5 Changes between 0.9.6b and 0.9.6c [XX xxx XXXX]
7 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8 OBJ_ld_ce should be OBJ_id_ce.
9 Also some ip-pda OIDs in crypto/objects/objects.txt were
10 incorrect (cf. RFC 3039).
11 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
13 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14 returns early because it has nothing to do.
15 [Andy Schneider <andy.schneider@bjss.co.uk>]
17 *) [In 0.9.6c-engine release:]
18 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
19 [Andy Schneider <andy.schneider@bjss.co.uk>]
21 *) [In 0.9.6c-engine release:]
22 Add support for Cryptographic Appliance's keyserver technology.
23 (Use engine 'keyclient')
24 [Cryptographic Appliances and Geoff Thorpe]
26 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
27 is called via tools/c89.sh because arguments have to be
28 rearranged (all '-L' options must appear before the first object
30 [Richard Shapiro <rshapiro@abinitio.com>]
32 *) [In 0.9.6c-engine release:]
33 Add support for Broadcom crypto accelerator cards, backported
35 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
37 *) [In 0.9.6c-engine release:]
38 Add support for SureWare crypto accelerator cards from
39 Baltimore Technologies. (Use engine 'sureware')
40 [Baltimore Technologies and Mark Cox]
42 *) [In 0.9.6c-engine release:]
43 Add support for crypto accelerator cards from Accelerated
44 Encryption Processing, www.aep.ie. (Use engine 'aep')
45 [AEP Inc. and Mark Cox]
47 *) Add a configuration entry for gcc on UnixWare.
48 [Gary Benson <gbenson@redhat.com>]
50 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
51 messages are stored in a single piece (fixed-length part and
52 variable-length part combined) and fix various bugs found on the way.
55 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
56 instead. BIO_gethostbyname() does not know what timeouts are
57 appropriate, so entries would stay in cache even when they have
59 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
61 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
62 faced with a pathologically small ClientHello fragment that does
63 not contain client_version: Instead of aborting with an error,
64 simply choose the highest available protocol version (i.e.,
65 TLS 1.0 unless it is disabled). In practice, ClientHello
66 messages are never sent like this, but this change gives us
67 strictly correct behaviour at least for TLS.
70 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
71 never resets s->method to s->ctx->method when called from within
72 one of the SSL handshake functions.
73 [Bodo Moeller; problem pointed out by Niko Baric]
75 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
76 (sent using the client's version number) if client_version is
77 smaller than the protocol version in use. Also change
78 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
79 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
80 the client will at least see that alert.
83 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
87 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
88 client receives HelloRequest while in a handshake.
89 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
91 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
92 should end in 'break', not 'goto end' which circuments various
93 cleanups done in state SSL_ST_OK. But session related stuff
94 must be disabled for SSL_ST_OK in the case that we just sent a
97 Also avoid some overhead by not calling ssl_init_wbio_buffer()
98 before just sending a HelloRequest.
99 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
101 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
102 reveal whether illegal block cipher padding was found or a MAC
103 verification error occured. (Neither SSLerr() codes nor alerts
104 are directly visible to potential attackers, but the information
105 may leak via logfiles.)
107 Similar changes are not required for the SSL 2.0 implementation
108 because the number of padding bytes is sent in clear for SSL 2.0,
109 and the extra bytes are just ignored. However ssl/s2_pkt.c
110 failed to verify that the purported number of padding bytes is in
114 *) Add OpenUNIX-8 support including shared libraries
115 (Boyd Lynn Gerber <gerberb@zenez.com>).
118 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
119 'wristwatch attack' using huge encoding parameters (cf.
120 James H. Manger's CRYPTO 2001 paper). Note that the
121 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
122 encoding parameters and hence was not vulnerable.
126 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
128 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
129 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
130 followed by modular reduction.
131 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
133 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
134 equivalent based on BN_pseudo_rand() instead of BN_rand().
137 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
138 This function was broken, as the check for a new client hello message
139 to handle SGC did not allow these large messages.
140 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
143 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
146 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
147 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
150 *) Rework the configuration and shared library support for Tru64 Unix.
151 The configuration part makes use of modern compiler features and
152 still retains old compiler behavior for those that run older versions
153 of the OS. The shared library support part includes a variant that
154 uses the RPATH feature, and is available through the special
155 configuration target "alpha-cc-rpath", which will never be selected
157 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
159 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
160 with the same message size as in ssl3_get_certificate_request().
161 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
162 messages might inadvertently be reject as too long.
163 [Petr Lampa <lampa@fee.vutbr.cz>]
165 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
168 *) Modified SSL library such that the verify_callback that has been set
169 specificly for an SSL object with SSL_set_verify() is actually being
170 used. Before the change, a verify_callback set with this function was
171 ignored and the verify_callback() set in the SSL_CTX at the time of
172 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
173 to allow the necessary settings.
176 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
177 explicitly to NULL, as at least on Solaris 8 this seems not always to be
178 done automatically (in contradiction to the requirements of the C
179 standard). This made problems when used from OpenSSH.
182 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
183 dh->length and always used
185 BN_rand_range(priv_key, dh->p).
187 BN_rand_range() is not necessary for Diffie-Hellman, and this
188 specific range makes Diffie-Hellman unnecessarily inefficient if
189 dh->length (recommended exponent length) is much smaller than the
190 length of dh->p. We could use BN_rand_range() if the order of
191 the subgroup was stored in the DH structure, but we only have
196 BN_rand(priv_key, l, ...)
198 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
204 RSA_eay_public_encrypt
205 RSA_eay_private_decrypt
206 RSA_eay_private_encrypt (signing)
207 RSA_eay_public_decrypt (signature verification)
209 (default implementations for RSA_public_encrypt,
210 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
211 always reject numbers >= n.
214 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
215 to synchronize access to 'locking_thread'. This is necessary on
216 systems where access to 'locking_thread' (an 'unsigned long'
217 variable) is not atomic.
220 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
221 *before* setting the 'crypto_lock_rand' flag. The previous code had
222 a race condition if 0 is a valid thread ID.
223 [Travis Vitek <vitek@roguewave.com>]
225 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
227 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
228 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
229 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
230 PRNG state recovery was possible based on the output of
231 one PRNG request appropriately sized to gain knowledge on
232 'md' followed by enough consecutive 1-byte PRNG requests
233 to traverse all of 'state'.
235 1. When updating 'md_local' (the current thread's copy of 'md')
236 during PRNG output generation, hash all of the previous
237 'md_local' value, not just the half used for PRNG output.
239 2. Make the number of bytes from 'state' included into the hash
240 independent from the number of PRNG bytes requested.
242 The first measure alone would be sufficient to avoid
243 Markku-Juhani's attack. (Actually it had never occurred
244 to me that the half of 'md_local' used for chaining was the
245 half from which PRNG output bytes were taken -- I had always
246 assumed that the secret half would be used.) The second
247 measure makes sure that additional data from 'state' is never
248 mixed into 'md_local' in small portions; this heuristically
249 further strengthens the PRNG.
252 *) Fix crypto/bn/asm/mips3.s.
255 *) When only the key is given to "enc", the IV is undefined. Print out
256 an error message in this case.
259 *) Handle special case when X509_NAME is empty in X509 printing routines.
262 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
263 positive and less than q.
266 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
267 used: it isn't thread safe and the add_lock_callback should handle
269 [Paul Rose <Paul.Rose@bridge.com>]
271 *) Verify that incoming data obeys the block size in
272 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
276 [Ulf Möller, Bodo Möller]
278 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
279 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
280 when fixing the server behaviour for backwards-compatible 'client
281 hello' messages. (Note that the attack is impractical against
282 SSL 3.0 and TLS 1.0 anyway because length and version checking
283 means that the probability of guessing a valid ciphertext is
284 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
287 Before 0.9.5, the countermeasure (hide the error by generating a
288 random 'decryption result') did not work properly because
289 ERR_clear_error() was missing, meaning that SSL_get_error() would
290 detect the supposedly ignored error.
292 Both problems are now fixed.
295 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
296 (previously it was 1024).
299 *) Fix for compatibility mode trust settings: ignore trust settings
300 unless some valid trust or reject settings are present.
303 *) Fix for blowfish EVP: its a variable length cipher.
306 *) Fix various bugs related to DSA S/MIME verification. Handle missing
307 parameters in DSA public key structures and return an error in the
308 DSA routines if parameters are absent.
311 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
312 in the current directory if neither $RANDFILE nor $HOME was set.
313 RAND_file_name() in 0.9.6a returned NULL in this case. This has
314 caused some confusion to Windows users who haven't defined $HOME.
315 Thus RAND_file_name() is changed again: e_os.h can define a
316 DEFAULT_HOME, which will be used if $HOME is not set.
317 For Windows, we use "C:"; on other platforms, we still require
318 environment variables.
320 *) Move 'if (!initialized) RAND_poll()' into regions protected by
321 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
322 having multiple threads call RAND_poll() concurrently.
325 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
326 combination of a flag and a thread ID variable.
327 Otherwise while one thread is in ssleay_rand_bytes (which sets the
328 flag), *other* threads can enter ssleay_add_bytes without obeying
329 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
330 that they do not hold after the first thread unsets add_do_not_lock).
333 *) Change bctest again: '-x' expressions are not available in all
337 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
339 *) Fix a couple of memory leaks in PKCS7_dataDecode()
340 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
342 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
343 the default extension for executables, if any. Also, make the perl
344 scripts that use symlink() to test if it really exists and use "cp"
345 if it doesn't. All this made OpenSSL compilable and installable in
349 *) Fix for asn1_GetSequence() for indefinite length constructed data.
350 If SEQUENCE is length is indefinite just set c->slen to the total
351 amount of data available.
352 [Steve Henson, reported by shige@FreeBSD.org]
353 [This change does not apply to 0.9.7.]
355 *) Change bctest to avoid here-documents inside command substitution
356 (workaround for FreeBSD /bin/sh bug).
357 For compatibility with Ultrix, avoid shell functions (introduced
358 in the bctest version that searches along $PATH).
361 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
362 with des_encrypt() defined on some operating systems, like Solaris
366 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
367 On the Importance of Eliminating Errors in Cryptographic
368 Computations, J. Cryptology 14 (2001) 2, 101-119,
369 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
372 *) MIPS assembler BIGNUM division bug fix.
375 *) Disabled incorrect Alpha assembler code.
378 *) Fix PKCS#7 decode routines so they correctly update the length
379 after reading an EOC for the EXPLICIT tag.
381 [This change does not apply to 0.9.7.]
383 *) Fix bug in PKCS#12 key generation routines. This was triggered
384 if a 3DES key was generated with a 0 initial byte. Include
385 PKCS12_BROKEN_KEYGEN compilation option to retain the old
386 (but broken) behaviour.
389 *) Enhance bctest to search for a working bc along $PATH and print
391 [Tim Rice <tim@multitalents.net> via Richard Levitte]
393 *) Fix memory leaks in err.c: free err_data string if necessary;
394 don't write to the wrong index in ERR_set_error_data.
397 *) Implement ssl23_peek (analogous to ssl23_read), which previously
401 *) Replace rdtsc with _emit statements for VC++ version 5.
402 [Jeremy Cooper <jeremy@baymoo.org>]
404 *) Make it possible to reuse SSLv2 sessions.
407 *) In copy_email() check for >= 0 as a return value for
408 X509_NAME_get_index_by_NID() since 0 is a valid index.
409 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
411 *) Avoid coredump with unsupported or invalid public keys by checking if
412 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
413 PKCS7_verify() fails with non detached data.
416 *) Don't use getenv in library functions when run as setuid/setgid.
417 New function OPENSSL_issetugid().
420 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
421 due to incorrect handling of multi-threading:
423 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
425 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
427 3. Count how many times MemCheck_off() has been called so that
428 nested use can be treated correctly. This also avoids
429 inband-signalling in the previous code (which relied on the
430 assumption that thread ID 0 is impossible).
433 *) Add "-rand" option also to s_client and s_server.
436 *) Fix CPU detection on Irix 6.x.
437 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
438 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
440 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
443 [This change does not apply to 0.9.7.]
445 *) Use the cached encoding of an X509_NAME structure rather than
446 copying it. This is apparently the reason for the libsafe "errors"
447 but the code is actually correct.
450 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
451 Bleichenbacher's DSA attack.
452 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
453 to be set and top=0 forces the highest bit to be set; top=-1 is new
454 and leaves the highest bit random.
455 [Ulf Moeller, Bodo Moeller]
457 *) In the NCONF_...-based implementations for CONF_... queries
458 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
459 a temporary CONF structure with the data component set to NULL
460 (which gives segmentation faults in lh_retrieve).
461 Instead, use NULL for the CONF pointer in CONF_get_string and
462 CONF_get_number (which may use environment variables) and directly
463 return NULL from CONF_get_section.
466 *) Fix potential buffer overrun for EBCDIC.
469 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
470 keyUsage if basicConstraints absent for a CA.
473 *) Make SMIME_write_PKCS7() write mail header values with a format that
474 is more generally accepted (no spaces before the semicolon), since
475 some programs can't parse those values properly otherwise. Also make
476 sure BIO's that break lines after each write do not create invalid
480 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
481 macros previously used would not encode an empty SEQUENCE OF
482 and break the signature.
484 [This change does not apply to 0.9.7.]
486 *) Zero the premaster secret after deriving the master secret in
490 *) Add some EVP_add_digest_alias registrations (as found in
491 OpenSSL_add_all_digests()) to SSL_library_init()
492 aka OpenSSL_add_ssl_algorithms(). This provides improved
493 compatibility with peers using X.509 certificates
494 with unconventional AlgorithmIdentifier OIDs.
497 *) Fix for Irix with NO_ASM.
498 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
500 *) ./config script fixes.
501 [Ulf Moeller, Richard Levitte]
503 *) Fix 'openssl passwd -1'.
506 *) Change PKCS12_key_gen_asc() so it can cope with non null
507 terminated strings whose length is passed in the passlen
508 parameter, for example from PEM callbacks. This was done
509 by adding an extra length parameter to asc2uni().
510 [Steve Henson, reported by <oddissey@samsung.co.kr>]
512 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
513 call failed, free the DSA structure.
516 *) Fix to uni2asc() to cope with zero length Unicode strings.
517 These are present in some PKCS#12 files.
520 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
521 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
522 when writing a 32767 byte record.
523 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
525 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
526 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
528 (RSA objects have a reference count access to which is protected
529 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
530 so they are meant to be shared between threads.)
531 [Bodo Moeller, Geoff Thorpe; original patch submitted by
532 "Reddie, Steven" <Steven.Reddie@ca.com>]
534 *) Fix a deadlock in CRYPTO_mem_leaks().
537 *) Use better test patterns in bntest.
540 *) rand_win.c fix for Borland C.
543 *) BN_rshift bugfix for n == 0.
546 *) Add a 'bctest' script that checks for some known 'bc' bugs
547 so that 'make test' does not abort just because 'bc' is broken.
550 *) Store verify_result within SSL_SESSION also for client side to
551 avoid potential security hole. (Re-used sessions on the client side
552 always resulted in verify_result==X509_V_OK, not using the original
553 result of the server certificate verification.)
556 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
557 SSL3_RT_APPLICATION_DATA, return 0.
558 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
562 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
563 releases, have been re-implemented by renaming the previous
564 implementations of ssl2_read and ssl3_read to ssl2_read_internal
565 and ssl3_read_internal, respectively, and adding 'peek' parameters
566 to them. The new ssl[23]_{read,peek} functions are calls to
567 ssl[23]_read_internal with the 'peek' flag set appropriately.
568 A 'peek' parameter has also been added to ssl3_read_bytes, which
569 does the actual work for ssl3_read_internal.
572 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
573 the method-specific "init()" handler. Also clean up ex_data after
574 calling the method-specific "finish()" handler. Previously, this was
575 happening the other way round.
578 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
579 The previous value, 12, was not always sufficient for BN_mod_exp().
582 *) Make sure that shared libraries get the internal name engine with
583 the full version number and not just 0. This should mark the
584 shared libraries as not backward compatible. Of course, this should
585 be changed again when we can guarantee backward binary compatibility.
588 *) Add support for shared libraries under Irix.
589 [Albert Chin-A-Young <china@thewrittenword.com>]
591 *) Add configuration option to build on Linux on both big-endian and
593 [Ralf Baechle <ralf@uni-koblenz.de>]
595 *) Add the possibility to create shared libraries on HP-UX
598 *) Fix typo in get_cert_by_subject() in by_dir.c
599 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
601 *) Rework the system to generate shared libraries:
603 - Make note of the expected extension for the shared libraries and
604 if there is a need for symbolic links from for example libcrypto.so.0
605 to libcrypto.so.0.9.7. There is extended info in Configure for
608 - Make as few rebuilds of the shared libraries as possible.
610 - Still avoid linking the OpenSSL programs with the shared libraries.
612 - When installing, install the shared libraries separately from the
616 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
618 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
619 and not in SSL_clear because the latter is also used by the
620 accept/connect functions; previously, the settings made by
621 SSL_set_read_ahead would be lost during the handshake.
622 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
624 *) Correct util/mkdef.pl to be selective about disabled algorithms.
625 Previously, it would create entries for disableed algorithms no
629 *) Added several new manual pages for SSL_* function.
632 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
634 *) In ssl23_get_client_hello, generate an error message when faced
635 with an initial SSL 3.0/TLS record that is too small to contain the
636 first two bytes of the ClientHello message, i.e. client_version.
637 (Note that this is a pathologic case that probably has never happened
638 in real life.) The previous approach was to use the version number
639 from the record header as a substitute; but our protocol choice
640 should not depend on that one because it is not authenticated
641 by the Finished messages.
644 *) More robust randomness gathering functions for Windows.
645 [Jeffrey Altman <jaltman@columbia.edu>]
647 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
648 not set then we don't setup the error code for issuer check errors
649 to avoid possibly overwriting other errors which the callback does
650 handle. If an application does set the flag then we assume it knows
651 what it is doing and can handle the new informational codes
655 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
656 a general "ANY" type, as such it should be able to decode anything
657 including tagged types. However it didn't check the class so it would
658 wrongly interpret tagged types in the same way as their universal
659 counterpart and unknown types were just rejected. Changed so that the
660 tagged and unknown types are handled in the same way as a SEQUENCE:
661 that is the encoding is stored intact. There is also a new type
662 "V_ASN1_OTHER" which is used when the class is not universal, in this
663 case we have no idea what the actual type is so we just lump them all
667 *) On VMS, stdout may very well lead to a file that is written to
668 in a record-oriented fashion. That means that every write() will
669 write a separate record, which will be read separately by the
670 programs trying to read from it. This can be very confusing.
672 The solution is to put a BIO filter in the way that will buffer
673 text until a linefeed is reached, and then write everything a
674 line at a time, so every record written will be an actual line,
675 not chunks of lines and not (usually doesn't happen, but I've
676 seen it once) several lines in one record. BIO_f_linebuffer() is
679 Currently, it's a VMS-only method, because that's where it has
680 been tested well enough.
683 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
684 it can return incorrect results.
685 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
686 but it was in 0.9.6-beta[12].)
689 *) Disable the check for content being present when verifying detached
690 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
691 include zero length content when signing messages.
694 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
695 BIO_ctrl (for BIO pairs).
698 *) Add DSO method for VMS.
701 *) Bug fix: Montgomery multiplication could produce results with the
705 *) Add RPM specification openssl.spec and modify it to build three
706 packages. The default package contains applications, application
707 documentation and run-time libraries. The devel package contains
708 include files, static libraries and function documentation. The
709 doc package contains the contents of the doc directory. The original
710 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
713 *) Add a large number of documentation files for many SSL routines.
714 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
716 *) Add a configuration entry for Sony News 4.
717 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
719 *) Don't set the two most significant bits to one when generating a
720 random number < q in the DSA library.
723 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
724 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
725 the underlying transport is blocking) if a handshake took place.
726 (The default behaviour is needed by applications such as s_client
727 and s_server that use select() to determine when to use SSL_read;
728 but for applications that know in advance when to expect data, it
729 just makes things more complicated.)
732 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
736 *) Add a few more EBCDIC conditionals that make `req' and `x509'
737 work better on such systems.
738 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
740 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
741 Update PKCS12_parse() so it copies the friendlyName and the
742 keyid to the certificates aux info.
745 *) Fix bug in PKCS7_verify() which caused an infinite loop
746 if there was more than one signature.
747 [Sven Uszpelkat <su@celocom.de>]
749 *) Major change in util/mkdef.pl to include extra information
750 about each symbol, as well as presentig variables as well
751 as functions. This change means that there's n more need
752 to rebuild the .num files when some algorithms are excluded.
755 *) Allow the verify time to be set by an application,
756 rather than always using the current time.
759 *) Phase 2 verify code reorganisation. The certificate
760 verify code now looks up an issuer certificate by a
761 number of criteria: subject name, authority key id
762 and key usage. It also verifies self signed certificates
763 by the same criteria. The main comparison function is
764 X509_check_issued() which performs these checks.
766 Lot of changes were necessary in order to support this
767 without completely rewriting the lookup code.
769 Authority and subject key identifier are now cached.
771 The LHASH 'certs' is X509_STORE has now been replaced
772 by a STACK_OF(X509_OBJECT). This is mainly because an
773 LHASH can't store or retrieve multiple objects with
776 As a result various functions (which were all internal
777 use only) have changed to handle the new X509_STORE
778 structure. This will break anything that messed round
779 with X509_STORE internally.
781 The functions X509_STORE_add_cert() now checks for an
782 exact match, rather than just subject name.
784 The X509_STORE API doesn't directly support the retrieval
785 of multiple certificates matching a given criteria, however
786 this can be worked round by performing a lookup first
787 (which will fill the cache with candidate certificates)
788 and then examining the cache for matches. This is probably
789 the best we can do without throwing out X509_LOOKUP
790 entirely (maybe later...).
792 The X509_VERIFY_CTX structure has been enhanced considerably.
794 All certificate lookup operations now go via a get_issuer()
795 callback. Although this currently uses an X509_STORE it
796 can be replaced by custom lookups. This is a simple way
797 to bypass the X509_STORE hackery necessary to make this
798 work and makes it possible to use more efficient techniques
799 in future. A very simple version which uses a simple
800 STACK for its trusted certificate store is also provided
801 using X509_STORE_CTX_trusted_stack().
803 The verify_cb() and verify() callbacks now have equivalents
804 in the X509_STORE_CTX structure.
806 X509_STORE_CTX also has a 'flags' field which can be used
807 to customise the verify behaviour.
810 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
811 excludes S/MIME capabilities.
814 *) When a certificate request is read in keep a copy of the
815 original encoding of the signed data and use it when outputing
816 again. Signatures then use the original encoding rather than
817 a decoded, encoded version which may cause problems if the
818 request is improperly encoded.
821 *) For consistency with other BIO_puts implementations, call
822 buffer_write(b, ...) directly in buffer_puts instead of calling
825 In BIO_puts, increment b->num_write as in BIO_write.
826 [Peter.Sylvester@EdelWeb.fr]
828 *) Fix BN_mul_word for the case where the word is 0. (We have to use
829 BN_zero, we may not return a BIGNUM with an array consisting of
833 *) Avoid calling abort() from within the library when problems are
834 detected, except if preprocessor symbols have been defined
835 (such as REF_CHECK, BN_DEBUG etc.).
838 *) New openssl application 'rsautl'. This utility can be
839 used for low level RSA operations. DER public key
840 BIO/fp routines also added.
843 *) New Configure entry and patches for compiling on QNX 4.
844 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
846 *) A demo state-machine implementation was sponsored by
847 Nuron (http://www.nuron.com/) and is now available in
851 *) New options added to the 'dgst' utility for signature
852 generation and verification.
855 *) Unrecognized PKCS#7 content types are now handled via a
856 catch all ASN1_TYPE structure. This allows unsupported
857 types to be stored as a "blob" and an application can
858 encode and decode it manually.
861 *) Fix various signed/unsigned issues to make a_strex.c
863 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
865 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
866 length if passed a buffer. ASN1_INTEGER_to_BN failed
867 if passed a NULL BN and its argument was negative.
868 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
870 *) Modification to PKCS#7 encoding routines to output definite
871 length encoding. Since currently the whole structures are in
872 memory there's not real point in using indefinite length
873 constructed encoding. However if OpenSSL is compiled with
874 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
877 *) Added BIO_vprintf() and BIO_vsnprintf().
880 *) Added more prefixes to parse for in the the strings written
881 through a logging bio, to cover all the levels that are available
882 through syslog. The prefixes are now:
884 PANIC, EMERG, EMR => LOG_EMERG
885 ALERT, ALR => LOG_ALERT
886 CRIT, CRI => LOG_CRIT
887 ERROR, ERR => LOG_ERR
888 WARNING, WARN, WAR => LOG_WARNING
889 NOTICE, NOTE, NOT => LOG_NOTICE
890 INFO, INF => LOG_INFO
891 DEBUG, DBG => LOG_DEBUG
893 and as before, if none of those prefixes are present at the
894 beginning of the string, LOG_ERR is chosen.
896 On Win32, the LOG_* levels are mapped according to this:
898 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
899 LOG_WARNING => EVENTLOG_WARNING_TYPE
900 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
904 *) Made it possible to reconfigure with just the configuration
905 argument "reconf" or "reconfigure". The command line arguments
906 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
907 and are retrieved from there when reconfiguring.
911 [Assar Westerlund <assar@sics.se>, Richard Levitte]
913 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
916 *) The obj_dat.pl script was messing up the sorting of object
917 names. The reason was that it compared the quoted version
918 of strings as a result "OCSP" > "OCSP Signing" because
919 " > SPACE. Changed script to store unquoted versions of
920 names and add quotes on output. It was also omitting some
921 names from the lookup table if they were given a default
922 value (that is if SN is missing it is given the same
923 value as LN and vice versa), these are now added on the
924 grounds that if an object has a name we should be able to
925 look it up. Finally added warning output when duplicate
926 short or long names are found.
929 *) Changes needed for Tandem NSK.
930 [Scott Uroff <scott@xypro.com>]
932 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
933 RSA_padding_check_SSLv23(), special padding was never detected
934 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
935 version rollback attacks was not effective.
937 In s23_clnt.c, don't use special rollback-attack detection padding
938 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
939 client; similarly, in s23_srvr.c, don't do the rollback check if
940 SSL 2.0 is the only protocol enabled in the server.
943 *) Make it possible to get hexdumps of unprintable data with 'openssl
944 asn1parse'. By implication, the functions ASN1_parse_dump() and
945 BIO_dump_indent() are added.
948 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
949 these print out strings and name structures based on various
950 flags including RFC2253 support and proper handling of
951 multibyte characters. Added options to the 'x509' utility
952 to allow the various flags to be set.
955 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
956 Also change the functions X509_cmp_current_time() and
957 X509_gmtime_adj() work with an ASN1_TIME structure,
958 this will enable certificates using GeneralizedTime in validity
962 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
963 negative public key encodings) on by default,
964 NO_NEG_PUBKEY_BUG can be set to disable it.
967 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
968 content octets. An i2c_ASN1_OBJECT is unnecessary because
969 the encoding can be trivially obtained from the structure.
972 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
973 not read locks (CRYPTO_r_[un]lock).
976 *) A first attempt at creating official support for shared
977 libraries through configuration. I've kept it so the
978 default is static libraries only, and the OpenSSL programs
979 are always statically linked for now, but there are
980 preparations for dynamic linking in place.
981 This has been tested on Linux and Tru64.
984 *) Randomness polling function for Win9x, as described in:
985 Peter Gutmann, Software Generation of Practically Strong
989 *) Fix so PRNG is seeded in req if using an already existing
993 *) New options to smime application. -inform and -outform
994 allow alternative formats for the S/MIME message including
995 PEM and DER. The -content option allows the content to be
996 specified separately. This should allow things like Netscape
997 form signing output easier to verify.
1000 *) Fix the ASN1 encoding of tags using the 'long form'.
1003 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1004 STRING types. These convert content octets to and from the
1005 underlying type. The actual tag and length octets are
1006 already assumed to have been read in and checked. These
1007 are needed because all other string types have virtually
1008 identical handling apart from the tag. By having versions
1009 of the ASN1 functions that just operate on content octets
1010 IMPLICIT tagging can be handled properly. It also allows
1011 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1012 and ASN1_INTEGER are identical apart from the tag.
1015 *) Change the handling of OID objects as follows:
1017 - New object identifiers are inserted in objects.txt, following
1018 the syntax given in objects.README.
1019 - objects.pl is used to process obj_mac.num and create a new
1021 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1024 This is currently kind of a hack, and the perl code in objects.pl
1025 isn't very elegant, but it works as I intended. The simplest way
1026 to check that it worked correctly is to look in obj_dat.h and
1027 check the array nid_objs and make sure the objects haven't moved
1028 around (this is important!). Additions are OK, as well as
1029 consistent name changes.
1032 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1035 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1036 The given file adds to whatever has already been seeded into the
1037 random pool through the RANDFILE configuration file option or
1038 environment variable, or the default random state file.
1041 *) mkstack.pl now sorts each macro group into lexical order.
1042 Previously the output order depended on the order the files
1043 appeared in the directory, resulting in needless rewriting
1047 *) Patches to make OpenSSL compile under Win32 again. Mostly
1048 work arounds for the VC++ problem that it treats func() as
1049 func(void). Also stripped out the parts of mkdef.pl that
1050 added extra typesafe functions: these no longer exist.
1053 *) Reorganisation of the stack code. The macros are now all
1054 collected in safestack.h . Each macro is defined in terms of
1055 a "stack macro" of the form SKM_<name>(type, a, b). The
1056 DEBUG_SAFESTACK is now handled in terms of function casts,
1057 this has the advantage of retaining type safety without the
1058 use of additional functions. If DEBUG_SAFESTACK is not defined
1059 then the non typesafe macros are used instead. Also modified the
1060 mkstack.pl script to handle the new form. Needs testing to see
1061 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1062 the default if no major problems. Similar behaviour for ASN1_SET_OF
1063 and PKCS12_STACK_OF.
1066 *) When some versions of IIS use the 'NET' form of private key the
1067 key derivation algorithm is different. Normally MD5(password) is
1068 used as a 128 bit RC4 key. In the modified case
1069 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1070 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1071 as the old Netscape_RSA functions except they have an additional
1072 'sgckey' parameter which uses the modified algorithm. Also added
1073 an -sgckey command line option to the rsa utility. Thanks to
1074 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1075 algorithm to openssl-dev.
1078 *) The evp_local.h macros were using 'c.##kname' which resulted in
1079 invalid expansion on some systems (SCO 5.0.5 for example).
1080 Corrected to 'c.kname'.
1081 [Phillip Porch <root@theporch.com>]
1083 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1084 a STACK of email addresses from a certificate or request, these look
1085 in the subject name and the subject alternative name extensions and
1086 omit any duplicate addresses.
1089 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1090 This makes DSA verification about 2 % faster.
1093 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1094 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1095 plus overhead for 1024 bit moduli).
1096 This makes exponentiations about 0.5 % faster for 1024 bit
1097 exponents (as measured by "openssl speed rsa2048").
1100 *) Rename memory handling macros to avoid conflicts with other
1102 Malloc => OPENSSL_malloc
1103 Malloc_locked => OPENSSL_malloc_locked
1104 Realloc => OPENSSL_realloc
1105 Free => OPENSSL_free
1108 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1109 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1112 *) CygWin32 support.
1113 [John Jarvie <jjarvie@newsguy.com>]
1115 *) The type-safe stack code has been rejigged. It is now only compiled
1116 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1117 by default all type-specific stack functions are "#define"d back to
1118 standard stack functions. This results in more streamlined output
1119 but retains the type-safety checking possibilities of the original
1123 *) The STACK code has been cleaned up, and certain type declarations
1124 that didn't make a lot of sense have been brought in line. This has
1125 also involved a cleanup of sorts in safestack.h to more correctly
1126 map type-safe stack functions onto their plain stack counterparts.
1127 This work has also resulted in a variety of "const"ifications of
1128 lots of the code, especially "_cmp" operations which should normally
1129 be prototyped with "const" parameters anyway.
1132 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1133 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1134 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1135 where all of 'md' is used each time the PRNG is used, but 'state'
1136 is used only indexed by a cyclic counter. As entropy may not be
1137 well distributed from the beginning, 'md' is important as a
1138 chaining variable. However, the output function chains only half
1139 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1140 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1141 in all of 'state' being rewritten, with the new values depending
1142 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1145 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1146 the handshake is continued after ssl_verify_cert_chain();
1147 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1148 can lead to 'unexplainable' connection aborts later.
1149 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1151 *) Major EVP API cipher revision.
1152 Add hooks for extra EVP features. This allows various cipher
1153 parameters to be set in the EVP interface. Support added for variable
1154 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1155 setting of RC2 and RC5 parameters.
1157 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1160 Remove lots of duplicated code from the EVP library. For example *every*
1161 cipher init() function handles the 'iv' in the same way according to the
1162 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1163 for CFB and OFB modes they zero ctx->num.
1165 New functionality allows removal of S/MIME code RC2 hack.
1167 Most of the routines have the same form and so can be declared in terms
1170 By shifting this to the top level EVP_CipherInit() it can be removed from
1171 all individual ciphers. If the cipher wants to handle IVs or keys
1172 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1175 Change lots of functions like EVP_EncryptUpdate() to now return a
1176 value: although software versions of the algorithms cannot fail
1177 any installed hardware versions can.
1180 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1181 this option is set, tolerate broken clients that send the negotiated
1182 protocol version number instead of the requested protocol version
1186 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1187 i.e. non-zero for export ciphersuites, zero otherwise.
1188 Previous versions had this flag inverted, inconsistent with
1189 rsa_tmp_cb (..._TMP_RSA_CB).
1190 [Bodo Moeller; problem reported by Amit Chopra]
1192 *) Add missing DSA library text string. Work around for some IIS
1193 key files with invalid SEQUENCE encoding.
1196 *) Add a document (doc/standards.txt) that list all kinds of standards
1197 and so on that are implemented in OpenSSL.
1200 *) Enhance c_rehash script. Old version would mishandle certificates
1201 with the same subject name hash and wouldn't handle CRLs at all.
1202 Added -fingerprint option to crl utility, to support new c_rehash
1206 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1209 *) Fix for SSL server purpose checking. Server checking was
1210 rejecting certificates which had extended key usage present
1211 but no ssl client purpose.
1212 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1214 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1215 is a little unclear about how a blank password is handled.
1216 Since the password in encoded as a BMPString with terminating
1217 double NULL a zero length password would end up as just the
1218 double NULL. However no password at all is different and is
1219 handled differently in the PKCS#12 key generation code. NS
1220 treats a blank password as zero length. MSIE treats it as no
1221 password on export: but it will try both on import. We now do
1222 the same: PKCS12_parse() tries zero length and no password if
1223 the password is set to "" or NULL (NULL is now a valid password:
1224 it wasn't before) as does the pkcs12 application.
1227 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1228 perror when PEM_read_bio_X509_REQ fails, the error message must
1229 be obtained from the error queue.
1232 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1233 it in ERR_remove_state if appropriate, and change ERR_get_state
1234 accordingly to avoid race conditions (this is necessary because
1235 thread_hash is no longer constant once set).
1238 *) Bugfix for linux-elf makefile.one.
1241 *) RSA_get_default_method() will now cause a default
1242 RSA_METHOD to be chosen if one doesn't exist already.
1243 Previously this was only set during a call to RSA_new()
1244 or RSA_new_method(NULL) meaning it was possible for
1245 RSA_get_default_method() to return NULL.
1248 *) Added native name translation to the existing DSO code
1249 that will convert (if the flag to do so is set) filenames
1250 that are sufficiently small and have no path information
1251 into a canonical native form. Eg. "blah" converted to
1252 "libblah.so" or "blah.dll" etc.
1255 *) New function ERR_error_string_n(e, buf, len) which is like
1256 ERR_error_string(e, buf), but writes at most 'len' bytes
1257 including the 0 terminator. For ERR_error_string_n, 'buf'
1259 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1261 *) CONF library reworked to become more general. A new CONF
1262 configuration file reader "class" is implemented as well as a
1263 new functions (NCONF_*, for "New CONF") to handle it. The now
1264 old CONF_* functions are still there, but are reimplemented to
1265 work in terms of the new functions. Also, a set of functions
1266 to handle the internal storage of the configuration data is
1267 provided to make it easier to write new configuration file
1268 reader "classes" (I can definitely see something reading a
1269 configuration file in XML format, for example), called _CONF_*,
1270 or "the configuration storage API"...
1272 The new configuration file reading functions are:
1274 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1275 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1277 NCONF_default, NCONF_WIN32
1279 NCONF_dump_fp, NCONF_dump_bio
1281 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1282 NCONF_new creates a new CONF object. This works in the same way
1283 as other interfaces in OpenSSL, like the BIO interface.
1284 NCONF_dump_* dump the internal storage of the configuration file,
1285 which is useful for debugging. All other functions take the same
1286 arguments as the old CONF_* functions wth the exception of the
1287 first that must be a `CONF *' instead of a `LHASH *'.
1289 To make it easer to use the new classes with the old CONF_* functions,
1290 the function CONF_set_default_method is provided.
1293 *) Add '-tls1' option to 'openssl ciphers', which was already
1294 mentioned in the documentation but had not been implemented.
1295 (This option is not yet really useful because even the additional
1296 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1299 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1300 OpenSSL-based applications) load shared libraries and bind to
1301 them in a portable way.
1302 [Geoff Thorpe, with contributions from Richard Levitte]
1304 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1306 *) Make sure _lrotl and _lrotr are only used with MSVC.
1308 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1309 (the default implementation of RAND_status).
1311 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1312 to '-clrext' (= clear extensions), as intended and documented.
1313 [Bodo Moeller; inconsistency pointed out by Michael Attili
1316 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1317 was larger than the MD block size.
1318 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1320 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1321 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1322 using the passed key: if the passed key was a private key the result
1323 of X509_print(), for example, would be to print out all the private key
1327 *) des_quad_cksum() byte order bug fix.
1328 [Ulf Möller, using the problem description in krb4-0.9.7, where
1329 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1331 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1333 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1335 *) For easily testing in shell scripts whether some command
1336 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1337 returns with exit code 0 iff no command of the given name is available.
1338 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1339 the output goes to stdout and nothing is printed to stderr.
1340 Additional arguments are always ignored.
1342 Since for each cipher there is a command of the same name,
1343 the 'no-cipher' compilation switches can be tested this way.
1345 ('openssl no-XXX' is not able to detect pseudo-commands such
1346 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1349 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1352 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1353 is set; it will be thrown away anyway because each handshake creates
1355 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1356 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1357 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1358 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1361 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1362 'Q' and 'R' lose their special meanings (quit/renegotiate).
1363 This is part of what -quiet does; unlike -quiet, -ign_eof
1364 does not suppress any output.
1367 *) Add compatibility options to the purpose and trust code. The
1368 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1369 accepts a certificate or CA, this was the previous behaviour,
1370 with all the associated security issues.
1372 X509_TRUST_COMPAT is the old trust behaviour: only and
1373 automatically trust self signed roots in certificate store. A
1374 new trust setting X509_TRUST_DEFAULT is used to specify that
1375 a purpose has no associated trust setting and it should instead
1376 use the value in the default purpose.
1379 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1380 and fix a memory leak.
1383 *) In util/mkerr.pl (which implements 'make errors'), preserve
1384 reason strings from the previous version of the .c file, as
1385 the default to have only downcase letters (and digits) in
1386 automatically generated reasons codes is not always appropriate.
1389 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1390 using strerror. Previously, ERR_reason_error_string() returned
1391 library names as reason strings for SYSerr; but SYSerr is a special
1392 case where small numbers are errno values, not library numbers.
1395 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1396 converts DSA parameters into DH parameters. (When creating parameters,
1397 DSA_generate_parameters is used.)
1400 *) Include 'length' (recommended exponent length) in C code generated
1401 by 'openssl dhparam -C'.
1404 *) The second argument to set_label in perlasm was already being used
1405 so couldn't be used as a "file scope" flag. Moved to third argument
1409 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1410 instead of RAND_bytes for encryption IVs and salts.
1413 *) Include RAND_status() into RAND_METHOD instead of implementing
1414 it only for md_rand.c Otherwise replacing the PRNG by calling
1415 RAND_set_rand_method would be impossible.
1418 *) Don't let DSA_generate_key() enter an infinite loop if the random
1419 number generation fails.
1422 *) New 'rand' application for creating pseudo-random output.
1425 *) Added configuration support for Linux/IA64
1426 [Rolf Haberrecker <rolf@suse.de>]
1428 *) Assembler module support for Mingw32.
1431 *) Shared library support for HPUX (in shlib/).
1432 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1434 *) Shared library support for Solaris gcc.
1435 [Lutz Behnke <behnke@trustcenter.de>]
1437 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1439 *) PKCS7_encrypt() was adding text MIME headers twice because they
1440 were added manually and by SMIME_crlf_copy().
1443 *) In bntest.c don't call BN_rand with zero bits argument.
1444 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1446 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1447 case was implemented. This caused BN_div_recp() to fail occasionally.
1450 *) Add an optional second argument to the set_label() in the perl
1451 assembly language builder. If this argument exists and is set
1452 to 1 it signals that the assembler should use a symbol whose
1453 scope is the entire file, not just the current function. This
1454 is needed with MASM which uses the format label:: for this scope.
1455 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1457 *) Change the ASN1 types so they are typedefs by default. Before
1458 almost all types were #define'd to ASN1_STRING which was causing
1459 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1463 *) Change names of new functions to the new get1/get0 naming
1464 convention: After 'get1', the caller owns a reference count
1465 and has to call ..._free; 'get0' returns a pointer to some
1466 data structure without incrementing reference counters.
1467 (Some of the existing 'get' functions increment a reference
1468 counter, some don't.)
1469 Similarly, 'set1' and 'add1' functions increase reference
1470 counters or duplicate objects.
1473 *) Allow for the possibility of temp RSA key generation failure:
1474 the code used to assume it always worked and crashed on failure.
1477 *) Fix potential buffer overrun problem in BIO_printf().
1478 [Ulf Möller, using public domain code by Patrick Powell; problem
1479 pointed out by David Sacerdote <das33@cornell.edu>]
1481 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1482 RAND_egd() and RAND_status(). In the command line application,
1483 the EGD socket can be specified like a seed file using RANDFILE
1487 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1488 Some CAs (e.g. Verisign) distribute certificates in this form.
1491 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1492 list to exclude them. This means that no special compilation option
1493 is needed to use anonymous DH: it just needs to be included in the
1497 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1498 EVP_MD_type. The old functionality is available in a new macro called
1499 EVP_MD_md(). Change code that uses it and update docs.
1502 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1503 where the 'void *' argument is replaced by a function pointer argument.
1504 Previously 'void *' was abused to point to functions, which works on
1505 many platforms, but is not correct. As these functions are usually
1506 called by macros defined in OpenSSL header files, most source code
1507 should work without changes.
1510 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1511 sections with information on -D... compiler switches used for
1512 compiling the library so that applications can see them. To enable
1513 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1514 must be defined. E.g.,
1515 #define OPENSSL_ALGORITHM_DEFINES
1516 #include <openssl/opensslconf.h>
1517 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1518 [Richard Levitte, Ulf and Bodo Möller]
1520 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1524 *) Change the 'other' type in certificate aux info to a STACK_OF
1525 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1526 the required ASN1 format: arbitrary types determined by an OID.
1529 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1530 argument to 'req'. This is not because the function is newer or
1531 better than others it just uses the work 'NEW' in the certificate
1532 request header lines. Some software needs this.
1535 *) Reorganise password command line arguments: now passwords can be
1536 obtained from various sources. Delete the PEM_cb function and make
1537 it the default behaviour: i.e. if the callback is NULL and the
1538 usrdata argument is not NULL interpret it as a null terminated pass
1539 phrase. If usrdata and the callback are NULL then the pass phrase
1540 is prompted for as usual.
1543 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1544 the support is automatically enabled. The resulting binaries will
1545 autodetect the card and use it if present.
1546 [Ben Laurie and Compaq Inc.]
1548 *) Work around for Netscape hang bug. This sends certificate request
1549 and server done in one record. Since this is perfectly legal in the
1550 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1551 the bugs/SSLv3 entry for more info.
1554 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1557 *) Add -rand argument to smime and pkcs12 applications and read/write
1561 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1564 *) Add command line password options to the remaining applications.
1567 *) Bug fix for BN_div_recp() for numerators with an even number of
1571 *) More tests in bntest.c, and changed test_bn output.
1574 *) ./config recognizes MacOS X now.
1577 *) Bug fix for BN_div() when the first words of num and divsor are
1578 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1581 *) Add support for various broken PKCS#8 formats, and command line
1582 options to produce them.
1585 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1586 get temporary BIGNUMs from a BN_CTX.
1589 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1593 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1594 include a #define from the old name to the new. The original intent
1595 was that statically linked binaries could for example just call
1596 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1597 link with digests. This never worked becayse SSLeay_add_all_digests()
1598 and SSLeay_add_all_ciphers() were in the same source file so calling
1599 one would link with the other. They are now in separate source files.
1602 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1605 *) Use a less unusual form of the Miller-Rabin primality test (it used
1606 a binary algorithm for exponentiation integrated into the Miller-Rabin
1607 loop, our standard modexp algorithms are faster).
1610 *) Support for the EBCDIC character set completed.
1611 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1613 *) Source code cleanups: use const where appropriate, eliminate casts,
1614 use void * instead of char * in lhash.
1617 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1618 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1619 this the server could overwrite ephemeral keys that the client
1623 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1624 using 50 iterations of the Rabin-Miller test.
1626 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1627 iterations of the Rabin-Miller test as required by the appendix
1628 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1629 As BN_is_prime_fasttest includes trial division, DSA parameter
1630 generation becomes much faster.
1632 This implies a change for the callback functions in DSA_is_prime
1633 and DSA_generate_parameters: The callback function is called once
1634 for each positive witness in the Rabin-Miller test, not just
1635 occasionally in the inner loop; and the parameters to the
1636 callback function now provide an iteration count for the outer
1637 loop rather than for the current invocation of the inner loop.
1638 DSA_generate_parameters additionally can call the callback
1639 function with an 'iteration count' of -1, meaning that a
1640 candidate has passed the trial division test (when q is generated
1641 from an application-provided seed, trial division is skipped).
1644 *) New function BN_is_prime_fasttest that optionally does trial
1645 division before starting the Rabin-Miller test and has
1646 an additional BN_CTX * argument (whereas BN_is_prime always
1647 has to allocate at least one BN_CTX).
1648 'callback(1, -1, cb_arg)' is called when a number has passed the
1649 trial division stage.
1652 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1656 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1659 *) New function BN_pseudo_rand().
1662 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1663 bignum version of BN_from_montgomery() with the working code from
1664 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1668 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1669 made it impossible to use the same SSL_SESSION data structure in
1670 SSL2 clients in multiple threads.
1673 *) The return value of RAND_load_file() no longer counts bytes obtained
1674 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1675 to seed the PRNG (previously an explicit byte count was required).
1676 [Ulf Möller, Bodo Möller]
1678 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1679 used (char *) instead of (void *) and had casts all over the place.
1682 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1685 *) Retain source code compatibility for BN_prime_checks macro:
1686 BN_is_prime(..., BN_prime_checks, ...) now uses
1687 BN_prime_checks_for_size to determine the appropriate number of
1688 Rabin-Miller iterations.
1691 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1692 DH_CHECK_P_NOT_SAFE_PRIME.
1693 (Check if this is true? OpenPGP calls them "strong".)
1696 *) Merge the functionality of "dh" and "gendh" programs into a new program
1697 "dhparam". The old programs are retained for now but will handle DH keys
1698 (instead of parameters) in future.
1701 *) Make the ciphers, s_server and s_client programs check the return values
1702 when a new cipher list is set.
1705 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1706 ciphers. Before when the 56bit ciphers were enabled the sorting was
1709 The syntax for the cipher sorting has been extended to support sorting by
1710 cipher-strength (using the strength_bits hard coded in the tables).
1711 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1713 Fix a bug in the cipher-command parser: when supplying a cipher command
1714 string with an "undefined" symbol (neither command nor alphanumeric
1715 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1716 an error is flagged.
1718 Due to the strength-sorting extension, the code of the
1719 ssl_create_cipher_list() function was completely rearranged. I hope that
1720 the readability was also increased :-)
1721 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1723 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1724 for the first serial number and places 2 in the serial number file. This
1725 avoids problems when the root CA is created with serial number zero and
1726 the first user certificate has the same issuer name and serial number
1730 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1731 the new code. Add documentation for this stuff.
1734 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1735 X509_*() to X509at_*() on the grounds that they don't handle X509
1736 structures and behave in an analagous way to the X509v3 functions:
1737 they shouldn't be called directly but wrapper functions should be used
1740 So we also now have some wrapper functions that call the X509at functions
1741 when passed certificate requests. (TO DO: similar things can be done with
1742 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1743 things. Some of these need some d2i or i2d and print functionality
1744 because they handle more complex structures.)
1747 *) Add missing #ifndefs that caused missing symbols when building libssl
1748 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1749 NO_RSA in ssl/s2*.c.
1750 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1752 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1753 has a return value which indicates the quality of the random data
1754 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1755 error queue. New function RAND_pseudo_bytes() generates output that is
1756 guaranteed to be unique but not unpredictable. RAND_add is like
1757 RAND_seed, but takes an extra argument for an entropy estimate
1758 (RAND_seed always assumes full entropy).
1761 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1762 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1763 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1764 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1765 false-positive rate of at most 2^-80 for random input.
1768 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1771 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1772 in the 0.9.5 release), this returns the chain
1773 from an X509_CTX structure with a dup of the stack and all
1774 the X509 reference counts upped: so the stack will exist
1775 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1778 Also make SSL_SESSION_print() print out the verify return
1782 *) Add manpage for the pkcs12 command. Also change the default
1783 behaviour so MAC iteration counts are used unless the new
1784 -nomaciter option is used. This improves file security and
1785 only older versions of MSIE (4.0 for example) need it.
1788 *) Honor the no-xxx Configure options when creating .DEF files.
1791 *) Add PKCS#10 attributes to field table: challengePassword,
1792 unstructuredName and unstructuredAddress. These are taken from
1793 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1794 international characters are used.
1796 More changes to X509_ATTRIBUTE code: allow the setting of types
1797 based on strings. Remove the 'loc' parameter when adding
1798 attributes because these will be a SET OF encoding which is sorted
1802 *) Initial changes to the 'req' utility to allow request generation
1803 automation. This will allow an application to just generate a template
1804 file containing all the field values and have req construct the
1807 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1808 used all over the place including certificate requests and PKCS#7
1809 structures. They are currently handled manually where necessary with
1810 some primitive wrappers for PKCS#7. The new functions behave in a
1811 manner analogous to the X509 extension functions: they allow
1812 attributes to be looked up by NID and added.
1814 Later something similar to the X509V3 code would be desirable to
1815 automatically handle the encoding, decoding and printing of the
1816 more complex types. The string types like challengePassword can
1817 be handled by the string table functions.
1819 Also modified the multi byte string table handling. Now there is
1820 a 'global mask' which masks out certain types. The table itself
1821 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1822 is useful when for example there is only one permissible type
1823 (as in countryName) and using the mask might result in no valid
1827 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1828 SSL_get_peer_finished to allow applications to obtain the latest
1829 Finished messages sent to the peer or expected from the peer,
1830 respectively. (SSL_get_peer_finished is usually the Finished message
1831 actually received from the peer, otherwise the protocol will be aborted.)
1833 As the Finished message are message digests of the complete handshake
1834 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1835 be used for external authentication procedures when the authentication
1836 provided by SSL/TLS is not desired or is not enough.
1839 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1840 the host supports BWX extension and if Compaq C is present on the
1841 $PATH. Just exploiting of the BWX extension results in 20-30%
1842 performance kick for some algorithms, e.g. DES and RC4 to mention
1843 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1847 *) Add support for MS "fast SGC". This is arguably a violation of the
1848 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1849 weak crypto and after checking the certificate is SGC a second one
1850 with strong crypto. MS SGC stops the first handshake after receiving
1851 the server certificate message and sends a second client hello. Since
1852 a server will typically do all the time consuming operations before
1853 expecting any further messages from the client (server key exchange
1854 is the most expensive) there is little difference between the two.
1856 To get OpenSSL to support MS SGC we have to permit a second client
1857 hello message after we have sent server done. In addition we have to
1858 reset the MAC if we do get this second client hello.
1861 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1862 if a DER encoded private key is RSA or DSA traditional format. Changed
1863 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1864 format DER encoded private key. Newer code should use PKCS#8 format which
1865 has the key type encoded in the ASN1 structure. Added DER private key
1866 support to pkcs8 application.
1869 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1870 ciphersuites has been selected (as required by the SSL 3/TLS 1
1871 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1872 is set, we interpret this as a request to violate the specification
1873 (the worst that can happen is a handshake failure, and 'correct'
1874 behaviour would result in a handshake failure anyway).
1877 *) In SSL_CTX_add_session, take into account that there might be multiple
1878 SSL_SESSION structures with the same session ID (e.g. when two threads
1879 concurrently obtain them from an external cache).
1880 The internal cache can handle only one SSL_SESSION with a given ID,
1881 so if there's a conflict, we now throw out the old one to achieve
1885 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1886 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1887 some routines that use cipher OIDs: some ciphers do not have OIDs
1888 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1892 *) Simplify the trust setting structure and code. Now we just have
1893 two sequences of OIDs for trusted and rejected settings. These will
1894 typically have values the same as the extended key usage extension
1895 and any application specific purposes.
1897 The trust checking code now has a default behaviour: it will just
1898 check for an object with the same NID as the passed id. Functions can
1899 be provided to override either the default behaviour or the behaviour
1900 for a given id. SSL client, server and email already have functions
1901 in place for compatibility: they check the NID and also return "trusted"
1902 if the certificate is self signed.
1905 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1906 traditional format into an EVP_PKEY structure.
1909 *) Add a password callback function PEM_cb() which either prompts for
1910 a password if usr_data is NULL or otherwise assumes it is a null
1911 terminated password. Allow passwords to be passed on command line
1912 environment or config files in a few more utilities.
1915 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1916 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1917 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1918 Update documentation.
1921 *) Support for ASN1 "NULL" type. This could be handled before by using
1922 ASN1_TYPE but there wasn't any function that would try to read a NULL
1923 and produce an error if it couldn't. For compatibility we also have
1924 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1925 don't allocate anything because they don't need to.
1928 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1930 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1932 *) Rebuild of the memory allocation routines used by OpenSSL code and
1933 possibly others as well. The purpose is to make an interface that
1934 provide hooks so anyone can build a separate set of allocation and
1935 deallocation routines to be used by OpenSSL, for example memory
1936 pool implementations, or something else, which was previously hard
1937 since Malloc(), Realloc() and Free() were defined as macros having
1938 the values malloc, realloc and free, respectively (except for Win32
1939 compilations). The same is provided for memory debugging code.
1940 OpenSSL already comes with functionality to find memory leaks, but
1941 this gives people a chance to debug other memory problems.
1943 With these changes, a new set of functions and macros have appeared:
1945 CRYPTO_set_mem_debug_functions() [F]
1946 CRYPTO_get_mem_debug_functions() [F]
1947 CRYPTO_dbg_set_options() [F]
1948 CRYPTO_dbg_get_options() [F]
1949 CRYPTO_malloc_debug_init() [M]
1951 The memory debug functions are NULL by default, unless the library
1952 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1953 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1954 gives the standard debugging functions that come with OpenSSL) or
1955 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1956 provided by the library user) must be used. When the standard
1957 debugging functions are used, CRYPTO_dbg_set_options can be used to
1958 request additional information:
1959 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
1960 the CRYPTO_MDEBUG_xxx macro when compiling the library.
1962 Also, things like CRYPTO_set_mem_functions will always give the
1963 expected result (the new set of functions is used for allocation
1964 and deallocation) at all times, regardless of platform and compiler
1967 To finish it up, some functions that were never use in any other
1968 way than through macros have a new API and new semantic:
1971 CRYPTO_dbg_realloc()
1974 All macros of value have retained their old syntax.
1975 [Richard Levitte and Bodo Moeller]
1977 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
1978 ordering of SMIMECapabilities wasn't in "strength order" and there
1979 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
1983 *) Some ASN1 types with illegal zero length encoding (INTEGER,
1984 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
1985 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
1987 *) Merge in my S/MIME library for OpenSSL. This provides a simple
1988 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
1989 functionality to handle multipart/signed properly) and a utility
1990 called 'smime' to call all this stuff. This is based on code I
1991 originally wrote for Celo who have kindly allowed it to be
1992 included in OpenSSL.
1995 *) Add variants des_set_key_checked and des_set_key_unchecked of
1996 des_set_key (aka des_key_sched). Global variable des_check_key
1997 decides which of these is called by des_set_key; this way
1998 des_check_key behaves as it always did, but applications and
1999 the library itself, which was buggy for des_check_key == 1,
2000 have a cleaner way to pick the version they need.
2003 *) New function PKCS12_newpass() which changes the password of a
2007 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2008 dynamic mix. In both cases the ids can be used as an index into the
2009 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2010 functions so they accept a list of the field values and the
2011 application doesn't need to directly manipulate the X509_TRUST
2015 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2019 *) Modify the way the V3 extension code looks up extensions. This now
2020 works in a similar way to the object code: we have some "standard"
2021 extensions in a static table which is searched with OBJ_bsearch()
2022 and the application can add dynamic ones if needed. The file
2023 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2024 updated whenever a new extension is added to the core code and kept
2025 in ext_nid order. There is a simple program 'tabtest.c' which checks
2026 this. New extensions are not added too often so this file can readily
2027 be maintained manually.
2029 There are two big advantages in doing things this way. The extensions
2030 can be looked up immediately and no longer need to be "added" using
2031 X509V3_add_standard_extensions(): this function now does nothing.
2032 [Side note: I get *lots* of email saying the extension code doesn't
2033 work because people forget to call this function]
2034 Also no dynamic allocation is done unless new extensions are added:
2035 so if we don't add custom extensions there is no need to call
2036 X509V3_EXT_cleanup().
2039 *) Modify enc utility's salting as follows: make salting the default. Add a
2040 magic header, so unsalted files fail gracefully instead of just decrypting
2041 to garbage. This is because not salting is a big security hole, so people
2042 should be discouraged from doing it.
2045 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2046 digest to be passed on the command line but it only used this
2047 parameter when signing a certificate. Modified so all relevant
2048 operations are affected by the digest parameter including the
2049 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2050 DSA key was used because it didn't fix the digest.
2053 *) Initial certificate chain verify code. Currently tests the untrusted
2054 certificates for consistency with the verify purpose (which is set
2055 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2057 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2058 this is because it will reject chains with invalid extensions whereas
2059 every previous version of OpenSSL and SSLeay made no checks at all.
2061 Trust code: checks the root CA for the relevant trust settings. Trust
2062 settings have an initial value consistent with the verify purpose: e.g.
2063 if the verify purpose is for SSL client use it expects the CA to be
2064 trusted for SSL client use. However the default value can be changed to
2065 permit custom trust settings: one example of this would be to only trust
2066 certificates from a specific "secure" set of CAs.
2068 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2069 which should be used for version portability: especially since the
2070 verify structure is likely to change more often now.
2072 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2073 to set them. If not set then assume SSL clients will verify SSL servers
2076 Two new options to the verify program: -untrusted allows a set of
2077 untrusted certificates to be passed in and -purpose which sets the
2078 intended purpose of the certificate. If a purpose is set then the
2079 new chain verify code is used to check extension consistency.
2082 *) Support for the authority information access extension.
2085 *) Modify RSA and DSA PEM read routines to transparently handle
2086 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2087 public keys in a format compatible with certificate
2088 SubjectPublicKeyInfo structures. Unfortunately there were already
2089 functions called *_PublicKey_* which used various odd formats so
2090 these are retained for compatibility: however the DSA variants were
2091 never in a public release so they have been deleted. Changed dsa/rsa
2092 utilities to handle the new format: note no releases ever handled public
2093 keys so we should be OK.
2095 The primary motivation for this change is to avoid the same fiasco
2096 that dogs private keys: there are several incompatible private key
2097 formats some of which are standard and some OpenSSL specific and
2098 require various evil hacks to allow partial transparent handling and
2099 even then it doesn't work with DER formats. Given the option anything
2100 other than PKCS#8 should be dumped: but the other formats have to
2101 stay in the name of compatibility.
2103 With public keys and the benefit of hindsight one standard format
2104 is used which works with EVP_PKEY, RSA or DSA structures: though
2105 it clearly returns an error if you try to read the wrong kind of key.
2107 Added a -pubkey option to the 'x509' utility to output the public key.
2108 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2109 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2110 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2111 that do the same as the EVP_PKEY_assign_*() except they up the
2112 reference count of the added key (they don't "swallow" the
2116 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2117 CRLs would fail if the file contained no certificates or no CRLs:
2118 added a new function to read in both types and return the number
2119 read: this means that if none are read it will be an error. The
2120 DER versions of the certificate and CRL reader would always fail
2121 because it isn't possible to mix certificates and CRLs in DER format
2122 without choking one or the other routine. Changed this to just read
2123 a certificate: this is the best we can do. Also modified the code
2124 in apps/verify.c to take notice of return codes: it was previously
2125 attempting to read in certificates from NULL pointers and ignoring
2126 any errors: this is one reason why the cert and CRL reader seemed
2127 to work. It doesn't check return codes from the default certificate
2128 routines: these may well fail if the certificates aren't installed.
2131 *) Code to support otherName option in GeneralName.
2134 *) First update to verify code. Change the verify utility
2135 so it warns if it is passed a self signed certificate:
2136 for consistency with the normal behaviour. X509_verify
2137 has been modified to it will now verify a self signed
2138 certificate if *exactly* the same certificate appears
2139 in the store: it was previously impossible to trust a
2140 single self signed certificate. This means that:
2141 openssl verify ss.pem
2142 now gives a warning about a self signed certificate but
2143 openssl verify -CAfile ss.pem ss.pem
2147 *) For servers, store verify_result in SSL_SESSION data structure
2148 (and add it to external session representation).
2149 This is needed when client certificate verifications fails,
2150 but an application-provided verification callback (set by
2151 SSL_CTX_set_cert_verify_callback) allows accepting the session
2152 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2153 but returns 1): When the session is reused, we have to set
2154 ssl->verify_result to the appropriate error code to avoid
2156 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2158 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2159 case in PKCS7_dataInit() where the signed PKCS7 structure
2160 didn't contain any existing data because it was being created.
2161 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2163 *) Add a salt to the key derivation routines in enc.c. This
2164 forms the first 8 bytes of the encrypted file. Also add a
2165 -S option to allow a salt to be input on the command line.
2168 *) New function X509_cmp(). Oddly enough there wasn't a function
2169 to compare two certificates. We do this by working out the SHA1
2170 hash and comparing that. X509_cmp() will be needed by the trust
2174 *) SSL_get1_session() is like SSL_get_session(), but increments
2175 the reference count in the SSL_SESSION returned.
2176 [Geoff Thorpe <geoff@eu.c2.net>]
2178 *) Fix for 'req': it was adding a null to request attributes.
2179 Also change the X509_LOOKUP and X509_INFO code to handle
2180 certificate auxiliary information.
2183 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2187 *) Add the possibility to add extra information to the memory leak
2188 detecting output, to form tracebacks, showing from where each
2189 allocation was originated: CRYPTO_push_info("constant string") adds
2190 the string plus current file name and line number to a per-thread
2191 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2192 is like calling CYRPTO_pop_info() until the stack is empty.
2193 Also updated memory leak detection code to be multi-thread-safe.
2196 *) Add options -text and -noout to pkcs7 utility and delete the
2197 encryption options which never did anything. Update docs.
2200 *) Add options to some of the utilities to allow the pass phrase
2201 to be included on either the command line (not recommended on
2202 OSes like Unix) or read from the environment. Update the
2203 manpages and fix a few bugs.
2206 *) Add a few manpages for some of the openssl commands.
2209 *) Fix the -revoke option in ca. It was freeing up memory twice,
2210 leaking and not finding already revoked certificates.
2213 *) Extensive changes to support certificate auxiliary information.
2214 This involves the use of X509_CERT_AUX structure and X509_AUX
2215 functions. An X509_AUX function such as PEM_read_X509_AUX()
2216 can still read in a certificate file in the usual way but it
2217 will also read in any additional "auxiliary information". By
2218 doing things this way a fair degree of compatibility can be
2219 retained: existing certificates can have this information added
2220 using the new 'x509' options.
2222 Current auxiliary information includes an "alias" and some trust
2223 settings. The trust settings will ultimately be used in enhanced
2224 certificate chain verification routines: currently a certificate
2225 can only be trusted if it is self signed and then it is trusted
2229 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2230 The problem was that one of the replacement routines had not been working
2231 since SSLeay releases. For now the offending routine has been replaced
2232 with non-optimised assembler. Even so, this now gives around 95%
2233 performance improvement for 1024 bit RSA signs.
2236 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2237 handling. Most clients have the effective key size in bits equal to
2238 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2239 A few however don't do this and instead use the size of the decrypted key
2240 to determine the RC2 key length and the AlgorithmIdentifier to determine
2241 the effective key length. In this case the effective key length can still
2242 be 40 bits but the key length can be 168 bits for example. This is fixed
2243 by manually forcing an RC2 key into the EVP_PKEY structure because the
2244 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2245 the key length and effective key length are equal.
2248 *) Add a bunch of functions that should simplify the creation of
2249 X509_NAME structures. Now you should be able to do:
2250 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2251 and have it automatically work out the correct field type and fill in
2252 the structures. The more adventurous can try:
2253 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2254 and it will (hopefully) work out the correct multibyte encoding.
2257 *) Change the 'req' utility to use the new field handling and multibyte
2258 copy routines. Before the DN field creation was handled in an ad hoc
2259 way in req, ca, and x509 which was rather broken and didn't support
2260 BMPStrings or UTF8Strings. Since some software doesn't implement
2261 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2262 using the dirstring_type option. See the new comment in the default
2263 openssl.cnf for more info.
2266 *) Make crypto/rand/md_rand.c more robust:
2267 - Assure unique random numbers after fork().
2268 - Make sure that concurrent threads access the global counter and
2269 md serializably so that we never lose entropy in them
2270 or use exactly the same state in multiple threads.
2271 Access to the large state is not always serializable because
2272 the additional locking could be a performance killer, and
2273 md should be large enough anyway.
2276 *) New file apps/app_rand.c with commonly needed functionality
2277 for handling the random seed file.
2279 Use the random seed file in some applications that previously did not:
2281 dsaparam -genkey (which also ignored its '-rand' option),
2284 x509 (when signing).
2285 Except on systems with /dev/urandom, it is crucial to have a random
2286 seed file at least for key creation, DSA signing, and for DH exchanges;
2287 for RSA signatures we could do without one.
2289 gendh and gendsa (unlike genrsa) used to read only the first byte
2290 of each file listed in the '-rand' option. The function as previously
2291 found in genrsa is now in app_rand.c and is used by all programs
2292 that support '-rand'.
2295 *) In RAND_write_file, use mode 0600 for creating files;
2296 don't just chmod when it may be too late.
2299 *) Report an error from X509_STORE_load_locations
2300 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2303 *) New function ASN1_mbstring_copy() this copies a string in either
2304 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2305 into an ASN1_STRING type. A mask of permissible types is passed
2306 and it chooses the "minimal" type to use or an error if not type
2310 *) Add function equivalents to the various macros in asn1.h. The old
2311 macros are retained with an M_ prefix. Code inside the library can
2312 use the M_ macros. External code (including the openssl utility)
2313 should *NOT* in order to be "shared library friendly".
2316 *) Add various functions that can check a certificate's extensions
2317 to see if it usable for various purposes such as SSL client,
2318 server or S/MIME and CAs of these types. This is currently
2319 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2320 verification. Also added a -purpose flag to x509 utility to
2321 print out all the purposes.
2324 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2328 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2329 for, obtain and decode and extension and obtain its critical flag.
2330 This allows all the necessary extension code to be handled in a
2331 single function call.
2334 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2335 platforms. See crypto/rc4/rc4_enc.c for further details.
2338 *) New -noout option to asn1parse. This causes no output to be produced
2339 its main use is when combined with -strparse and -out to extract data
2340 from a file (which may not be in ASN.1 format).
2343 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2344 when producing the local key id.
2345 [Richard Levitte <levitte@stacken.kth.se>]
2347 *) New option -dhparam in s_server. This allows a DH parameter file to be
2348 stated explicitly. If it is not stated then it tries the first server
2349 certificate file. The previous behaviour hard coded the filename
2353 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2354 a public key to be input or output. For example:
2355 openssl rsa -in key.pem -pubout -out pubkey.pem
2356 Also added necessary DSA public key functions to handle this.
2359 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2360 in the message. This was handled by allowing
2361 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2362 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2364 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2365 to the end of the strings whereas this didn't. This would cause problems
2366 if strings read with d2i_ASN1_bytes() were later modified.
2367 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2369 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2370 data and it contains EOF it will end up returning an error. This is
2371 caused by input 46 bytes long. The cause is due to the way base64
2372 BIOs find the start of base64 encoded data. They do this by trying a
2373 trial decode on each line until they find one that works. When they
2374 do a flag is set and it starts again knowing it can pass all the
2375 data directly through the decoder. Unfortunately it doesn't reset
2376 the context it uses. This means that if EOF is reached an attempt
2377 is made to pass two EOFs through the context and this causes the
2378 resulting error. This can also cause other problems as well. As is
2379 usual with these problems it takes *ages* to find and the fix is
2380 trivial: move one line.
2381 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2383 *) Ugly workaround to get s_client and s_server working under Windows. The
2384 old code wouldn't work because it needed to select() on sockets and the
2385 tty (for keypresses and to see if data could be written). Win32 only
2386 supports select() on sockets so we select() with a 1s timeout on the
2387 sockets and then see if any characters are waiting to be read, if none
2388 are present then we retry, we also assume we can always write data to
2389 the tty. This isn't nice because the code then blocks until we've
2390 received a complete line of data and it is effectively polling the
2391 keyboard at 1s intervals: however it's quite a bit better than not
2392 working at all :-) A dedicated Windows application might handle this
2393 with an event loop for example.
2396 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2397 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2398 will be called when RSA_sign() and RSA_verify() are used. This is useful
2399 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2400 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2401 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2402 This necessitated the support of an extra signature type NID_md5_sha1
2403 for SSL signatures and modifications to the SSL library to use it instead
2404 of calling RSA_public_decrypt() and RSA_private_encrypt().
2407 *) Add new -verify -CAfile and -CApath options to the crl program, these
2408 will lookup a CRL issuers certificate and verify the signature in a
2409 similar way to the verify program. Tidy up the crl program so it
2410 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2411 less strict. It will now permit CRL extensions even if it is not
2412 a V2 CRL: this will allow it to tolerate some broken CRLs.
2415 *) Initialize all non-automatic variables each time one of the openssl
2416 sub-programs is started (this is necessary as they may be started
2417 multiple times from the "OpenSSL>" prompt).
2418 [Lennart Bang, Bodo Moeller]
2420 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2421 removing all other RSA functionality (this is what NO_RSA does). This
2422 is so (for example) those in the US can disable those operations covered
2423 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2427 *) Non-copying interface to BIO pairs.
2428 (still largely untested)
2431 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2432 ASCII string. This was handled independently in various places before.
2435 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2436 UTF8 strings a character at a time.
2439 *) Use client_version from client hello to select the protocol
2440 (s23_srvr.c) and for RSA client key exchange verification
2441 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2444 *) Add various utility functions to handle SPKACs, these were previously
2445 handled by poking round in the structure internals. Added new function
2446 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2447 print, verify and generate SPKACs. Based on an original idea from
2448 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2451 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2454 *) Allow the config file extension section to be overwritten on the
2455 command line. Based on an original idea from Massimiliano Pala
2456 <madwolf@comune.modena.it>. The new option is called -extensions
2457 and can be applied to ca, req and x509. Also -reqexts to override
2458 the request extensions in req and -crlexts to override the crl extensions
2462 *) Add new feature to the SPKAC handling in ca. Now you can include
2463 the same field multiple times by preceding it by "XXXX." for example:
2466 this is the same syntax as used in the req config file.
2469 *) Allow certificate extensions to be added to certificate requests. These
2470 are specified in a 'req_extensions' option of the req section of the
2471 config file. They can be printed out with the -text option to req but
2472 are otherwise ignored at present.
2475 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2476 data read consists of only the final block it would not decrypted because
2477 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2478 A misplaced 'break' also meant the decrypted final block might not be
2479 copied until the next read.
2482 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2483 a few extra parameters to the DH structure: these will be useful if
2484 for example we want the value of 'q' or implement X9.42 DH.
2487 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2488 provides hooks that allow the default DSA functions or functions on a
2489 "per key" basis to be replaced. This allows hardware acceleration and
2490 hardware key storage to be handled without major modification to the
2491 library. Also added low level modexp hooks and CRYPTO_EX structure and
2492 associated functions.
2495 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2496 as "read only": it can't be written to and the buffer it points to will
2497 not be freed. Reading from a read only BIO is much more efficient than
2498 a normal memory BIO. This was added because there are several times when
2499 an area of memory needs to be read from a BIO. The previous method was
2500 to create a memory BIO and write the data to it, this results in two
2501 copies of the data and an O(n^2) reading algorithm. There is a new
2502 function BIO_new_mem_buf() which creates a read only memory BIO from
2503 an area of memory. Also modified the PKCS#7 routines to use read only
2507 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2508 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2509 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2510 but a retry condition occured while trying to read the rest.
2513 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2514 NID_pkcs7_encrypted by default: this was wrong since this should almost
2515 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2516 the encrypted data type: this is a more sensible place to put it and it
2517 allows the PKCS#12 code to be tidied up that duplicated this
2521 *) Changed obj_dat.pl script so it takes its input and output files on
2522 the command line. This should avoid shell escape redirection problems
2526 *) Initial support for certificate extension requests, these are included
2527 in things like Xenroll certificate requests. Included functions to allow
2528 extensions to be obtained and added.
2531 *) -crlf option to s_client and s_server for sending newlines as
2532 CRLF (as required by many protocols).
2535 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2537 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2538 [Ralf S. Engelschall]
2540 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2541 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2543 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2547 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2548 DH parameters/keys (q is lost during that conversion, but the resulting
2549 DH parameters contain its length).
2551 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2552 much faster than DH_generate_parameters (which creates parameters
2553 where p = 2*q + 1), and also the smaller q makes DH computations
2554 much more efficient (160-bit exponentiation instead of 1024-bit
2555 exponentiation); so this provides a convenient way to support DHE
2556 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2557 utter importance to use
2558 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2560 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2561 when such DH parameters are used, because otherwise small subgroup
2562 attacks may become possible!
2565 *) Avoid memory leak in i2d_DHparams.
2568 *) Allow the -k option to be used more than once in the enc program:
2569 this allows the same encrypted message to be read by multiple recipients.
2572 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2573 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2574 it will always use the numerical form of the OID, even if it has a short
2578 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2579 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2580 otherwise bn_mod_exp was called. In the case of hardware keys for example
2581 no private key components need be present and it might store extra data
2582 in the RSA structure, which cannot be accessed from bn_mod_exp.
2583 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2584 private key operations.
2587 *) Added support for SPARC Linux.
2590 *) pem_password_cb function type incompatibly changed from
2591 typedef int pem_password_cb(char *buf, int size, int rwflag);
2593 ....(char *buf, int size, int rwflag, void *userdata);
2594 so that applications can pass data to their callbacks:
2595 The PEM[_ASN1]_{read,write}... functions and macros now take an
2596 additional void * argument, which is just handed through whenever
2597 the password callback is called.
2598 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2600 New function SSL_CTX_set_default_passwd_cb_userdata.
2602 Compatibility note: As many C implementations push function arguments
2603 onto the stack in reverse order, the new library version is likely to
2604 interoperate with programs that have been compiled with the old
2605 pem_password_cb definition (PEM_whatever takes some data that
2606 happens to be on the stack as its last argument, and the callback
2607 just ignores this garbage); but there is no guarantee whatsoever that
2610 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2611 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2612 problems not only on Windows, but also on some Unix platforms.
2613 To avoid problematic command lines, these definitions are now in an
2614 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2615 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2618 *) MIPS III/IV assembler module is reimplemented.
2621 *) More DES library cleanups: remove references to srand/rand and
2622 delete an unused file.
2625 *) Add support for the the free Netwide assembler (NASM) under Win32,
2626 since not many people have MASM (ml) and it can be hard to obtain.
2627 This is currently experimental but it seems to work OK and pass all
2628 the tests. Check out INSTALL.W32 for info.
2631 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2632 without temporary keys kept an extra copy of the server key,
2633 and connections with temporary keys did not free everything in case
2637 *) New function RSA_check_key and new openssl rsa option -check
2638 for verifying the consistency of RSA keys.
2639 [Ulf Moeller, Bodo Moeller]
2641 *) Various changes to make Win32 compile work:
2642 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2643 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2644 comparison" warnings.
2645 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2648 *) Add a debugging option to PKCS#5 v2 key generation function: when
2649 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2650 derived keys are printed to stderr.
2653 *) Copy the flags in ASN1_STRING_dup().
2654 [Roman E. Pavlov <pre@mo.msk.ru>]
2656 *) The x509 application mishandled signing requests containing DSA
2657 keys when the signing key was also DSA and the parameters didn't match.
2659 It was supposed to omit the parameters when they matched the signing key:
2660 the verifying software was then supposed to automatically use the CA's
2661 parameters if they were absent from the end user certificate.
2663 Omitting parameters is no longer recommended. The test was also
2664 the wrong way round! This was probably due to unusual behaviour in
2665 EVP_cmp_parameters() which returns 1 if the parameters match.
2666 This meant that parameters were omitted when they *didn't* match and
2667 the certificate was useless. Certificates signed with 'ca' didn't have
2669 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2671 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2672 The interface is as follows:
2673 Applications can use
2674 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2675 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2676 "off" is now the default.
2677 The library internally uses
2678 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2679 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2680 to disable memory-checking temporarily.
2682 Some inconsistent states that previously were possible (and were
2683 even the default) are now avoided.
2685 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2686 with each memory chunk allocated; this is occasionally more helpful
2687 than just having a counter.
2689 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2691 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2695 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2696 which largely parallels "options", but is for changing API behaviour,
2697 whereas "options" are about protocol behaviour.
2698 Initial "mode" flags are:
2700 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2701 a single record has been written.
2702 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2703 retries use the same buffer location.
2704 (But all of the contents must be
2708 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
2711 *) Fix problems with no-hmac etc.
2712 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2714 *) New functions RSA_get_default_method(), RSA_set_method() and
2715 RSA_get_method(). These allows replacement of RSA_METHODs without having
2716 to mess around with the internals of an RSA structure.
2719 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2720 Also really enable memory leak checks in openssl.c and in some
2722 [Chad C. Mulligan, Bodo Moeller]
2724 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2725 up the length of negative integers. This has now been simplified to just
2726 store the length when it is first determined and use it later, rather
2727 than trying to keep track of where data is copied and updating it to
2729 [Steve Henson, reported by Brien Wheeler
2730 <bwheeler@authentica-security.com>]
2732 *) Add a new function PKCS7_signatureVerify. This allows the verification
2733 of a PKCS#7 signature but with the signing certificate passed to the
2734 function itself. This contrasts with PKCS7_dataVerify which assumes the
2735 certificate is present in the PKCS#7 structure. This isn't always the
2736 case: certificates can be omitted from a PKCS#7 structure and be
2737 distributed by "out of band" means (such as a certificate database).
2740 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2741 function prototypes in pem.h, also change util/mkdef.pl to add the
2742 necessary function names.
2745 *) mk1mf.pl (used by Windows builds) did not properly read the
2746 options set by Configure in the top level Makefile, and Configure
2747 was not even able to write more than one option correctly.
2748 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2751 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2752 file to be loaded from a BIO or FILE pointer. The BIO version will
2753 for example allow memory BIOs to contain config info.
2756 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2757 Whoever hopes to achieve shared-library compatibility across versions
2758 must use this, not the compile-time macro.
2759 (Exercise 0.9.4: Which is the minimum library version required by
2761 Note: All this applies only to multi-threaded programs, others don't
2765 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2766 through a BIO pair triggered the default case, i.e.
2767 SSLerr(...,SSL_R_UNKNOWN_STATE).
2770 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2771 can use the SSL library even if none of the specific BIOs is
2775 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2776 for the encoded length.
2777 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2779 *) Add initial documentation of the X509V3 functions.
2782 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2783 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2784 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2785 secure PKCS#8 private key format with a high iteration count.
2788 *) Fix determination of Perl interpreter: A perl or perl5
2789 _directory_ in $PATH was also accepted as the interpreter.
2790 [Ralf S. Engelschall]
2792 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2793 wrong with it but it was very old and did things like calling
2794 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2798 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2799 to use the new extension code.
2802 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2803 with macros. This should make it easier to change their form, add extra
2804 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2808 *) Add to configuration table a new entry that can specify an alternative
2809 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2810 according to Mark Crispin <MRC@Panda.COM>.
2814 *) DES CBC did not update the IV. Weird.
2817 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2818 Changing the behaviour of the former might break existing programs --
2819 where IV updating is needed, des_ncbc_encrypt can be used.
2822 *) When bntest is run from "make test" it drives bc to check its
2823 calculations, as well as internally checking them. If an internal check
2824 fails, it needs to cause bc to give a non-zero result or make test carries
2825 on without noticing the failure. Fixed.
2828 *) DES library cleanups.
2831 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2832 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2833 ciphers. NOTE: although the key derivation function has been verified
2834 against some published test vectors it has not been extensively tested
2835 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2839 *) Instead of "mkdir -p", which is not fully portable, use new
2840 Perl script "util/mkdir-p.pl".
2843 *) Rewrite the way password based encryption (PBE) is handled. It used to
2844 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2845 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2846 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2847 the 'parameter' field of the AlgorithmIdentifier is passed to the
2848 underlying key generation function so it must do its own ASN1 parsing.
2849 This has also changed the EVP_PBE_CipherInit() function which now has a
2850 'parameter' argument instead of literal salt and iteration count values
2851 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2854 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2855 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2856 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2857 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2858 value was just used as a "magic string" and not used directly its
2859 value doesn't matter.
2862 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2866 *) "linux-sparc64" configuration (ultrapenguin).
2867 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2868 "linux-sparc" configuration.
2869 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2871 *) config now generates no-xxx options for missing ciphers.
2874 *) Support the EBCDIC character set (work in progress).
2875 File ebcdic.c not yet included because it has a different license.
2876 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2878 *) Support BS2000/OSD-POSIX.
2879 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2881 *) Make callbacks for key generation use void * instead of char *.
2884 *) Make S/MIME samples compile (not yet tested).
2887 *) Additional typesafe stacks.
2890 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2894 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2896 *) New configuration variant "sco5-gcc".
2898 *) Updated some demos.
2899 [Sean O Riordain, Wade Scholine]
2901 *) Add missing BIO_free at exit of pkcs12 application.
2904 *) Fix memory leak in conf.c.
2907 *) Updates for Win32 to assembler version of MD5.
2910 *) Set #! path to perl in apps/der_chop to where we found it
2911 instead of using a fixed path.
2914 *) SHA library changes for irix64-mips4-cc.
2917 *) Improvements for VMS support.
2921 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2923 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2924 This also avoids the problems with SC4.2 and unpatched SC5.
2925 [Andy Polyakov <appro@fy.chalmers.se>]
2927 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2928 These are required because of the typesafe stack would otherwise break
2929 existing code. If old code used a structure member which used to be STACK
2930 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2931 sk_num or sk_value it would produce an error because the num, data members
2932 are not present in STACK_OF. Now it just produces a warning. sk_set
2933 replaces the old method of assigning a value to sk_value
2934 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2935 that does this will no longer work (and should use sk_set instead) but
2936 this could be regarded as a "questionable" behaviour anyway.
2939 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2940 correctly handle encrypted S/MIME data.
2943 *) Change type of various DES function arguments from des_cblock
2944 (which means, in function argument declarations, pointer to char)
2945 to des_cblock * (meaning pointer to array with 8 char elements),
2946 which allows the compiler to do more typechecking; it was like
2947 that back in SSLeay, but with lots of ugly casts.
2949 Introduce new type const_des_cblock.
2952 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2953 problems: find RecipientInfo structure that matches recipient certificate
2954 and initialise the ASN1 structures properly based on passed cipher.
2957 *) Belatedly make the BN tests actually check the results.
2960 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
2961 to and from BNs: it was completely broken. New compilation option
2962 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
2963 key elements as negative integers.
2966 *) Reorganize and speed up MD5.
2967 [Andy Polyakov <appro@fy.chalmers.se>]
2970 [Richard Levitte <richard@levitte.org>]
2972 *) New option -out to asn1parse to allow the parsed structure to be
2973 output to a file. This is most useful when combined with the -strparse
2974 option to examine the output of things like OCTET STRINGS.
2977 *) Make SSL library a little more fool-proof by not requiring any longer
2978 that SSL_set_{accept,connect}_state be called before
2979 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
2980 in many applications because usually everything *appeared* to work as
2981 intended anyway -- now it really works as intended).
2984 *) Move openssl.cnf out of lib/.
2987 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
2988 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
2989 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
2990 [Ralf S. Engelschall]
2992 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
2993 handle PKCS#7 enveloped data properly.
2994 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
2996 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
2997 copying pointers. The cert_st handling is changed by this in
2998 various ways (and thus what used to be known as ctx->default_cert
2999 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3000 any longer when s->cert does not give us what we need).
3001 ssl_cert_instantiate becomes obsolete by this change.
3002 As soon as we've got the new code right (possibly it already is?),
3003 we have solved a couple of bugs of the earlier code where s->cert
3004 was used as if it could not have been shared with other SSL structures.
3006 Note that using the SSL API in certain dirty ways now will result
3007 in different behaviour than observed with earlier library versions:
3008 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3009 does not influence s as it used to.
3011 In order to clean up things more thoroughly, inside SSL_SESSION
3012 we don't use CERT any longer, but a new structure SESS_CERT
3013 that holds per-session data (if available); currently, this is
3014 the peer's certificate chain and, for clients, the server's certificate
3015 and temporary key. CERT holds only those values that can have
3016 meaningful defaults in an SSL_CTX.
3019 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3020 from the internal representation. Various PKCS#7 fixes: remove some
3021 evil casts and set the enc_dig_alg field properly based on the signing
3025 *) Allow PKCS#12 password to be set from the command line or the
3026 environment. Let 'ca' get its config file name from the environment
3027 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3031 *) Allow certificate policies extension to use an IA5STRING for the
3032 organization field. This is contrary to the PKIX definition but
3033 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3037 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3038 without disallowing inline assembler and the like for non-pedantic builds.
3041 *) Support Borland C++ builder.
3042 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3047 *) SHA-1 cleanups and performance enhancements.
3048 [Andy Polyakov <appro@fy.chalmers.se>]
3050 *) Sparc v8plus assembler for the bignum library.
3051 [Andy Polyakov <appro@fy.chalmers.se>]
3053 *) Accept any -xxx and +xxx compiler options in Configure.
3056 *) Update HPUX configuration.
3059 *) Add missing sk_<type>_unshift() function to safestack.h
3060 [Ralf S. Engelschall]
3062 *) New function SSL_CTX_use_certificate_chain_file that sets the
3063 "extra_cert"s in addition to the certificate. (This makes sense
3064 only for "PEM" format files, as chains as a whole are not
3068 *) Support verify_depth from the SSL API.
3069 x509_vfy.c had what can be considered an off-by-one-error:
3070 Its depth (which was not part of the external interface)
3071 was actually counting the number of certificates in a chain;
3072 now it really counts the depth.
3075 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3076 instead of X509err, which often resulted in confusing error
3077 messages since the error codes are not globally unique
3078 (e.g. an alleged error in ssl3_accept when a certificate
3079 didn't match the private key).
3081 *) New function SSL_CTX_set_session_id_context that allows to set a default
3082 value (so that you don't need SSL_set_session_id_context for each
3083 connection using the SSL_CTX).
3086 *) OAEP decoding bug fix.
3089 *) Support INSTALL_PREFIX for package builders, as proposed by
3093 *) New Configure options "threads" and "no-threads". For systems
3094 where the proper compiler options are known (currently Solaris
3095 and Linux), "threads" is the default.
3098 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3101 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3102 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3103 such as /usr/local/bin.
3106 *) "make linux-shared" to build shared libraries.
3107 [Niels Poppe <niels@netbox.org>]
3109 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3112 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3113 extension adding in x509 utility.
3116 *) Remove NOPROTO sections and error code comments.
3119 *) Partial rewrite of the DEF file generator to now parse the ANSI
3123 *) New Configure options --prefix=DIR and --openssldir=DIR.
3126 *) Complete rewrite of the error code script(s). It is all now handled
3127 by one script at the top level which handles error code gathering,
3128 header rewriting and C source file generation. It should be much better
3129 than the old method: it now uses a modified version of Ulf's parser to
3130 read the ANSI prototypes in all header files (thus the old K&R definitions
3131 aren't needed for error creation any more) and do a better job of
3132 translating function codes into names. The old 'ASN1 error code imbedded
3133 in a comment' is no longer necessary and it doesn't use .err files which
3134 have now been deleted. Also the error code call doesn't have to appear all
3135 on one line (which resulted in some large lines...).
3138 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3141 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3142 0 (which usually indicates a closed connection), but continue reading.
3145 *) Fix some race conditions.
3148 *) Add support for CRL distribution points extension. Add Certificate
3149 Policies and CRL distribution points documentation.
3152 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3155 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3156 8 of keying material. Merlin has also confirmed interop with this fix
3157 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3158 [Merlin Hughes <merlin@baltimore.ie>]
3160 *) Fix lots of warnings.
3161 [Richard Levitte <levitte@stacken.kth.se>]
3163 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3164 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3165 [Richard Levitte <levitte@stacken.kth.se>]
3167 *) Fix problems with sizeof(long) == 8.
3168 [Andy Polyakov <appro@fy.chalmers.se>]
3170 *) Change functions to ANSI C.
3173 *) Fix typos in error codes.
3174 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3176 *) Remove defunct assembler files from Configure.
3179 *) SPARC v8 assembler BIGNUM implementation.
3180 [Andy Polyakov <appro@fy.chalmers.se>]
3182 *) Support for Certificate Policies extension: both print and set.
3183 Various additions to support the r2i method this uses.
3186 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3187 return a const string when you are expecting an allocated buffer.
3190 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3191 types DirectoryString and DisplayText.
3194 *) Add code to allow r2i extensions to access the configuration database,
3195 add an LHASH database driver and add several ctx helper functions.
3198 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3199 fail when they extended the size of a BIGNUM.
3202 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3203 support typesafe stack.
3206 *) Fix typo in SSL_[gs]et_options().
3207 [Nils Frostberg <nils@medcom.se>]
3209 *) Delete various functions and files that belonged to the (now obsolete)
3210 old X509V3 handling code.
3213 *) New Configure option "rsaref".
3216 *) Don't auto-generate pem.h.
3219 *) Introduce type-safe ASN.1 SETs.
3222 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3223 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3225 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3226 that links with OpenSSL (well at least cause lots of warnings), but fear
3227 not: the conversion is trivial, and it eliminates loads of evil casts. A
3228 few STACKed things have been converted already. Feel free to convert more.
3229 In the fullness of time, I'll do away with the STACK type altogether.
3232 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3233 specified in <certfile> by updating the entry in the index.txt file.
3234 This way one no longer has to edit the index.txt file manually for
3235 revoking a certificate. The -revoke option does the gory details now.
3236 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3238 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3239 `-text' option at all and this way the `-noout -text' combination was
3240 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3241 [Ralf S. Engelschall]
3243 *) Make sure a corresponding plain text error message exists for the
3244 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3245 verify callback function determined that a certificate was revoked.
3246 [Ralf S. Engelschall]
3248 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3249 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3250 all available cipers including rc5, which was forgotten until now.
3251 In order to let the testing shell script know which algorithms
3252 are available, a new (up to now undocumented) command
3253 "openssl list-cipher-commands" is used.
3256 *) Bugfix: s_client occasionally would sleep in select() when
3257 it should have checked SSL_pending() first.
3260 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3261 the raw DSA values prior to ASN.1 encoding.
3264 *) Tweaks to Configure
3265 [Niels Poppe <niels@netbox.org>]
3267 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3271 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3274 *) New config option to avoid instructions that are illegal on the 80386.
3275 The default code is faster, but requires at least a 486.
3278 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3279 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3280 same as SSL2_VERSION anyway.
3283 *) New "-showcerts" option for s_client.
3286 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3287 application. Various cleanups and fixes.
3290 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3291 modify error routines to work internally. Add error codes and PBE init
3292 to library startup routines.
3295 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3296 packing functions to asn1 and evp. Changed function names and error
3297 codes along the way.
3300 *) PKCS12 integration: and so it begins... First of several patches to
3301 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3302 objects to objects.h
3305 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3306 and display support for Thawte strong extranet extension.
3309 *) Add LinuxPPC support.
3310 [Jeff Dubrule <igor@pobox.org>]
3312 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3313 bn_div_words in alpha.s.
3314 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3316 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3317 OAEP isn't supported when OpenSSL is built with RSAref.
3318 [Ulf Moeller <ulf@fitug.de>]
3320 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3321 so they no longer are missing under -DNOPROTO.
3322 [Soren S. Jorvang <soren@t.dk>]
3325 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3327 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3328 doesn't work when the session is reused. Coming soon!
3331 *) Fix a security hole, that allows sessions to be reused in the wrong
3332 context thus bypassing client cert protection! All software that uses
3333 client certs and session caches in multiple contexts NEEDS PATCHING to
3334 allow session reuse! A fuller solution is in the works.
3335 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3337 *) Some more source tree cleanups (removed obsolete files
3338 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3339 permission on "config" script to be executable) and a fix for the INSTALL
3341 [Ulf Moeller <ulf@fitug.de>]
3343 *) Remove some legacy and erroneous uses of malloc, free instead of
3345 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3347 *) Make rsa_oaep_test return non-zero on error.
3348 [Ulf Moeller <ulf@fitug.de>]
3350 *) Add support for native Solaris shared libraries. Configure
3351 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3352 if someone would make that last step automatic.
3353 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3355 *) ctx_size was not built with the right compiler during "make links". Fixed.
3358 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3359 except NULL ciphers". This means the default cipher list will no longer
3360 enable NULL ciphers. They need to be specifically enabled e.g. with
3361 the string "DEFAULT:eNULL".
3364 *) Fix to RSA private encryption routines: if p < q then it would
3365 occasionally produce an invalid result. This will only happen with
3366 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3369 *) Be less restrictive and allow also `perl util/perlpath.pl
3370 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3371 because this way one can also use an interpreter named `perl5' (which is
3372 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3373 installed as `perl').
3374 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3376 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3377 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3379 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3380 advapi32.lib to Win32 build and change the pem test comparision
3381 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3382 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3383 and crypto/des/ede_cbcm_enc.c.
3386 *) DES quad checksum was broken on big-endian architectures. Fixed.
3389 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3390 Win32 test batch file so it (might) work again. The Win32 test batch file
3391 is horrible: I feel ill....
3394 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3395 in e_os.h. Audit of header files to check ANSI and non ANSI
3396 sections: 10 functions were absent from non ANSI section and not exported
3397 from Windows DLLs. Fixed up libeay.num for new functions.
3400 *) Make `openssl version' output lines consistent.
3401 [Ralf S. Engelschall]
3403 *) Fix Win32 symbol export lists for BIO functions: Added
3404 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3405 to ms/libeay{16,32}.def.
3406 [Ralf S. Engelschall]
3408 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3409 fine under Unix and passes some trivial tests I've now added. But the
3410 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3411 added to make sure no one expects that this stuff really works in the
3412 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3413 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3415 [Ralf S. Engelschall]
3417 *) Fix the generation of two part addresses in perl.
3418 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3420 *) Add config entry for Linux on MIPS.
3421 [John Tobey <jtobey@channel1.com>]
3423 *) Make links whenever Configure is run, unless we are on Windoze.
3426 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3427 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3431 *) Add a useful kludge to allow package maintainers to specify compiler and
3432 other platforms details on the command line without having to patch the
3433 Configure script everytime: One now can use ``perl Configure
3434 <id>:<details>'', i.e. platform ids are allowed to have details appended
3435 to them (seperated by colons). This is treated as there would be a static
3436 pre-configured entry in Configure's %table under key <id> with value
3437 <details> and ``perl Configure <id>'' is called. So, when you want to
3438 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3439 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3440 now, which overrides the FreeBSD-elf entry on-the-fly.
3441 [Ralf S. Engelschall]
3443 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3446 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3447 on the `perl Configure ...' command line. This way one can compile
3448 OpenSSL libraries with Position Independent Code (PIC) which is needed
3449 for linking it into DSOs.
3450 [Ralf S. Engelschall]
3452 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3456 *) Cleaned up the LICENSE document: The official contact for any license
3457 questions now is the OpenSSL core team under openssl-core@openssl.org.
3458 And add a paragraph about the dual-license situation to make sure people
3459 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3460 to the OpenSSL toolkit.
3461 [Ralf S. Engelschall]
3463 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3464 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3465 Additonally cleaned up the `make links' target: Remove unnecessary
3466 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3467 to speed processing and no longer clutter the display with confusing
3468 stuff. Instead only the actually done links are displayed.
3469 [Ralf S. Engelschall]
3471 *) Permit null encryption ciphersuites, used for authentication only. It used
3472 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3473 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3477 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3478 signed attributes when verifying signatures (this would break them),
3479 the detached data encoding was wrong and public keys obtained using
3480 X509_get_pubkey() weren't freed.
3483 *) Add text documentation for the BUFFER functions. Also added a work around
3484 to a Win95 console bug. This was triggered by the password read stuff: the
3485 last character typed gets carried over to the next fread(). If you were
3486 generating a new cert request using 'req' for example then the last
3487 character of the passphrase would be CR which would then enter the first
3491 *) Added the new `Includes OpenSSL Cryptography Software' button as
3492 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3493 button and can be used by applications based on OpenSSL to show the
3494 relationship to the OpenSSL project.
3495 [Ralf S. Engelschall]
3497 *) Remove confusing variables in function signatures in files
3498 ssl/ssl_lib.c and ssl/ssl.h.
3499 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3501 *) Don't install bss_file.c under PREFIX/include/
3502 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3504 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3505 functions that return function pointers and has support for NT specific
3506 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3507 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3508 unsigned to signed types: this was killing the Win32 compile.
3511 *) Add new certificate file to stack functions,
3512 SSL_add_dir_cert_subjects_to_stack() and
3513 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3514 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3515 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3516 This means that Apache-SSL and similar packages don't have to mess around
3517 to add as many CAs as they want to the preferred list.
3520 *) Experiment with doxygen documentation. Currently only partially applied to
3522 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3523 openssl.doxy as the configuration file.
3526 *) Get rid of remaining C++-style comments which strict C compilers hate.
3527 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3529 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3530 compiled in by default: it has problems with large keys.
3533 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3534 DH private keys and/or callback functions which directly correspond to
3535 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3536 is needed for applications which have to configure certificates on a
3537 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3539 For the RSA certificate situation is makes no difference, but
3540 for the DSA certificate situation this fixes the "no shared cipher"
3541 problem where the OpenSSL cipher selection procedure failed because the
3542 temporary keys were not overtaken from the context and the API provided
3543 no way to reconfigure them.
3544 The new functions now let applications reconfigure the stuff and they
3545 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3546 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3547 non-public-API function ssl_cert_instantiate() is used as a helper
3548 function and also to reduce code redundancy inside ssl_rsa.c.
3549 [Ralf S. Engelschall]
3551 *) Move s_server -dcert and -dkey options out of the undocumented feature
3552 area because they are useful for the DSA situation and should be
3553 recognized by the users.
3554 [Ralf S. Engelschall]
3556 *) Fix the cipher decision scheme for export ciphers: the export bits are
3557 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3558 SSL_EXP_MASK. So, the original variable has to be used instead of the
3559 already masked variable.
3560 [Richard Levitte <levitte@stacken.kth.se>]
3562 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3563 [Richard Levitte <levitte@stacken.kth.se>]
3565 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3566 from `int' to `unsigned int' because it's a length and initialized by
3567 EVP_DigestFinal() which expects an `unsigned int *'.
3568 [Richard Levitte <levitte@stacken.kth.se>]
3570 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3571 script. Instead use the usual Shell->Perl transition trick.
3572 [Ralf S. Engelschall]
3574 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3575 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3576 -noout -modulus' as it's already the case for `openssl rsa -noout
3577 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3578 currently the public key is printed (a decision which was already done by
3579 `openssl dsa -modulus' in the past) which serves a similar purpose.
3580 Additionally the NO_RSA no longer completely removes the whole -modulus
3581 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3583 [Ralf S. Engelschall]
3585 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3586 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3587 [Arne Ansper <arne@ats.cyber.ee>]
3589 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3590 to be added. Now both 'req' and 'ca' can use new objects defined in the
3594 *) Add cool BIO that does syslog (or event log on NT).
3595 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3597 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3598 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3599 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3600 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3603 *) Add preliminary config info for new extension code.
3606 *) Make RSA_NO_PADDING really use no padding.
3607 [Ulf Moeller <ulf@fitug.de>]
3609 *) Generate errors when private/public key check is done.
3612 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3613 for some CRL extensions and new objects added.
3616 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3617 key usage extension and fuller support for authority key id.
3620 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3621 padding method for RSA, which is recommended for new applications in PKCS
3622 #1 v2.0 (RFC 2437, October 1998).
3623 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3624 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3625 against Bleichbacher's attack on RSA.
3626 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3629 *) Updates to the new SSL compression code
3630 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3632 *) Fix so that the version number in the master secret, when passed
3633 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3634 (because the server will not accept higher), that the version number
3635 is 0x03,0x01, not 0x03,0x00
3636 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3638 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3639 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3640 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3643 *) Support for RAW extensions where an arbitrary extension can be
3644 created by including its DER encoding. See apps/openssl.cnf for
3648 *) Make sure latest Perl versions don't interpret some generated C array
3649 code as Perl array code in the crypto/err/err_genc.pl script.
3650 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3652 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3653 not many people have the assembler. Various Win32 compilation fixes and
3654 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3658 *) Modify configure script 'Configure' to automatically create crypto/date.h
3659 file under Win32 and also build pem.h from pem.org. New script
3660 util/mkfiles.pl to create the MINFO file on environments that can't do a
3661 'make files': perl util/mkfiles.pl >MINFO should work.
3664 *) Major rework of DES function declarations, in the pursuit of correctness
3665 and purity. As a result, many evil casts evaporated, and some weirdness,
3666 too. You may find this causes warnings in your code. Zapping your evil
3667 casts will probably fix them. Mostly.
3670 *) Fix for a typo in asn1.h. Bug fix to object creation script
3671 obj_dat.pl. It considered a zero in an object definition to mean
3672 "end of object": none of the objects in objects.h have any zeros
3673 so it wasn't spotted.
3674 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3676 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3677 Masking (CBCM). In the absence of test vectors, the best I have been able
3678 to do is check that the decrypt undoes the encrypt, so far. Send me test
3679 vectors if you have them.
3682 *) Correct calculation of key length for export ciphers (too much space was
3683 allocated for null ciphers). This has not been tested!
3686 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3687 message is now correct (it understands "crypto" and "ssl" on its
3688 command line). There is also now an "update" option. This will update
3689 the util/ssleay.num and util/libeay.num files with any new functions.
3691 perl util/mkdef.pl crypto ssl update
3692 it will update them.
3695 *) Overhauled the Perl interface (perl/*):
3696 - ported BN stuff to OpenSSL's different BN library
3697 - made the perl/ source tree CVS-aware
3698 - renamed the package from SSLeay to OpenSSL (the files still contain
3699 their history because I've copied them in the repository)
3700 - removed obsolete files (the test scripts will be replaced
3701 by better Test::Harness variants in the future)
3702 [Ralf S. Engelschall]
3704 *) First cut for a very conservative source tree cleanup:
3705 1. merge various obsolete readme texts into doc/ssleay.txt
3706 where we collect the old documents and readme texts.
3707 2. remove the first part of files where I'm already sure that we no
3708 longer need them because of three reasons: either they are just temporary
3709 files which were left by Eric or they are preserved original files where
3710 I've verified that the diff is also available in the CVS via "cvs diff
3711 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3712 the crypto/md/ stuff).
3713 [Ralf S. Engelschall]
3715 *) More extension code. Incomplete support for subject and issuer alt
3716 name, issuer and authority key id. Change the i2v function parameters
3717 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3718 what that's for :-) Fix to ASN1 macro which messed up
3719 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3722 *) Preliminary support for ENUMERATED type. This is largely copied from the
3726 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3727 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3729 *) Make sure `make rehash' target really finds the `openssl' program.
3730 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3732 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3733 like to hear about it if this slows down other processors.
3736 *) Add CygWin32 platform information to Configure script.
3737 [Alan Batie <batie@aahz.jf.intel.com>]
3739 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3740 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3742 *) New program nseq to manipulate netscape certificate sequences
3745 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3749 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3750 but the BN code had some problems that would cause failures when
3751 doing certificate verification and some other functions.
3752 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3754 *) Add ASN1 and PEM code to support netscape certificate sequences.
3757 *) Add ASN1 and PEM code to support netscape certificate sequences.
3760 *) Add several PKIX and private extended key usage OIDs.
3763 *) Modify the 'ca' program to handle the new extension code. Modify
3764 openssl.cnf for new extension format, add comments.
3767 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3768 and add a sample to openssl.cnf so req -x509 now adds appropriate
3772 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3773 error code, add initial support to X509_print() and x509 application.
3776 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3777 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3778 stuff is currently isolated and isn't even compiled yet.
3781 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3782 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3783 Removed the versions check from X509 routines when loading extensions:
3784 this allows certain broken certificates that don't set the version
3785 properly to be processed.
3788 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3789 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3790 can still be regenerated with "make depend".
3793 *) Spelling mistake in C version of CAST-128.
3794 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3796 *) Changes to the error generation code. The perl script err-code.pl
3797 now reads in the old error codes and retains the old numbers, only
3798 adding new ones if necessary. It also only changes the .err files if new
3799 codes are added. The makefiles have been modified to only insert errors
3800 when needed (to avoid needlessly modifying header files). This is done
3801 by only inserting errors if the .err file is newer than the auto generated
3802 C file. To rebuild all the error codes from scratch (the old behaviour)
3803 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3804 or delete all the .err files.
3807 *) CAST-128 was incorrectly implemented for short keys. The C version has
3808 been fixed, but is untested. The assembler versions are also fixed, but
3809 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3810 to regenerate it if needed.
3811 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3812 Hagino <itojun@kame.net>]
3814 *) File was opened incorrectly in randfile.c.
3815 [Ulf Möller <ulf@fitug.de>]
3817 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3818 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3819 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3820 al: it's just almost always a UTCTime. Note this patch adds new error
3821 codes so do a "make errors" if there are problems.
3824 *) Correct Linux 1 recognition in config.
3825 [Ulf Möller <ulf@fitug.de>]
3827 *) Remove pointless MD5 hash when using DSA keys in ca.
3828 [Anonymous <nobody@replay.com>]
3830 *) Generate an error if given an empty string as a cert directory. Also
3831 generate an error if handed NULL (previously returned 0 to indicate an
3832 error, but didn't set one).
3833 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3835 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3838 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3839 parameters. This was causing a warning which killed off the Win32 compile.
3842 *) Remove C++ style comments from crypto/bn/bn_local.h.
3843 [Neil Costigan <neil.costigan@celocom.com>]
3845 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3846 based on a text string, looking up short and long names and finally
3847 "dot" format. The "dot" format stuff didn't work. Added new function
3848 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3849 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3850 OID is not part of the table.
3853 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3854 X509_LOOKUP_by_alias().
3857 *) Sort openssl functions by name.
3860 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3861 encryption from sample DSA keys (in case anyone is interested the password
3865 *) Make _all_ *_free functions accept a NULL pointer.
3866 [Frans Heymans <fheymans@isaserver.be>]
3868 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3870 [Anonymous <nobody@replay.com>]
3872 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3873 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3875 *) Don't blow it for numeric -newkey arguments to apps/req.
3876 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3878 *) Temp key "for export" tests were wrong in s3_srvr.c.
3879 [Anonymous <nobody@replay.com>]
3881 *) Add prototype for temp key callback functions
3882 SSL_CTX_set_tmp_{rsa,dh}_callback().
3885 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3886 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3889 *) X509_name_add_entry() freed the wrong thing after an error.
3890 [Arne Ansper <arne@ats.cyber.ee>]
3892 *) rsa_eay.c would attempt to free a NULL context.
3893 [Arne Ansper <arne@ats.cyber.ee>]
3895 *) BIO_s_socket() had a broken should_retry() on Windoze.
3896 [Arne Ansper <arne@ats.cyber.ee>]
3898 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3899 [Arne Ansper <arne@ats.cyber.ee>]
3901 *) Make sure the already existing X509_STORE->depth variable is initialized
3902 in X509_STORE_new(), but document the fact that this variable is still
3903 unused in the certificate verification process.
3904 [Ralf S. Engelschall]
3906 *) Fix the various library and apps files to free up pkeys obtained from
3907 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3910 *) Fix reference counting in X509_PUBKEY_get(). This makes
3911 demos/maurice/example2.c work, amongst others, probably.
3912 [Steve Henson and Ben Laurie]
3914 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3915 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3916 are no longer created. This way we have a single and consistent command
3917 line interface `openssl <command>', similar to `cvs <command>'.
3918 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3920 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3921 BIT STRING wrapper always have zero unused bits.
3924 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3927 *) Make the top-level INSTALL documentation easier to understand.
3930 *) Makefiles updated to exit if an error occurs in a sub-directory
3931 make (including if user presses ^C) [Paul Sutton]
3933 *) Make Montgomery context stuff explicit in RSA data structure.
3936 *) Fix build order of pem and err to allow for generated pem.h.
3939 *) Fix renumbering bug in X509_NAME_delete_entry().
3942 *) Enhanced the err-ins.pl script so it makes the error library number
3943 global and can add a library name. This is needed for external ASN1 and
3944 other error libraries.
3947 *) Fixed sk_insert which never worked properly.
3950 *) Fix ASN1 macros so they can handle indefinite length construted
3951 EXPLICIT tags. Some non standard certificates use these: they can now
3955 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3956 into a single doc/ssleay.txt bundle. This way the information is still
3957 preserved but no longer messes up this directory. Now it's new room for
3958 the new set of documenation files.
3959 [Ralf S. Engelschall]
3961 *) SETs were incorrectly DER encoded. This was a major pain, because they
3962 shared code with SEQUENCEs, which aren't coded the same. This means that
3963 almost everything to do with SETs or SEQUENCEs has either changed name or
3964 number of arguments.
3965 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
3967 *) Fix test data to work with the above.
3970 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
3971 was already fixed by Eric for 0.9.1 it seems.
3972 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
3974 *) Autodetect FreeBSD3.
3977 *) Fix various bugs in Configure. This affects the following platforms:
3981 unixware-2.0-pentium
3985 *) Eliminate generated files from CVS. Reorder tests to regenerate files
3986 before they are needed.
3989 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
3993 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
3995 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
3996 changed SSLeay to OpenSSL in version strings.
3997 [Ralf S. Engelschall]
3999 *) Some fixups to the top-level documents.
4002 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4003 because the symlink to include/ was missing.
4004 [Ralf S. Engelschall]
4006 *) Incorporated the popular no-RSA/DSA-only patches
4007 which allow to compile a RSA-free SSLeay.
4008 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4010 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4011 when "ssleay" is still not found.
4012 [Ralf S. Engelschall]
4014 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4015 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4017 *) Updated the README file.
4018 [Ralf S. Engelschall]
4020 *) Added various .cvsignore files in the CVS repository subdirs
4021 to make a "cvs update" really silent.
4022 [Ralf S. Engelschall]
4024 *) Recompiled the error-definition header files and added
4025 missing symbols to the Win32 linker tables.
4026 [Ralf S. Engelschall]
4028 *) Cleaned up the top-level documents;
4029 o new files: CHANGES and LICENSE
4030 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4031 o merged COPYRIGHT into LICENSE
4032 o removed obsolete TODO file
4033 o renamed MICROSOFT to INSTALL.W32
4034 [Ralf S. Engelschall]
4036 *) Removed dummy files from the 0.9.1b source tree:
4037 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4038 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4039 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4040 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4041 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4042 [Ralf S. Engelschall]
4044 *) Added various platform portability fixes.
4047 *) The Genesis of the OpenSSL rpject:
4048 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4049 Young and Tim J. Hudson created while they were working for C2Net until
4051 [The OpenSSL Project]
4054 Changes between 0.9.0b and 0.9.1b [not released]
4056 *) Updated a few CA certificates under certs/
4059 *) Changed some BIGNUM api stuff.
4062 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4063 DGUX x86, Linux Alpha, etc.
4066 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4067 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4071 *) Add -strparse option to asn1pars program which parses nested
4073 [Dr Stephen Henson <shenson@bigfoot.com>]
4075 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4078 *) DSA fix for "ca" program.
4081 *) Added "-genkey" option to "dsaparam" program.
4084 *) Added RIPE MD160 (rmd160) message digest.
4087 *) Added -a (all) option to "ssleay version" command.
4090 *) Added PLATFORM define which is the id given to Configure.
4093 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4096 *) Extended the ASN.1 parser routines.
4099 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4102 *) Added a BN_CTX to the BN library.
4105 *) Fixed the weak key values in DES library
4108 *) Changed API in EVP library for cipher aliases.
4111 *) Added support for RC2/64bit cipher.
4114 *) Converted the lhash library to the crypto/mem.c functions.
4117 *) Added more recognized ASN.1 object ids.
4120 *) Added more RSA padding checks for SSL/TLS.
4123 *) Added BIO proxy/filter functionality.
4126 *) Added extra_certs to SSL_CTX which can be used
4127 send extra CA certificates to the client in the CA cert chain sending
4128 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4131 *) Now Fortezza is denied in the authentication phase because
4132 this is key exchange mechanism is not supported by SSLeay at all.
4135 *) Additional PKCS1 checks.
4138 *) Support the string "TLSv1" for all TLS v1 ciphers.
4141 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4142 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4145 *) Fixed a few memory leaks.
4148 *) Fixed various code and comment typos.
4151 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4152 bytes sent in the client random.
4153 [Edward Bishop <ebishop@spyglass.com>]