorithms and allow them
2040 to be specified on the command line for the pkcs8 and pkcs12 utilities.
2041 Update documentation.
2044 *) Support for ASN1 "NULL" type. This could be handled before by using
2045 ASN1_TYPE but there wasn't any function that would try to read a NULL
2046 and produce an error if it couldn't. For compatibility we also have
2047 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
2048 don't allocate anything because they don't need to.
2051 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
2053 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
2055 *) Rebuild of the memory allocation routines used by OpenSSL code and
2056 possibly others as well. The purpose is to make an interface that
2057 provide hooks so anyone can build a separate set of allocation and
2058 deallocation routines to be used by OpenSSL, for example memory
2059 pool implementations, or something else, which was previously hard
2060 since Malloc(), Realloc() and Free() were defined as macros having
2061 the values malloc, realloc and free, respectively (except for Win32
2062 compilations). The same is provided for memory debugging code.
2063 OpenSSL already comes with functionality to find memory leaks, but
2064 this gives people a chance to debug other memory problems.
2066 With these changes, a new set of functions and macros have appeared:
2068 CRYPTO_set_mem_debug_functions() [F]
2069 CRYPTO_get_mem_debug_functions() [F]
2070 CRYPTO_dbg_set_options() [F]
2071 CRYPTO_dbg_get_options() [F]
2072 CRYPTO_malloc_debug_init() [M]
2074 The memory debug functions are NULL by default, unless the library
2075 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
2076 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
2077 gives the standard debugging functions that come with OpenSSL) or
2078 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
2079 provided by the library user) must be used. When the standard
2080 debugging functions are used, CRYPTO_dbg_set_options can be used to
2081 request additional information:
2082 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2083 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2085 Also, things like CRYPTO_set_mem_functions will always give the
2086 expected result (the new set of functions is used for allocation
2087 and deallocation) at all times, regardless of platform and compiler
2090 To finish it up, some functions that were never use in any other
2091 way than through macros have a new API and new semantic:
2094 CRYPTO_dbg_realloc()
2097 All macros of value have retained their old syntax.
2098 [Richard Levitte and Bodo Moeller]
2100 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2101 ordering of SMIMECapabilities wasn't in "strength order" and there
2102 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2106 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2107 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2108 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2110 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2111 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2112 functionality to handle multipart/signed properly) and a utility
2113 called 'smime' to call all this stuff. This is based on code I
2114 originally wrote for Celo who have kindly allowed it to be
2115 included in OpenSSL.
2118 *) Add variants des_set_key_checked and des_set_key_unchecked of
2119 des_set_key (aka des_key_sched). Global variable des_check_key
2120 decides which of these is called by des_set_key; this way
2121 des_check_key behaves as it always did, but applications and
2122 the library itself, which was buggy for des_check_key == 1,
2123 have a cleaner way to pick the version they need.
2126 *) New function PKCS12_newpass() which changes the password of a
2130 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2131 dynamic mix. In both cases the ids can be used as an index into the
2132 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2133 functions so they accept a list of the field values and the
2134 application doesn't need to directly manipulate the X509_TRUST
2138 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2142 *) Modify the way the V3 extension code looks up extensions. This now
2143 works in a similar way to the object code: we have some "standard"
2144 extensions in a static table which is searched with OBJ_bsearch()
2145 and the application can add dynamic ones if needed. The file
2146 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2147 updated whenever a new extension is added to the core code and kept
2148 in ext_nid order. There is a simple program 'tabtest.c' which checks
2149 this. New extensions are not added too often so this file can readily
2150 be maintained manually.
2152 There are two big advantages in doing things this way. The extensions
2153 can be looked up immediately and no longer need to be "added" using
2154 X509V3_add_standard_extensions(): this function now does nothing.
2155 [Side note: I get *lots* of email saying the extension code doesn't
2156 work because people forget to call this function]
2157 Also no dynamic allocation is done unless new extensions are added:
2158 so if we don't add custom extensions there is no need to call
2159 X509V3_EXT_cleanup().
2162 *) Modify enc utility's salting as follows: make salting the default. Add a
2163 magic header, so unsalted files fail gracefully instead of just decrypting
2164 to garbage. This is because not salting is a big security hole, so people
2165 should be discouraged from doing it.
2168 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2169 digest to be passed on the command line but it only used this
2170 parameter when signing a certificate. Modified so all relevant
2171 operations are affected by the digest parameter including the
2172 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2173 DSA key was used because it didn't fix the digest.
2176 *) Initial certificate chain verify code. Currently tests the untrusted
2177 certificates for consistency with the verify purpose (which is set
2178 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2180 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2181 this is because it will reject chains with invalid extensions whereas
2182 every previous version of OpenSSL and SSLeay made no checks at all.
2184 Trust code: checks the root CA for the relevant trust settings. Trust
2185 settings have an initial value consistent with the verify purpose: e.g.
2186 if the verify purpose is for SSL client use it expects the CA to be
2187 trusted for SSL client use. However the default value can be changed to
2188 permit custom trust settings: one example of this would be to only trust
2189 certificates from a specific "secure" set of CAs.
2191 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2192 which should be used for version portability: especially since the
2193 verify structure is likely to change more often now.
2195 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2196 to set them. If not set then assume SSL clients will verify SSL servers
2199 Two new options to the verify program: -untrusted allows a set of
2200 untrusted certificates to be passed in and -purpose which sets the
2201 intended purpose of the certificate. If a purpose is set then the
2202 new chain verify code is used to check extension consistency.
2205 *) Support for the authority information access extension.
2208 *) Modify RSA and DSA PEM read routines to transparently handle
2209 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2210 public keys in a format compatible with certificate
2211 SubjectPublicKeyInfo structures. Unfortunately there were already
2212 functions called *_PublicKey_* which used various odd formats so
2213 these are retained for compatibility: however the DSA variants were
2214 never in a public release so they have been deleted. Changed dsa/rsa
2215 utilities to handle the new format: note no releases ever handled public
2216 keys so we should be OK.
2218 The primary motivation for this change is to avoid the same fiasco
2219 that dogs private keys: there are several incompatible private key
2220 formats some of which are standard and some OpenSSL specific and
2221 require various evil hacks to allow partial transparent handling and
2222 even then it doesn't work with DER formats. Given the option anything
2223 other than PKCS#8 should be dumped: but the other formats have to
2224 stay in the name of compatibility.
2226 With public keys and the benefit of hindsight one standard format
2227 is used which works with EVP_PKEY, RSA or DSA structures: though
2228 it clearly returns an error if you try to read the wrong kind of key.
2230 Added a -pubkey option to the 'x509' utility to output the public key.
2231 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2232 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2233 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2234 that do the same as the EVP_PKEY_assign_*() except they up the
2235 reference count of the added key (they don't "swallow" the
2239 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2240 CRLs would fail if the file contained no certificates or no CRLs:
2241 added a new function to read in both types and return the number
2242 read: this means that if none are read it will be an error. The
2243 DER versions of the certificate and CRL reader would always fail
2244 because it isn't possible to mix certificates and CRLs in DER format
2245 without choking one or the other routine. Changed this to just read
2246 a certificate: this is the best we can do. Also modified the code
2247 in apps/verify.c to take notice of return codes: it was previously
2248 attempting to read in certificates from NULL pointers and ignoring
2249 any errors: this is one reason why the cert and CRL reader seemed
2250 to work. It doesn't check return codes from the default certificate
2251 routines: these may well fail if the certificates aren't installed.
2254 *) Code to support otherName option in GeneralName.
2257 *) First update to verify code. Change the verify utility
2258 so it warns if it is passed a self signed certificate:
2259 for consistency with the normal behaviour. X509_verify
2260 has been modified to it will now verify a self signed
2261 certificate if *exactly* the same certificate appears
2262 in the store: it was previously impossible to trust a
2263 single self signed certificate. This means that:
2264 openssl verify ss.pem
2265 now gives a warning about a self signed certificate but
2266 openssl verify -CAfile ss.pem ss.pem
2270 *) For servers, store verify_result in SSL_SESSION data structure
2271 (and add it to external session representation).
2272 This is needed when client certificate verifications fails,
2273 but an application-provided verification callback (set by
2274 SSL_CTX_set_cert_verify_callback) allows accepting the session
2275 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2276 but returns 1): When the session is reused, we have to set
2277 ssl->verify_result to the appropriate error code to avoid
2279 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2281 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2282 case in PKCS7_dataInit() where the signed PKCS7 structure
2283 didn't contain any existing data because it was being created.
2284 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2286 *) Add a salt to the key derivation routines in enc.c. This
2287 forms the first 8 bytes of the encrypted file. Also add a
2288 -S option to allow a salt to be input on the command line.
2291 *) New function X509_cmp(). Oddly enough there wasn't a function
2292 to compare two certificates. We do this by working out the SHA1
2293 hash and comparing that. X509_cmp() will be needed by the trust
2297 *) SSL_get1_session() is like SSL_get_session(), but increments
2298 the reference count in the SSL_SESSION returned.
2299 [Geoff Thorpe <geoff@eu.c2.net>]
2301 *) Fix for 'req': it was adding a null to request attributes.
2302 Also change the X509_LOOKUP and X509_INFO code to handle
2303 certificate auxiliary information.
2306 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2310 *) Add the possibility to add extra information to the memory leak
2311 detecting output, to form tracebacks, showing from where each
2312 allocation was originated: CRYPTO_push_info("constant string") adds
2313 the string plus current file name and line number to a per-thread
2314 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2315 is like calling CYRPTO_pop_info() until the stack is empty.
2316 Also updated memory leak detection code to be multi-thread-safe.
2319 *) Add options -text and -noout to pkcs7 utility and delete the
2320 encryption options which never did anything. Update docs.
2323 *) Add options to some of the utilities to allow the pass phrase
2324 to be included on either the command line (not recommended on
2325 OSes like Unix) or read from the environment. Update the
2326 manpages and fix a few bugs.
2329 *) Add a few manpages for some of the openssl commands.
2332 *) Fix the -revoke option in ca. It was freeing up memory twice,
2333 leaking and not finding already revoked certificates.
2336 *) Extensive changes to support certificate auxiliary information.
2337 This involves the use of X509_CERT_AUX structure and X509_AUX
2338 functions. An X509_AUX function such as PEM_read_X509_AUX()
2339 can still read in a certificate file in the usual way but it
2340 will also read in any additional "auxiliary information". By
2341 doing things this way a fair degree of compatibility can be
2342 retained: existing certificates can have this information added
2343 using the new 'x509' options.
2345 Current auxiliary information includes an "alias" and some trust
2346 settings. The trust settings will ultimately be used in enhanced
2347 certificate chain verification routines: currently a certificate
2348 can only be trusted if it is self signed and then it is trusted
2352 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2353 The problem was that one of the replacement routines had not been working
2354 since SSLeay releases. For now the offending routine has been replaced
2355 with non-optimised assembler. Even so, this now gives around 95%
2356 performance improvement for 1024 bit RSA signs.
2359 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2360 handling. Most clients have the effective key size in bits equal to
2361 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2362 A few however don't do this and instead use the size of the decrypted key
2363 to determine the RC2 key length and the AlgorithmIdentifier to determine
2364 the effective key length. In this case the effective key length can still
2365 be 40 bits but the key length can be 168 bits for example. This is fixed
2366 by manually forcing an RC2 key into the EVP_PKEY structure because the
2367 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2368 the key length and effective key length are equal.
2371 *) Add a bunch of functions that should simplify the creation of
2372 X509_NAME structures. Now you should be able to do:
2373 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2374 and have it automatically work out the correct field type and fill in
2375 the structures. The more adventurous can try:
2376 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2377 and it will (hopefully) work out the correct multibyte encoding.
2380 *) Change the 'req' utility to use the new field handling and multibyte
2381 copy routines. Before the DN field creation was handled in an ad hoc
2382 way in req, ca, and x509 which was rather broken and didn't support
2383 BMPStrings or UTF8Strings. Since some software doesn't implement
2384 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2385 using the dirstring_type option. See the new comment in the default
2386 openssl.cnf for more info.
2389 *) Make crypto/rand/md_rand.c more robust:
2390 - Assure unique random numbers after fork().
2391 - Make sure that concurrent threads access the global counter and
2392 md serializably so that we never lose entropy in them
2393 or use exactly the same state in multiple threads.
2394 Access to the large state is not always serializable because
2395 the additional locking could be a performance killer, and
2396 md should be large enough anyway.
2399 *) New file apps/app_rand.c with commonly needed functionality
2400 for handling the random seed file.
2402 Use the random seed file in some applications that previously did not:
2404 dsaparam -genkey (which also ignored its '-rand' option),
2407 x509 (when signing).
2408 Except on systems with /dev/urandom, it is crucial to have a random
2409 seed file at least for key creation, DSA signing, and for DH exchanges;
2410 for RSA signatures we could do without one.
2412 gendh and gendsa (unlike genrsa) used to read only the first byte
2413 of each file listed in the '-rand' option. The function as previously
2414 found in genrsa is now in app_rand.c and is used by all programs
2415 that support '-rand'.
2418 *) In RAND_write_file, use mode 0600 for creating files;
2419 don't just chmod when it may be too late.
2422 *) Report an error from X509_STORE_load_locations
2423 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2426 *) New function ASN1_mbstring_copy() this copies a string in either
2427 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2428 into an ASN1_STRING type. A mask of permissible types is passed
2429 and it chooses the "minimal" type to use or an error if not type
2433 *) Add function equivalents to the various macros in asn1.h. The old
2434 macros are retained with an M_ prefix. Code inside the library can
2435 use the M_ macros. External code (including the openssl utility)
2436 should *NOT* in order to be "shared library friendly".
2439 *) Add various functions that can check a certificate's extensions
2440 to see if it usable for various purposes such as SSL client,
2441 server or S/MIME and CAs of these types. This is currently
2442 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2443 verification. Also added a -purpose flag to x509 utility to
2444 print out all the purposes.
2447 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2451 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2452 for, obtain and decode and extension and obtain its critical flag.
2453 This allows all the necessary extension code to be handled in a
2454 single function call.
2457 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2458 platforms. See crypto/rc4/rc4_enc.c for further details.
2461 *) New -noout option to asn1parse. This causes no output to be produced
2462 its main use is when combined with -strparse and -out to extract data
2463 from a file (which may not be in ASN.1 format).
2466 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2467 when producing the local key id.
2468 [Richard Levitte <levitte@stacken.kth.se>]
2470 *) New option -dhparam in s_server. This allows a DH parameter file to be
2471 stated explicitly. If it is not stated then it tries the first server
2472 certificate file. The previous behaviour hard coded the filename
2476 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2477 a public key to be input or output. For example:
2478 openssl rsa -in key.pem -pubout -out pubkey.pem
2479 Also added necessary DSA public key functions to handle this.
2482 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2483 in the message. This was handled by allowing
2484 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2485 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2487 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2488 to the end of the strings whereas this didn't. This would cause problems
2489 if strings read with d2i_ASN1_bytes() were later modified.
2490 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2492 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2493 data and it contains EOF it will end up returning an error. This is
2494 caused by input 46 bytes long. The cause is due to the way base64
2495 BIOs find the start of base64 encoded data. They do this by trying a
2496 trial decode on each line until they find one that works. When they
2497 do a flag is set and it starts again knowing it can pass all the
2498 data directly through the decoder. Unfortunately it doesn't reset
2499 the context it uses. This means that if EOF is reached an attempt
2500 is made to pass two EOFs through the context and this causes the
2501 resulting error. This can also cause other problems as well. As is
2502 usual with these problems it takes *ages* to find and the fix is
2503 trivial: move one line.
2504 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2506 *) Ugly workaround to get s_client and s_server working under Windows. The
2507 old code wouldn't work because it needed to select() on sockets and the
2508 tty (for keypresses and to see if data could be written). Win32 only
2509 supports select() on sockets so we select() with a 1s timeout on the
2510 sockets and then see if any characters are waiting to be read, if none
2511 are present then we retry, we also assume we can always write data to
2512 the tty. This isn't nice because the code then blocks until we've
2513 received a complete line of data and it is effectively polling the
2514 keyboard at 1s intervals: however it's quite a bit better than not
2515 working at all :-) A dedicated Windows application might handle this
2516 with an event loop for example.
2519 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2520 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2521 will be called when RSA_sign() and RSA_verify() are used. This is useful
2522 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2523 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2524 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2525 This necessitated the support of an extra signature type NID_md5_sha1
2526 for SSL signatures and modifications to the SSL library to use it instead
2527 of calling RSA_public_decrypt() and RSA_private_encrypt().
2530 *) Add new -verify -CAfile and -CApath options to the crl program, these
2531 will lookup a CRL issuers certificate and verify the signature in a
2532 similar way to the verify program. Tidy up the crl program so it
2533 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2534 less strict. It will now permit CRL extensions even if it is not
2535 a V2 CRL: this will allow it to tolerate some broken CRLs.
2538 *) Initialize all non-automatic variables each time one of the openssl
2539 sub-programs is started (this is necessary as they may be started
2540 multiple times from the "OpenSSL>" prompt).
2541 [Lennart Bang, Bodo Moeller]
2543 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2544 removing all other RSA functionality (this is what NO_RSA does). This
2545 is so (for example) those in the US can disable those operations covered
2546 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2550 *) Non-copying interface to BIO pairs.
2551 (still largely untested)
2554 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2555 ASCII string. This was handled independently in various places before.
2558 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2559 UTF8 strings a character at a time.
2562 *) Use client_version from client hello to select the protocol
2563 (s23_srvr.c) and for RSA client key exchange verification
2564 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2567 *) Add various utility functions to handle SPKACs, these were previously
2568 handled by poking round in the structure internals. Added new function
2569 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2570 print, verify and generate SPKACs. Based on an original idea from
2571 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2574 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2577 *) Allow the config file extension section to be overwritten on the
2578 command line. Based on an original idea from Massimiliano Pala
2579 <madwolf@comune.modena.it>. The new option is called -extensions
2580 and can be applied to ca, req and x509. Also -reqexts to override
2581 the request extensions in req and -crlexts to override the crl extensions
2585 *) Add new feature to the SPKAC handling in ca. Now you can include
2586 the same field multiple times by preceding it by "XXXX." for example:
2589 this is the same syntax as used in the req config file.
2592 *) Allow certificate extensions to be added to certificate requests. These
2593 are specified in a 'req_extensions' option of the req section of the
2594 config file. They can be printed out with the -text option to req but
2595 are otherwise ignored at present.
2598 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2599 data read consists of only the final block it would not decrypted because
2600 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2601 A misplaced 'break' also meant the decrypted final block might not be
2602 copied until the next read.
2605 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2606 a few extra parameters to the DH structure: these will be useful if
2607 for example we want the value of 'q' or implement X9.42 DH.
2610 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2611 provides hooks that allow the default DSA functions or functions on a
2612 "per key" basis to be replaced. This allows hardware acceleration and
2613 hardware key storage to be handled without major modification to the
2614 library. Also added low level modexp hooks and CRYPTO_EX structure and
2615 associated functions.
2618 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2619 as "read only": it can't be written to and the buffer it points to will
2620 not be freed. Reading from a read only BIO is much more efficient than
2621 a normal memory BIO. This was added because there are several times when
2622 an area of memory needs to be read from a BIO. The previous method was
2623 to create a memory BIO and write the data to it, this results in two
2624 copies of the data and an O(n^2) reading algorithm. There is a new
2625 function BIO_new_mem_buf() which creates a read only memory BIO from
2626 an area of memory. Also modified the PKCS#7 routines to use read only
2630 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2631 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2632 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2633 but a retry condition occured while trying to read the rest.
2636 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2637 NID_pkcs7_encrypted by default: this was wrong since this should almost
2638 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2639 the encrypted data type: this is a more sensible place to put it and it
2640 allows the PKCS#12 code to be tidied up that duplicated this
2644 *) Changed obj_dat.pl script so it takes its input and output files on
2645 the command line. This should avoid shell escape redirection problems
2649 *) Initial support for certificate extension requests, these are included
2650 in things like Xenroll certificate requests. Included functions to allow
2651 extensions to be obtained and added.
2654 *) -crlf option to s_client and s_server for sending newlines as
2655 CRLF (as required by many protocols).
2658 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2660 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2661 [Ralf S. Engelschall]
2663 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2664 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2666 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2670 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2671 DH parameters/keys (q is lost during that conversion, but the resulting
2672 DH parameters contain its length).
2674 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2675 much faster than DH_generate_parameters (which creates parameters
2676 where p = 2*q + 1), and also the smaller q makes DH computations
2677 much more efficient (160-bit exponentiation instead of 1024-bit
2678 exponentiation); so this provides a convenient way to support DHE
2679 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2680 utter importance to use
2681 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2683 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2684 when such DH parameters are used, because otherwise small subgroup
2685 attacks may become possible!
2688 *) Avoid memory leak in i2d_DHparams.
2691 *) Allow the -k option to be used more than once in the enc program:
2692 this allows the same encrypted message to be read by multiple recipients.
2695 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2696 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2697 it will always use the numerical form of the OID, even if it has a short
2701 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2702 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2703 otherwise bn_mod_exp was called. In the case of hardware keys for example
2704 no private key components need be present and it might store extra data
2705 in the RSA structure, which cannot be accessed from bn_mod_exp.
2706 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2707 private key operations.
2710 *) Added support for SPARC Linux.
2713 *) pem_password_cb function type incompatibly changed from
2714 typedef int pem_password_cb(char *buf, int size, int rwflag);
2716 ....(char *buf, int size, int rwflag, void *userdata);
2717 so that applications can pass data to their callbacks:
2718 The PEM[_ASN1]_{read,write}... functions and macros now take an
2719 additional void * argument, which is just handed through whenever
2720 the password callback is called.
2721 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2723 New function SSL_CTX_set_default_passwd_cb_userdata.
2725 Compatibility note: As many C implementations push function arguments
2726 onto the stack in reverse order, the new library version is likely to
2727 interoperate with programs that have been compiled with the old
2728 pem_password_cb definition (PEM_whatever takes some data that
2729 happens to be on the stack as its last argument, and the callback
2730 just ignores this garbage); but there is no guarantee whatsoever that
2733 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2734 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2735 problems not only on Windows, but also on some Unix platforms.
2736 To avoid problematic command lines, these definitions are now in an
2737 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2738 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2741 *) MIPS III/IV assembler module is reimplemented.
2744 *) More DES library cleanups: remove references to srand/rand and
2745 delete an unused file.
2748 *) Add support for the the free Netwide assembler (NASM) under Win32,
2749 since not many people have MASM (ml) and it can be hard to obtain.
2750 This is currently experimental but it seems to work OK and pass all
2751 the tests. Check out INSTALL.W32 for info.
2754 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2755 without temporary keys kept an extra copy of the server key,
2756 and connections with temporary keys did not free everything in case
2760 *) New function RSA_check_key and new openssl rsa option -check
2761 for verifying the consistency of RSA keys.
2762 [Ulf Moeller, Bodo Moeller]
2764 *) Various changes to make Win32 compile work:
2765 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2766 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2767 comparison" warnings.
2768 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2771 *) Add a debugging option to PKCS#5 v2 key generation function: when
2772 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2773 derived keys are printed to stderr.
2776 *) Copy the flags in ASN1_STRING_dup().
2777 [Roman E. Pavlov <pre@mo.msk.ru>]
2779 *) The x509 application mishandled signing requests containing DSA
2780 keys when the signing key was also DSA and the parameters didn't match.
2782 It was supposed to omit the parameters when they matched the signing key:
2783 the verifying software was then supposed to automatically use the CA's
2784 parameters if they were absent from the end user certificate.
2786 Omitting parameters is no longer recommended. The test was also
2787 the wrong way round! This was probably due to unusual behaviour in
2788 EVP_cmp_parameters() which returns 1 if the parameters match.
2789 This meant that parameters were omitted when they *didn't* match and
2790 the certificate was useless. Certificates signed with 'ca' didn't have
2792 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2794 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2795 The interface is as follows:
2796 Applications can use
2797 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2798 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2799 "off" is now the default.
2800 The library internally uses
2801 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2802 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2803 to disable memory-checking temporarily.
2805 Some inconsistent states that previously were possible (and were
2806 even the default) are now avoided.
2808 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2809 with each memory chunk allocated; this is occasionally more helpful
2810 than just having a counter.
2812 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2814 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2818 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2819 which largely parallels "options", but is for changing API behaviour,
2820 whereas "options" are about protocol behaviour.
2821 Initial "mode" flags are:
2823 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2824 a single record has been written.
2825 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2826 retries use the same buffer location.
2827 (But all of the contents must be
2831 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
2834 *) Fix problems with no-hmac etc.
2835 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2837 *) New functions RSA_get_default_method(), RSA_set_method() and
2838 RSA_get_method(). These allows replacement of RSA_METHODs without having
2839 to mess around with the internals of an RSA structure.
2842 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2843 Also really enable memory leak checks in openssl.c and in some
2845 [Chad C. Mulligan, Bodo Moeller]
2847 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2848 up the length of negative integers. This has now been simplified to just
2849 store the length when it is first determined and use it later, rather
2850 than trying to keep track of where data is copied and updating it to
2852 [Steve Henson, reported by Brien Wheeler
2853 <bwheeler@authentica-security.com>]
2855 *) Add a new function PKCS7_signatureVerify. This allows the verification
2856 of a PKCS#7 signature but with the signing certificate passed to the
2857 function itself. This contrasts with PKCS7_dataVerify which assumes the
2858 certificate is present in the PKCS#7 structure. This isn't always the
2859 case: certificates can be omitted from a PKCS#7 structure and be
2860 distributed by "out of band" means (such as a certificate database).
2863 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2864 function prototypes in pem.h, also change util/mkdef.pl to add the
2865 necessary function names.
2868 *) mk1mf.pl (used by Windows builds) did not properly read the
2869 options set by Configure in the top level Makefile, and Configure
2870 was not even able to write more than one option correctly.
2871 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2874 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2875 file to be loaded from a BIO or FILE pointer. The BIO version will
2876 for example allow memory BIOs to contain config info.
2879 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2880 Whoever hopes to achieve shared-library compatibility across versions
2881 must use this, not the compile-time macro.
2882 (Exercise 0.9.4: Which is the minimum library version required by
2884 Note: All this applies only to multi-threaded programs, others don't
2888 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2889 through a BIO pair triggered the default case, i.e.
2890 SSLerr(...,SSL_R_UNKNOWN_STATE).
2893 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2894 can use the SSL library even if none of the specific BIOs is
2898 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2899 for the encoded length.
2900 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2902 *) Add initial documentation of the X509V3 functions.
2905 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2906 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2907 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2908 secure PKCS#8 private key format with a high iteration count.
2911 *) Fix determination of Perl interpreter: A perl or perl5
2912 _directory_ in $PATH was also accepted as the interpreter.
2913 [Ralf S. Engelschall]
2915 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2916 wrong with it but it was very old and did things like calling
2917 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2921 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2922 to use the new extension code.
2925 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2926 with macros. This should make it easier to change their form, add extra
2927 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2931 *) Add to configuration table a new entry that can specify an alternative
2932 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2933 according to Mark Crispin <MRC@Panda.COM>.
2937 *) DES CBC did not update the IV. Weird.
2940 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2941 Changing the behaviour of the former might break existing programs --
2942 where IV updating is needed, des_ncbc_encrypt can be used.
2945 *) When bntest is run from "make test" it drives bc to check its
2946 calculations, as well as internally checking them. If an internal check
2947 fails, it needs to cause bc to give a non-zero result or make test carries
2948 on without noticing the failure. Fixed.
2951 *) DES library cleanups.
2954 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2955 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2956 ciphers. NOTE: although the key derivation function has been verified
2957 against some published test vectors it has not been extensively tested
2958 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2962 *) Instead of "mkdir -p", which is not fully portable, use new
2963 Perl script "util/mkdir-p.pl".
2966 *) Rewrite the way password based encryption (PBE) is handled. It used to
2967 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2968 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2969 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2970 the 'parameter' field of the AlgorithmIdentifier is passed to the
2971 underlying key generation function so it must do its own ASN1 parsing.
2972 This has also changed the EVP_PBE_CipherInit() function which now has a
2973 'parameter' argument instead of literal salt and iteration count values
2974 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2977 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2978 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2979 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2980 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2981 value was just used as a "magic string" and not used directly its
2982 value doesn't matter.
2985 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2989 *) "linux-sparc64" configuration (ultrapenguin).
2990 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2991 "linux-sparc" configuration.
2992 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2994 *) config now generates no-xxx options for missing ciphers.
2997 *) Support the EBCDIC character set (work in progress).
2998 File ebcdic.c not yet included because it has a different license.
2999 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3001 *) Support BS2000/OSD-POSIX.
3002 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3004 *) Make callbacks for key generation use void * instead of char *.
3007 *) Make S/MIME samples compile (not yet tested).
3010 *) Additional typesafe stacks.
3013 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
3017 Changes between 0.9.3 and 0.9.3a [29 May 1999]
3019 *) New configuration variant "sco5-gcc".
3021 *) Updated some demos.
3022 [Sean O Riordain, Wade Scholine]
3024 *) Add missing BIO_free at exit of pkcs12 application.
3027 *) Fix memory leak in conf.c.
3030 *) Updates for Win32 to assembler version of MD5.
3033 *) Set #! path to perl in apps/der_chop to where we found it
3034 instead of using a fixed path.
3037 *) SHA library changes for irix64-mips4-cc.
3040 *) Improvements for VMS support.
3044 Changes between 0.9.2b and 0.9.3 [24 May 1999]
3046 *) Bignum library bug fix. IRIX 6 passes "make test" now!
3047 This also avoids the problems with SC4.2 and unpatched SC5.
3048 [Andy Polyakov <appro@fy.chalmers.se>]
3050 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
3051 These are required because of the typesafe stack would otherwise break
3052 existing code. If old code used a structure member which used to be STACK
3053 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
3054 sk_num or sk_value it would produce an error because the num, data members
3055 are not present in STACK_OF. Now it just produces a warning. sk_set
3056 replaces the old method of assigning a value to sk_value
3057 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
3058 that does this will no longer work (and should use sk_set instead) but
3059 this could be regarded as a "questionable" behaviour anyway.
3062 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
3063 correctly handle encrypted S/MIME data.
3066 *) Change type of various DES function arguments from des_cblock
3067 (which means, in function argument declarations, pointer to char)
3068 to des_cblock * (meaning pointer to array with 8 char elements),
3069 which allows the compiler to do more typechecking; it was like
3070 that back in SSLeay, but with lots of ugly casts.
3072 Introduce new type const_des_cblock.
3075 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
3076 problems: find RecipientInfo structure that matches recipient certificate
3077 and initialise the ASN1 structures properly based on passed cipher.
3080 *) Belatedly make the BN tests actually check the results.
3083 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3084 to and from BNs: it was completely broken. New compilation option
3085 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3086 key elements as negative integers.
3089 *) Reorganize and speed up MD5.
3090 [Andy Polyakov <appro@fy.chalmers.se>]
3093 [Richard Levitte <richard@levitte.org>]
3095 *) New option -out to asn1parse to allow the parsed structure to be
3096 output to a file. This is most useful when combined with the -strparse
3097 option to examine the output of things like OCTET STRINGS.
3100 *) Make SSL library a little more fool-proof by not requiring any longer
3101 that SSL_set_{accept,connect}_state be called before
3102 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3103 in many applications because usually everything *appeared* to work as
3104 intended anyway -- now it really works as intended).
3107 *) Move openssl.cnf out of lib/.
3110 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3111 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3112 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3113 [Ralf S. Engelschall]
3115 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3116 handle PKCS#7 enveloped data properly.
3117 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3119 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3120 copying pointers. The cert_st handling is changed by this in
3121 various ways (and thus what used to be known as ctx->default_cert
3122 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3123 any longer when s->cert does not give us what we need).
3124 ssl_cert_instantiate becomes obsolete by this change.
3125 As soon as we've got the new code right (possibly it already is?),
3126 we have solved a couple of bugs of the earlier code where s->cert
3127 was used as if it could not have been shared with other SSL structures.
3129 Note that using the SSL API in certain dirty ways now will result
3130 in different behaviour than observed with earlier library versions:
3131 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3132 does not influence s as it used to.
3134 In order to clean up things more thoroughly, inside SSL_SESSION
3135 we don't use CERT any longer, but a new structure SESS_CERT
3136 that holds per-session data (if available); currently, this is
3137 the peer's certificate chain and, for clients, the server's certificate
3138 and temporary key. CERT holds only those values that can have
3139 meaningful defaults in an SSL_CTX.
3142 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3143 from the internal representation. Various PKCS#7 fixes: remove some
3144 evil casts and set the enc_dig_alg field properly based on the signing
3148 *) Allow PKCS#12 password to be set from the command line or the
3149 environment. Let 'ca' get its config file name from the environment
3150 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3154 *) Allow certificate policies extension to use an IA5STRING for the
3155 organization field. This is contrary to the PKIX definition but
3156 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3160 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3161 without disallowing inline assembler and the like for non-pedantic builds.
3164 *) Support Borland C++ builder.
3165 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3170 *) SHA-1 cleanups and performance enhancements.
3171 [Andy Polyakov <appro@fy.chalmers.se>]
3173 *) Sparc v8plus assembler for the bignum library.
3174 [Andy Polyakov <appro@fy.chalmers.se>]
3176 *) Accept any -xxx and +xxx compiler options in Configure.
3179 *) Update HPUX configuration.
3182 *) Add missing sk_<type>_unshift() function to safestack.h
3183 [Ralf S. Engelschall]
3185 *) New function SSL_CTX_use_certificate_chain_file that sets the
3186 "extra_cert"s in addition to the certificate. (This makes sense
3187 only for "PEM" format files, as chains as a whole are not
3191 *) Support verify_depth from the SSL API.
3192 x509_vfy.c had what can be considered an off-by-one-error:
3193 Its depth (which was not part of the external interface)
3194 was actually counting the number of certificates in a chain;
3195 now it really counts the depth.
3198 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3199 instead of X509err, which often resulted in confusing error
3200 messages since the error codes are not globally unique
3201 (e.g. an alleged error in ssl3_accept when a certificate
3202 didn't match the private key).
3204 *) New function SSL_CTX_set_session_id_context that allows to set a default
3205 value (so that you don't need SSL_set_session_id_context for each
3206 connection using the SSL_CTX).
3209 *) OAEP decoding bug fix.
3212 *) Support INSTALL_PREFIX for package builders, as proposed by
3216 *) New Configure options "threads" and "no-threads". For systems
3217 where the proper compiler options are known (currently Solaris
3218 and Linux), "threads" is the default.
3221 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3224 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3225 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3226 such as /usr/local/bin.
3229 *) "make linux-shared" to build shared libraries.
3230 [Niels Poppe <niels@netbox.org>]
3232 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3235 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3236 extension adding in x509 utility.
3239 *) Remove NOPROTO sections and error code comments.
3242 *) Partial rewrite of the DEF file generator to now parse the ANSI
3246 *) New Configure options --prefix=DIR and --openssldir=DIR.
3249 *) Complete rewrite of the error code script(s). It is all now handled
3250 by one script at the top level which handles error code gathering,
3251 header rewriting and C source file generation. It should be much better
3252 than the old method: it now uses a modified version of Ulf's parser to
3253 read the ANSI prototypes in all header files (thus the old K&R definitions
3254 aren't needed for error creation any more) and do a better job of
3255 translating function codes into names. The old 'ASN1 error code imbedded
3256 in a comment' is no longer necessary and it doesn't use .err files which
3257 have now been deleted. Also the error code call doesn't have to appear all
3258 on one line (which resulted in some large lines...).
3261 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3264 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3265 0 (which usually indicates a closed connection), but continue reading.
3268 *) Fix some race conditions.
3271 *) Add support for CRL distribution points extension. Add Certificate
3272 Policies and CRL distribution points documentation.
3275 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3278 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3279 8 of keying material. Merlin has also confirmed interop with this fix
3280 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3281 [Merlin Hughes <merlin@baltimore.ie>]
3283 *) Fix lots of warnings.
3284 [Richard Levitte <levitte@stacken.kth.se>]
3286 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3287 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3288 [Richard Levitte <levitte@stacken.kth.se>]
3290 *) Fix problems with sizeof(long) == 8.
3291 [Andy Polyakov <appro@fy.chalmers.se>]
3293 *) Change functions to ANSI C.
3296 *) Fix typos in error codes.
3297 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3299 *) Remove defunct assembler files from Configure.
3302 *) SPARC v8 assembler BIGNUM implementation.
3303 [Andy Polyakov <appro@fy.chalmers.se>]
3305 *) Support for Certificate Policies extension: both print and set.
3306 Various additions to support the r2i method this uses.
3309 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3310 return a const string when you are expecting an allocated buffer.
3313 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3314 types DirectoryString and DisplayText.
3317 *) Add code to allow r2i extensions to access the configuration database,
3318 add an LHASH database driver and add several ctx helper functions.
3321 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3322 fail when they extended the size of a BIGNUM.
3325 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3326 support typesafe stack.
3329 *) Fix typo in SSL_[gs]et_options().
3330 [Nils Frostberg <nils@medcom.se>]
3332 *) Delete various functions and files that belonged to the (now obsolete)
3333 old X509V3 handling code.
3336 *) New Configure option "rsaref".
3339 *) Don't auto-generate pem.h.
3342 *) Introduce type-safe ASN.1 SETs.
3345 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3346 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3348 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3349 that links with OpenSSL (well at least cause lots of warnings), but fear
3350 not: the conversion is trivial, and it eliminates loads of evil casts. A
3351 few STACKed things have been converted already. Feel free to convert more.
3352 In the fullness of time, I'll do away with the STACK type altogether.
3355 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3356 specified in <certfile> by updating the entry in the index.txt file.
3357 This way one no longer has to edit the index.txt file manually for
3358 revoking a certificate. The -revoke option does the gory details now.
3359 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3361 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3362 `-text' option at all and this way the `-noout -text' combination was
3363 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3364 [Ralf S. Engelschall]
3366 *) Make sure a corresponding plain text error message exists for the
3367 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3368 verify callback function determined that a certificate was revoked.
3369 [Ralf S. Engelschall]
3371 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3372 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3373 all available cipers including rc5, which was forgotten until now.
3374 In order to let the testing shell script know which algorithms
3375 are available, a new (up to now undocumented) command
3376 "openssl list-cipher-commands" is used.
3379 *) Bugfix: s_client occasionally would sleep in select() when
3380 it should have checked SSL_pending() first.
3383 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3384 the raw DSA values prior to ASN.1 encoding.
3387 *) Tweaks to Configure
3388 [Niels Poppe <niels@netbox.org>]
3390 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3394 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3397 *) New config option to avoid instructions that are illegal on the 80386.
3398 The default code is faster, but requires at least a 486.
3401 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3402 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3403 same as SSL2_VERSION anyway.
3406 *) New "-showcerts" option for s_client.
3409 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3410 application. Various cleanups and fixes.
3413 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3414 modify error routines to work internally. Add error codes and PBE init
3415 to library startup routines.
3418 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3419 packing functions to asn1 and evp. Changed function names and error
3420 codes along the way.
3423 *) PKCS12 integration: and so it begins... First of several patches to
3424 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3425 objects to objects.h
3428 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3429 and display support for Thawte strong extranet extension.
3432 *) Add LinuxPPC support.
3433 [Jeff Dubrule <igor@pobox.org>]
3435 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3436 bn_div_words in alpha.s.
3437 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3439 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3440 OAEP isn't supported when OpenSSL is built with RSAref.
3441 [Ulf Moeller <ulf@fitug.de>]
3443 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3444 so they no longer are missing under -DNOPROTO.
3445 [Soren S. Jorvang <soren@t.dk>]
3448 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3450 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3451 doesn't work when the session is reused. Coming soon!
3454 *) Fix a security hole, that allows sessions to be reused in the wrong
3455 context thus bypassing client cert protection! All software that uses
3456 client certs and session caches in multiple contexts NEEDS PATCHING to
3457 allow session reuse! A fuller solution is in the works.
3458 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3460 *) Some more source tree cleanups (removed obsolete files
3461 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3462 permission on "config" script to be executable) and a fix for the INSTALL
3464 [Ulf Moeller <ulf@fitug.de>]
3466 *) Remove some legacy and erroneous uses of malloc, free instead of
3468 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3470 *) Make rsa_oaep_test return non-zero on error.
3471 [Ulf Moeller <ulf@fitug.de>]
3473 *) Add support for native Solaris shared libraries. Configure
3474 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3475 if someone would make that last step automatic.
3476 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3478 *) ctx_size was not built with the right compiler during "make links". Fixed.
3481 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3482 except NULL ciphers". This means the default cipher list will no longer
3483 enable NULL ciphers. They need to be specifically enabled e.g. with
3484 the string "DEFAULT:eNULL".
3487 *) Fix to RSA private encryption routines: if p < q then it would
3488 occasionally produce an invalid result. This will only happen with
3489 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3492 *) Be less restrictive and allow also `perl util/perlpath.pl
3493 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3494 because this way one can also use an interpreter named `perl5' (which is
3495 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3496 installed as `perl').
3497 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3499 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3500 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3502 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3503 advapi32.lib to Win32 build and change the pem test comparision
3504 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3505 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3506 and crypto/des/ede_cbcm_enc.c.
3509 *) DES quad checksum was broken on big-endian architectures. Fixed.
3512 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3513 Win32 test batch file so it (might) work again. The Win32 test batch file
3514 is horrible: I feel ill....
3517 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3518 in e_os.h. Audit of header files to check ANSI and non ANSI
3519 sections: 10 functions were absent from non ANSI section and not exported
3520 from Windows DLLs. Fixed up libeay.num for new functions.
3523 *) Make `openssl version' output lines consistent.
3524 [Ralf S. Engelschall]
3526 *) Fix Win32 symbol export lists for BIO functions: Added
3527 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3528 to ms/libeay{16,32}.def.
3529 [Ralf S. Engelschall]
3531 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3532 fine under Unix and passes some trivial tests I've now added. But the
3533 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3534 added to make sure no one expects that this stuff really works in the
3535 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3536 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3538 [Ralf S. Engelschall]
3540 *) Fix the generation of two part addresses in perl.
3541 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3543 *) Add config entry for Linux on MIPS.
3544 [John Tobey <jtobey@channel1.com>]
3546 *) Make links whenever Configure is run, unless we are on Windoze.
3549 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3550 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3554 *) Add a useful kludge to allow package maintainers to specify compiler and
3555 other platforms details on the command line without having to patch the
3556 Configure script everytime: One now can use ``perl Configure
3557 <id>:<details>'', i.e. platform ids are allowed to have details appended
3558 to them (seperated by colons). This is treated as there would be a static
3559 pre-configured entry in Configure's %table under key <id> with value
3560 <details> and ``perl Configure <id>'' is called. So, when you want to
3561 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3562 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3563 now, which overrides the FreeBSD-elf entry on-the-fly.
3564 [Ralf S. Engelschall]
3566 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3569 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3570 on the `perl Configure ...' command line. This way one can compile
3571 OpenSSL libraries with Position Independent Code (PIC) which is needed
3572 for linking it into DSOs.
3573 [Ralf S. Engelschall]
3575 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3579 *) Cleaned up the LICENSE document: The official contact for any license
3580 questions now is the OpenSSL core team under openssl-core@openssl.org.
3581 And add a paragraph about the dual-license situation to make sure people
3582 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3583 to the OpenSSL toolkit.
3584 [Ralf S. Engelschall]
3586 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3587 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3588 Additonally cleaned up the `make links' target: Remove unnecessary
3589 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3590 to speed processing and no longer clutter the display with confusing
3591 stuff. Instead only the actually done links are displayed.
3592 [Ralf S. Engelschall]
3594 *) Permit null encryption ciphersuites, used for authentication only. It used
3595 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3596 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3600 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3601 signed attributes when verifying signatures (this would break them),
3602 the detached data encoding was wrong and public keys obtained using
3603 X509_get_pubkey() weren't freed.
3606 *) Add text documentation for the BUFFER functions. Also added a work around
3607 to a Win95 console bug. This was triggered by the password read stuff: the
3608 last character typed gets carried over to the next fread(). If you were
3609 generating a new cert request using 'req' for example then the last
3610 character of the passphrase would be CR which would then enter the first
3614 *) Added the new `Includes OpenSSL Cryptography Software' button as
3615 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3616 button and can be used by applications based on OpenSSL to show the
3617 relationship to the OpenSSL project.
3618 [Ralf S. Engelschall]
3620 *) Remove confusing variables in function signatures in files
3621 ssl/ssl_lib.c and ssl/ssl.h.
3622 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3624 *) Don't install bss_file.c under PREFIX/include/
3625 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3627 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3628 functions that return function pointers and has support for NT specific
3629 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3630 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3631 unsigned to signed types: this was killing the Win32 compile.
3634 *) Add new certificate file to stack functions,
3635 SSL_add_dir_cert_subjects_to_stack() and
3636 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3637 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3638 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3639 This means that Apache-SSL and similar packages don't have to mess around
3640 to add as many CAs as they want to the preferred list.
3643 *) Experiment with doxygen documentation. Currently only partially applied to
3645 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3646 openssl.doxy as the configuration file.
3649 *) Get rid of remaining C++-style comments which strict C compilers hate.
3650 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3652 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3653 compiled in by default: it has problems with large keys.
3656 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3657 DH private keys and/or callback functions which directly correspond to
3658 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3659 is needed for applications which have to configure certificates on a
3660 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3662 For the RSA certificate situation is makes no difference, but
3663 for the DSA certificate situation this fixes the "no shared cipher"
3664 problem where the OpenSSL cipher selection procedure failed because the
3665 temporary keys were not overtaken from the context and the API provided
3666 no way to reconfigure them.
3667 The new functions now let applications reconfigure the stuff and they
3668 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3669 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3670 non-public-API function ssl_cert_instantiate() is used as a helper
3671 function and also to reduce code redundancy inside ssl_rsa.c.
3672 [Ralf S. Engelschall]
3674 *) Move s_server -dcert and -dkey options out of the undocumented feature
3675 area because they are useful for the DSA situation and should be
3676 recognized by the users.
3677 [Ralf S. Engelschall]
3679 *) Fix the cipher decision scheme for export ciphers: the export bits are
3680 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3681 SSL_EXP_MASK. So, the original variable has to be used instead of the
3682 already masked variable.
3683 [Richard Levitte <levitte@stacken.kth.se>]
3685 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3686 [Richard Levitte <levitte@stacken.kth.se>]
3688 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3689 from `int' to `unsigned int' because it's a length and initialized by
3690 EVP_DigestFinal() which expects an `unsigned int *'.
3691 [Richard Levitte <levitte@stacken.kth.se>]
3693 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3694 script. Instead use the usual Shell->Perl transition trick.
3695 [Ralf S. Engelschall]
3697 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3698 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3699 -noout -modulus' as it's already the case for `openssl rsa -noout
3700 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3701 currently the public key is printed (a decision which was already done by
3702 `openssl dsa -modulus' in the past) which serves a similar purpose.
3703 Additionally the NO_RSA no longer completely removes the whole -modulus
3704 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3706 [Ralf S. Engelschall]
3708 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3709 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3710 [Arne Ansper <arne@ats.cyber.ee>]
3712 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3713 to be added. Now both 'req' and 'ca' can use new objects defined in the
3717 *) Add cool BIO that does syslog (or event log on NT).
3718 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3720 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3721 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3722 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3723 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3726 *) Add preliminary config info for new extension code.
3729 *) Make RSA_NO_PADDING really use no padding.
3730 [Ulf Moeller <ulf@fitug.de>]
3732 *) Generate errors when private/public key check is done.
3735 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3736 for some CRL extensions and new objects added.
3739 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3740 key usage extension and fuller support for authority key id.
3743 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3744 padding method for RSA, which is recommended for new applications in PKCS
3745 #1 v2.0 (RFC 2437, October 1998).
3746 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3747 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3748 against Bleichbacher's attack on RSA.
3749 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3752 *) Updates to the new SSL compression code
3753 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3755 *) Fix so that the version number in the master secret, when passed
3756 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3757 (because the server will not accept higher), that the version number
3758 is 0x03,0x01, not 0x03,0x00
3759 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3761 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3762 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3763 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3766 *) Support for RAW extensions where an arbitrary extension can be
3767 created by including its DER encoding. See apps/openssl.cnf for
3771 *) Make sure latest Perl versions don't interpret some generated C array
3772 code as Perl array code in the crypto/err/err_genc.pl script.
3773 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3775 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3776 not many people have the assembler. Various Win32 compilation fixes and
3777 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3781 *) Modify configure script 'Configure' to automatically create crypto/date.h
3782 file under Win32 and also build pem.h from pem.org. New script
3783 util/mkfiles.pl to create the MINFO file on environments that can't do a
3784 'make files': perl util/mkfiles.pl >MINFO should work.
3787 *) Major rework of DES function declarations, in the pursuit of correctness
3788 and purity. As a result, many evil casts evaporated, and some weirdness,
3789 too. You may find this causes warnings in your code. Zapping your evil
3790 casts will probably fix them. Mostly.
3793 *) Fix for a typo in asn1.h. Bug fix to object creation script
3794 obj_dat.pl. It considered a zero in an object definition to mean
3795 "end of object": none of the objects in objects.h have any zeros
3796 so it wasn't spotted.
3797 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3799 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3800 Masking (CBCM). In the absence of test vectors, the best I have been able
3801 to do is check that the decrypt undoes the encrypt, so far. Send me test
3802 vectors if you have them.
3805 *) Correct calculation of key length for export ciphers (too much space was
3806 allocated for null ciphers). This has not been tested!
3809 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3810 message is now correct (it understands "crypto" and "ssl" on its
3811 command line). There is also now an "update" option. This will update
3812 the util/ssleay.num and util/libeay.num files with any new functions.
3814 perl util/mkdef.pl crypto ssl update
3815 it will update them.
3818 *) Overhauled the Perl interface (perl/*):
3819 - ported BN stuff to OpenSSL's different BN library
3820 - made the perl/ source tree CVS-aware
3821 - renamed the package from SSLeay to OpenSSL (the files still contain
3822 their history because I've copied them in the repository)
3823 - removed obsolete files (the test scripts will be replaced
3824 by better Test::Harness variants in the future)
3825 [Ralf S. Engelschall]
3827 *) First cut for a very conservative source tree cleanup:
3828 1. merge various obsolete readme texts into doc/ssleay.txt
3829 where we collect the old documents and readme texts.
3830 2. remove the first part of files where I'm already sure that we no
3831 longer need them because of three reasons: either they are just temporary
3832 files which were left by Eric or they are preserved original files where
3833 I've verified that the diff is also available in the CVS via "cvs diff
3834 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3835 the crypto/md/ stuff).
3836 [Ralf S. Engelschall]
3838 *) More extension code. Incomplete support for subject and issuer alt
3839 name, issuer and authority key id. Change the i2v function parameters
3840 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3841 what that's for :-) Fix to ASN1 macro which messed up
3842 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3845 *) Preliminary support for ENUMERATED type. This is largely copied from the
3849 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3850 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3852 *) Make sure `make rehash' target really finds the `openssl' program.
3853 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3855 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3856 like to hear about it if this slows down other processors.
3859 *) Add CygWin32 platform information to Configure script.
3860 [Alan Batie <batie@aahz.jf.intel.com>]
3862 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3863 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3865 *) New program nseq to manipulate netscape certificate sequences
3868 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3872 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3873 but the BN code had some problems that would cause failures when
3874 doing certificate verification and some other functions.
3875 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3877 *) Add ASN1 and PEM code to support netscape certificate sequences.
3880 *) Add ASN1 and PEM code to support netscape certificate sequences.
3883 *) Add several PKIX and private extended key usage OIDs.
3886 *) Modify the 'ca' program to handle the new extension code. Modify
3887 openssl.cnf for new extension format, add comments.
3890 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3891 and add a sample to openssl.cnf so req -x509 now adds appropriate
3895 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3896 error code, add initial support to X509_print() and x509 application.
3899 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3900 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3901 stuff is currently isolated and isn't even compiled yet.
3904 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3905 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3906 Removed the versions check from X509 routines when loading extensions:
3907 this allows certain broken certificates that don't set the version
3908 properly to be processed.
3911 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3912 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3913 can still be regenerated with "make depend".
3916 *) Spelling mistake in C version of CAST-128.
3917 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3919 *) Changes to the error generation code. The perl script err-code.pl
3920 now reads in the old error codes and retains the old numbers, only
3921 adding new ones if necessary. It also only changes the .err files if new
3922 codes are added. The makefiles have been modified to only insert errors
3923 when needed (to avoid needlessly modifying header files). This is done
3924 by only inserting errors if the .err file is newer than the auto generated
3925 C file. To rebuild all the error codes from scratch (the old behaviour)
3926 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3927 or delete all the .err files.
3930 *) CAST-128 was incorrectly implemented for short keys. The C version has
3931 been fixed, but is untested. The assembler versions are also fixed, but
3932 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3933 to regenerate it if needed.
3934 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3935 Hagino <itojun@kame.net>]
3937 *) File was opened incorrectly in randfile.c.
3938 [Ulf Möller <ulf@fitug.de>]
3940 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3941 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3942 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3943 al: it's just almost always a UTCTime. Note this patch adds new error
3944 codes so do a "make errors" if there are problems.
3947 *) Correct Linux 1 recognition in config.
3948 [Ulf Möller <ulf@fitug.de>]
3950 *) Remove pointless MD5 hash when using DSA keys in ca.
3951 [Anonymous <nobody@replay.com>]
3953 *) Generate an error if given an empty string as a cert directory. Also
3954 generate an error if handed NULL (previously returned 0 to indicate an
3955 error, but didn't set one).
3956 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3958 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3961 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3962 parameters. This was causing a warning which killed off the Win32 compile.
3965 *) Remove C++ style comments from crypto/bn/bn_local.h.
3966 [Neil Costigan <neil.costigan@celocom.com>]
3968 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3969 based on a text string, looking up short and long names and finally
3970 "dot" format. The "dot" format stuff didn't work. Added new function
3971 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3972 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3973 OID is not part of the table.
3976 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3977 X509_LOOKUP_by_alias().
3980 *) Sort openssl functions by name.
3983 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3984 encryption from sample DSA keys (in case anyone is interested the password
3988 *) Make _all_ *_free functions accept a NULL pointer.
3989 [Frans Heymans <fheymans@isaserver.be>]
3991 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3993 [Anonymous <nobody@replay.com>]
3995 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3996 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3998 *) Don't blow it for numeric -newkey arguments to apps/req.
3999 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
4001 *) Temp key "for export" tests were wrong in s3_srvr.c.
4002 [Anonymous <nobody@replay.com>]
4004 *) Add prototype for temp key callback functions
4005 SSL_CTX_set_tmp_{rsa,dh}_callback().
4008 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
4009 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
4012 *) X509_name_add_entry() freed the wrong thing after an error.
4013 [Arne Ansper <arne@ats.cyber.ee>]
4015 *) rsa_eay.c would attempt to free a NULL context.
4016 [Arne Ansper <arne@ats.cyber.ee>]
4018 *) BIO_s_socket() had a broken should_retry() on Windoze.
4019 [Arne Ansper <arne@ats.cyber.ee>]
4021 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
4022 [Arne Ansper <arne@ats.cyber.ee>]
4024 *) Make sure the already existing X509_STORE->depth variable is initialized
4025 in X509_STORE_new(), but document the fact that this variable is still
4026 unused in the certificate verification process.
4027 [Ralf S. Engelschall]
4029 *) Fix the various library and apps files to free up pkeys obtained from
4030 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
4033 *) Fix reference counting in X509_PUBKEY_get(). This makes
4034 demos/maurice/example2.c work, amongst others, probably.
4035 [Steve Henson and Ben Laurie]
4037 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
4038 `openssl' and second, the shortcut symlinks for the `openssl <command>'
4039 are no longer created. This way we have a single and consistent command
4040 line interface `openssl <command>', similar to `cvs <command>'.
4041 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
4043 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
4044 BIT STRING wrapper always have zero unused bits.
4047 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
4050 *) Make the top-level INSTALL documentation easier to understand.
4053 *) Makefiles updated to exit if an error occurs in a sub-directory
4054 make (including if user presses ^C) [Paul Sutton]
4056 *) Make Montgomery context stuff explicit in RSA data structure.
4059 *) Fix build order of pem and err to allow for generated pem.h.
4062 *) Fix renumbering bug in X509_NAME_delete_entry().
4065 *) Enhanced the err-ins.pl script so it makes the error library number
4066 global and can add a library name. This is needed for external ASN1 and
4067 other error libraries.
4070 *) Fixed sk_insert which never worked properly.
4073 *) Fix ASN1 macros so they can handle indefinite length construted
4074 EXPLICIT tags. Some non standard certificates use these: they can now
4078 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
4079 into a single doc/ssleay.txt bundle. This way the information is still
4080 preserved but no longer messes up this directory. Now it's new room for
4081 the new set of documenation files.
4082 [Ralf S. Engelschall]
4084 *) SETs were incorrectly DER encoded. This was a major pain, because they
4085 shared code with SEQUENCEs, which aren't coded the same. This means that
4086 almost everything to do with SETs or SEQUENCEs has either changed name or
4087 number of arguments.
4088 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4090 *) Fix test data to work with the above.
4093 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4094 was already fixed by Eric for 0.9.1 it seems.
4095 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4097 *) Autodetect FreeBSD3.
4100 *) Fix various bugs in Configure. This affects the following platforms:
4104 unixware-2.0-pentium
4108 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4109 before they are needed.
4112 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4116 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4118 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4119 changed SSLeay to OpenSSL in version strings.
4120 [Ralf S. Engelschall]
4122 *) Some fixups to the top-level documents.
4125 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4126 because the symlink to include/ was missing.
4127 [Ralf S. Engelschall]
4129 *) Incorporated the popular no-RSA/DSA-only patches
4130 which allow to compile a RSA-free SSLeay.
4131 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4133 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4134 when "ssleay" is still not found.
4135 [Ralf S. Engelschall]
4137 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4138 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4140 *) Updated the README file.
4141 [Ralf S. Engelschall]
4143 *) Added various .cvsignore files in the CVS repository subdirs
4144 to make a "cvs update" really silent.
4145 [Ralf S. Engelschall]
4147 *) Recompiled the error-definition header files and added
4148 missing symbols to the Win32 linker tables.
4149 [Ralf S. Engelschall]
4151 *) Cleaned up the top-level documents;
4152 o new files: CHANGES and LICENSE
4153 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4154 o merged COPYRIGHT into LICENSE
4155 o removed obsolete TODO file
4156 o renamed MICROSOFT to INSTALL.W32
4157 [Ralf S. Engelschall]
4159 *) Removed dummy files from the 0.9.1b source tree:
4160 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4161 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4162 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4163 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4164 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4165 [Ralf S. Engelschall]
4167 *) Added various platform portability fixes.
4170 *) The Genesis of the OpenSSL rpject:
4171 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4172 Young and Tim J. Hudson created while they were working for C2Net until
4174 [The OpenSSL Project]
4177 Changes between 0.9.0b and 0.9.1b [not released]
4179 *) Updated a few CA certificates under certs/
4182 *) Changed some BIGNUM api stuff.
4185 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4186 DGUX x86, Linux Alpha, etc.
4189 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4190 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4194 *) Add -strparse option to asn1pars program which parses nested
4196 [Dr Stephen Henson <shenson@bigfoot.com>]
4198 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4201 *) DSA fix for "ca" program.
4204 *) Added "-genkey" option to "dsaparam" program.
4207 *) Added RIPE MD160 (rmd160) message digest.
4210 *) Added -a (all) option to "ssleay version" command.
4213 *) Added PLATFORM define which is the id given to Configure.
4216 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4219 *) Extended the ASN.1 parser routines.
4222 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4225 *) Added a BN_CTX to the BN library.
4228 *) Fixed the weak key values in DES library
4231 *) Changed API in EVP library for cipher aliases.
4234 *) Added support for RC2/64bit cipher.
4237 *) Converted the lhash library to the crypto/mem.c functions.
4240 *) Added more recognized ASN.1 object ids.
4243 *) Added more RSA padding checks for SSL/TLS.
4246 *) Added BIO proxy/filter functionality.
4249 *) Added extra_certs to SSL_CTX which can be used
4250 send extra CA certificates to the client in the CA cert chain sending
4251 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4254 *) Now Fortezza is denied in the authentication phase because
4255 this is key exchange mechanism is not supported by SSLeay at all.
4258 *) Additional PKCS1 checks.
4261 *) Support the string "TLSv1" for all TLS v1 ciphers.
4264 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4265 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4268 *) Fixed a few memory leaks.
4271 *) Fixed various code and comment typos.
4274 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4275 bytes sent in the client random.
4276 [Edward Bishop <ebishop@spyglass.com>]