5 Changes between 0.9.6c and 0.9.6d [XX xxx XXXX]
7 *) Add the configuration target linux-s390x.
8 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
10 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
11 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
12 variable as an indication that a ClientHello message has been
13 received. As the flag value will be lost between multiple
14 invocations of ssl3_accept when using non-blocking I/O, the
15 function may not be aware that a handshake has actually taken
16 place, thus preventing a new session from being added to the
19 To avoid this problem, we now set s->new_session to 2 instead of
20 using a local variable.
21 [Lutz Jaenicke, Bodo Moeller]
23 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
24 if the SSL_R_LENGTH_MISMATCH error is detected.
25 [Geoff Thorpe, Bodo Moeller]
27 *) New 'shared_ldflag' column in Configure platform table.
30 *) Fix EVP_CIPHER_mode macro.
31 ["Dan S. Camper" <dan@bti.net>]
33 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
34 type, we must throw them away by setting rr->length to 0.
35 [D P Chang <dpc@qualys.com>]
37 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
39 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
40 <Dominikus.Scherkl@biodata.com>. (The previous implementation
41 worked incorrectly for those cases where range = 10..._2 and
42 3*range is two bits longer than range.)
45 *) Only add signing time to PKCS7 structures if it is not already
49 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
50 OBJ_ld_ce should be OBJ_id_ce.
51 Also some ip-pda OIDs in crypto/objects/objects.txt were
52 incorrect (cf. RFC 3039).
53 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
55 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
56 returns early because it has nothing to do.
57 [Andy Schneider <andy.schneider@bjss.co.uk>]
59 *) [In 0.9.6c-engine release:]
60 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
61 [Andy Schneider <andy.schneider@bjss.co.uk>]
63 *) [In 0.9.6c-engine release:]
64 Add support for Cryptographic Appliance's keyserver technology.
65 (Use engine 'keyclient')
66 [Cryptographic Appliances and Geoff Thorpe]
68 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
69 is called via tools/c89.sh because arguments have to be
70 rearranged (all '-L' options must appear before the first object
72 [Richard Shapiro <rshapiro@abinitio.com>]
74 *) [In 0.9.6c-engine release:]
75 Add support for Broadcom crypto accelerator cards, backported
77 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
79 *) [In 0.9.6c-engine release:]
80 Add support for SureWare crypto accelerator cards from
81 Baltimore Technologies. (Use engine 'sureware')
82 [Baltimore Technologies and Mark Cox]
84 *) [In 0.9.6c-engine release:]
85 Add support for crypto accelerator cards from Accelerated
86 Encryption Processing, www.aep.ie. (Use engine 'aep')
87 [AEP Inc. and Mark Cox]
89 *) Add a configuration entry for gcc on UnixWare.
90 [Gary Benson <gbenson@redhat.com>]
92 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
93 messages are stored in a single piece (fixed-length part and
94 variable-length part combined) and fix various bugs found on the way.
97 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
98 instead. BIO_gethostbyname() does not know what timeouts are
99 appropriate, so entries would stay in cache even when they have
101 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
103 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
104 faced with a pathologically small ClientHello fragment that does
105 not contain client_version: Instead of aborting with an error,
106 simply choose the highest available protocol version (i.e.,
107 TLS 1.0 unless it is disabled). In practice, ClientHello
108 messages are never sent like this, but this change gives us
109 strictly correct behaviour at least for TLS.
112 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
113 never resets s->method to s->ctx->method when called from within
114 one of the SSL handshake functions.
115 [Bodo Moeller; problem pointed out by Niko Baric]
117 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
118 (sent using the client's version number) if client_version is
119 smaller than the protocol version in use. Also change
120 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
121 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
122 the client will at least see that alert.
125 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
129 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
130 client receives HelloRequest while in a handshake.
131 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
133 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
134 should end in 'break', not 'goto end' which circuments various
135 cleanups done in state SSL_ST_OK. But session related stuff
136 must be disabled for SSL_ST_OK in the case that we just sent a
139 Also avoid some overhead by not calling ssl_init_wbio_buffer()
140 before just sending a HelloRequest.
141 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
143 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
144 reveal whether illegal block cipher padding was found or a MAC
145 verification error occured. (Neither SSLerr() codes nor alerts
146 are directly visible to potential attackers, but the information
147 may leak via logfiles.)
149 Similar changes are not required for the SSL 2.0 implementation
150 because the number of padding bytes is sent in clear for SSL 2.0,
151 and the extra bytes are just ignored. However ssl/s2_pkt.c
152 failed to verify that the purported number of padding bytes is in
156 *) Add OpenUNIX-8 support including shared libraries
157 (Boyd Lynn Gerber <gerberb@zenez.com>).
160 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
161 'wristwatch attack' using huge encoding parameters (cf.
162 James H. Manger's CRYPTO 2001 paper). Note that the
163 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
164 encoding parameters and hence was not vulnerable.
168 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
170 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
171 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
172 followed by modular reduction.
173 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
175 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
176 equivalent based on BN_pseudo_rand() instead of BN_rand().
179 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
180 This function was broken, as the check for a new client hello message
181 to handle SGC did not allow these large messages.
182 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
185 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
188 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
189 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
192 *) Rework the configuration and shared library support for Tru64 Unix.
193 The configuration part makes use of modern compiler features and
194 still retains old compiler behavior for those that run older versions
195 of the OS. The shared library support part includes a variant that
196 uses the RPATH feature, and is available through the special
197 configuration target "alpha-cc-rpath", which will never be selected
199 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
201 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
202 with the same message size as in ssl3_get_certificate_request().
203 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
204 messages might inadvertently be reject as too long.
205 [Petr Lampa <lampa@fee.vutbr.cz>]
207 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
210 *) Modified SSL library such that the verify_callback that has been set
211 specificly for an SSL object with SSL_set_verify() is actually being
212 used. Before the change, a verify_callback set with this function was
213 ignored and the verify_callback() set in the SSL_CTX at the time of
214 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
215 to allow the necessary settings.
218 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
219 explicitly to NULL, as at least on Solaris 8 this seems not always to be
220 done automatically (in contradiction to the requirements of the C
221 standard). This made problems when used from OpenSSH.
224 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
225 dh->length and always used
227 BN_rand_range(priv_key, dh->p).
229 BN_rand_range() is not necessary for Diffie-Hellman, and this
230 specific range makes Diffie-Hellman unnecessarily inefficient if
231 dh->length (recommended exponent length) is much smaller than the
232 length of dh->p. We could use BN_rand_range() if the order of
233 the subgroup was stored in the DH structure, but we only have
238 BN_rand(priv_key, l, ...)
240 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
246 RSA_eay_public_encrypt
247 RSA_eay_private_decrypt
248 RSA_eay_private_encrypt (signing)
249 RSA_eay_public_decrypt (signature verification)
251 (default implementations for RSA_public_encrypt,
252 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
253 always reject numbers >= n.
256 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
257 to synchronize access to 'locking_thread'. This is necessary on
258 systems where access to 'locking_thread' (an 'unsigned long'
259 variable) is not atomic.
262 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
263 *before* setting the 'crypto_lock_rand' flag. The previous code had
264 a race condition if 0 is a valid thread ID.
265 [Travis Vitek <vitek@roguewave.com>]
267 *) Add support for shared libraries under Irix.
268 [Albert Chin-A-Young <china@thewrittenword.com>]
270 *) Add configuration option to build on Linux on both big-endian and
272 [Ralf Baechle <ralf@uni-koblenz.de>]
274 *) Add the possibility to create shared libraries on HP-UX.
277 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
279 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
280 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
281 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
282 PRNG state recovery was possible based on the output of
283 one PRNG request appropriately sized to gain knowledge on
284 'md' followed by enough consecutive 1-byte PRNG requests
285 to traverse all of 'state'.
287 1. When updating 'md_local' (the current thread's copy of 'md')
288 during PRNG output generation, hash all of the previous
289 'md_local' value, not just the half used for PRNG output.
291 2. Make the number of bytes from 'state' included into the hash
292 independent from the number of PRNG bytes requested.
294 The first measure alone would be sufficient to avoid
295 Markku-Juhani's attack. (Actually it had never occurred
296 to me that the half of 'md_local' used for chaining was the
297 half from which PRNG output bytes were taken -- I had always
298 assumed that the secret half would be used.) The second
299 measure makes sure that additional data from 'state' is never
300 mixed into 'md_local' in small portions; this heuristically
301 further strengthens the PRNG.
304 *) Fix crypto/bn/asm/mips3.s.
307 *) When only the key is given to "enc", the IV is undefined. Print out
308 an error message in this case.
311 *) Handle special case when X509_NAME is empty in X509 printing routines.
314 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
315 positive and less than q.
318 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
319 used: it isn't thread safe and the add_lock_callback should handle
321 [Paul Rose <Paul.Rose@bridge.com>]
323 *) Verify that incoming data obeys the block size in
324 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
328 [Ulf Möller, Bodo Möller]
330 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
331 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
332 when fixing the server behaviour for backwards-compatible 'client
333 hello' messages. (Note that the attack is impractical against
334 SSL 3.0 and TLS 1.0 anyway because length and version checking
335 means that the probability of guessing a valid ciphertext is
336 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
339 Before 0.9.5, the countermeasure (hide the error by generating a
340 random 'decryption result') did not work properly because
341 ERR_clear_error() was missing, meaning that SSL_get_error() would
342 detect the supposedly ignored error.
344 Both problems are now fixed.
347 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
348 (previously it was 1024).
351 *) Fix for compatibility mode trust settings: ignore trust settings
352 unless some valid trust or reject settings are present.
355 *) Fix for blowfish EVP: its a variable length cipher.
358 *) Fix various bugs related to DSA S/MIME verification. Handle missing
359 parameters in DSA public key structures and return an error in the
360 DSA routines if parameters are absent.
363 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
364 in the current directory if neither $RANDFILE nor $HOME was set.
365 RAND_file_name() in 0.9.6a returned NULL in this case. This has
366 caused some confusion to Windows users who haven't defined $HOME.
367 Thus RAND_file_name() is changed again: e_os.h can define a
368 DEFAULT_HOME, which will be used if $HOME is not set.
369 For Windows, we use "C:"; on other platforms, we still require
370 environment variables.
372 *) Move 'if (!initialized) RAND_poll()' into regions protected by
373 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
374 having multiple threads call RAND_poll() concurrently.
377 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
378 combination of a flag and a thread ID variable.
379 Otherwise while one thread is in ssleay_rand_bytes (which sets the
380 flag), *other* threads can enter ssleay_add_bytes without obeying
381 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
382 that they do not hold after the first thread unsets add_do_not_lock).
385 *) Change bctest again: '-x' expressions are not available in all
389 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
391 *) Fix a couple of memory leaks in PKCS7_dataDecode()
392 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
394 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
395 the default extension for executables, if any. Also, make the perl
396 scripts that use symlink() to test if it really exists and use "cp"
397 if it doesn't. All this made OpenSSL compilable and installable in
401 *) Fix for asn1_GetSequence() for indefinite length constructed data.
402 If SEQUENCE is length is indefinite just set c->slen to the total
403 amount of data available.
404 [Steve Henson, reported by shige@FreeBSD.org]
405 [This change does not apply to 0.9.7.]
407 *) Change bctest to avoid here-documents inside command substitution
408 (workaround for FreeBSD /bin/sh bug).
409 For compatibility with Ultrix, avoid shell functions (introduced
410 in the bctest version that searches along $PATH).
413 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
414 with des_encrypt() defined on some operating systems, like Solaris
418 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
419 On the Importance of Eliminating Errors in Cryptographic
420 Computations, J. Cryptology 14 (2001) 2, 101-119,
421 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
424 *) MIPS assembler BIGNUM division bug fix.
427 *) Disabled incorrect Alpha assembler code.
430 *) Fix PKCS#7 decode routines so they correctly update the length
431 after reading an EOC for the EXPLICIT tag.
433 [This change does not apply to 0.9.7.]
435 *) Fix bug in PKCS#12 key generation routines. This was triggered
436 if a 3DES key was generated with a 0 initial byte. Include
437 PKCS12_BROKEN_KEYGEN compilation option to retain the old
438 (but broken) behaviour.
441 *) Enhance bctest to search for a working bc along $PATH and print
443 [Tim Rice <tim@multitalents.net> via Richard Levitte]
445 *) Fix memory leaks in err.c: free err_data string if necessary;
446 don't write to the wrong index in ERR_set_error_data.
449 *) Implement ssl23_peek (analogous to ssl23_read), which previously
453 *) Replace rdtsc with _emit statements for VC++ version 5.
454 [Jeremy Cooper <jeremy@baymoo.org>]
456 *) Make it possible to reuse SSLv2 sessions.
459 *) In copy_email() check for >= 0 as a return value for
460 X509_NAME_get_index_by_NID() since 0 is a valid index.
461 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
463 *) Avoid coredump with unsupported or invalid public keys by checking if
464 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
465 PKCS7_verify() fails with non detached data.
468 *) Don't use getenv in library functions when run as setuid/setgid.
469 New function OPENSSL_issetugid().
472 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
473 due to incorrect handling of multi-threading:
475 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
477 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
479 3. Count how many times MemCheck_off() has been called so that
480 nested use can be treated correctly. This also avoids
481 inband-signalling in the previous code (which relied on the
482 assumption that thread ID 0 is impossible).
485 *) Add "-rand" option also to s_client and s_server.
488 *) Fix CPU detection on Irix 6.x.
489 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
490 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
492 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
495 [This change does not apply to 0.9.7.]
497 *) Use the cached encoding of an X509_NAME structure rather than
498 copying it. This is apparently the reason for the libsafe "errors"
499 but the code is actually correct.
502 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
503 Bleichenbacher's DSA attack.
504 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
505 to be set and top=0 forces the highest bit to be set; top=-1 is new
506 and leaves the highest bit random.
507 [Ulf Moeller, Bodo Moeller]
509 *) In the NCONF_...-based implementations for CONF_... queries
510 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
511 a temporary CONF structure with the data component set to NULL
512 (which gives segmentation faults in lh_retrieve).
513 Instead, use NULL for the CONF pointer in CONF_get_string and
514 CONF_get_number (which may use environment variables) and directly
515 return NULL from CONF_get_section.
518 *) Fix potential buffer overrun for EBCDIC.
521 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
522 keyUsage if basicConstraints absent for a CA.
525 *) Make SMIME_write_PKCS7() write mail header values with a format that
526 is more generally accepted (no spaces before the semicolon), since
527 some programs can't parse those values properly otherwise. Also make
528 sure BIO's that break lines after each write do not create invalid
532 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
533 macros previously used would not encode an empty SEQUENCE OF
534 and break the signature.
536 [This change does not apply to 0.9.7.]
538 *) Zero the premaster secret after deriving the master secret in
542 *) Add some EVP_add_digest_alias registrations (as found in
543 OpenSSL_add_all_digests()) to SSL_library_init()
544 aka OpenSSL_add_ssl_algorithms(). This provides improved
545 compatibility with peers using X.509 certificates
546 with unconventional AlgorithmIdentifier OIDs.
549 *) Fix for Irix with NO_ASM.
550 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
552 *) ./config script fixes.
553 [Ulf Moeller, Richard Levitte]
555 *) Fix 'openssl passwd -1'.
558 *) Change PKCS12_key_gen_asc() so it can cope with non null
559 terminated strings whose length is passed in the passlen
560 parameter, for example from PEM callbacks. This was done
561 by adding an extra length parameter to asc2uni().
562 [Steve Henson, reported by <oddissey@samsung.co.kr>]
564 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
565 call failed, free the DSA structure.
568 *) Fix to uni2asc() to cope with zero length Unicode strings.
569 These are present in some PKCS#12 files.
572 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
573 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
574 when writing a 32767 byte record.
575 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
577 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
578 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
580 (RSA objects have a reference count access to which is protected
581 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
582 so they are meant to be shared between threads.)
583 [Bodo Moeller, Geoff Thorpe; original patch submitted by
584 "Reddie, Steven" <Steven.Reddie@ca.com>]
586 *) Fix a deadlock in CRYPTO_mem_leaks().
589 *) Use better test patterns in bntest.
592 *) rand_win.c fix for Borland C.
595 *) BN_rshift bugfix for n == 0.
598 *) Add a 'bctest' script that checks for some known 'bc' bugs
599 so that 'make test' does not abort just because 'bc' is broken.
602 *) Store verify_result within SSL_SESSION also for client side to
603 avoid potential security hole. (Re-used sessions on the client side
604 always resulted in verify_result==X509_V_OK, not using the original
605 result of the server certificate verification.)
608 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
609 SSL3_RT_APPLICATION_DATA, return 0.
610 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
614 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
615 releases, have been re-implemented by renaming the previous
616 implementations of ssl2_read and ssl3_read to ssl2_read_internal
617 and ssl3_read_internal, respectively, and adding 'peek' parameters
618 to them. The new ssl[23]_{read,peek} functions are calls to
619 ssl[23]_read_internal with the 'peek' flag set appropriately.
620 A 'peek' parameter has also been added to ssl3_read_bytes, which
621 does the actual work for ssl3_read_internal.
624 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
625 the method-specific "init()" handler. Also clean up ex_data after
626 calling the method-specific "finish()" handler. Previously, this was
627 happening the other way round.
630 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
631 The previous value, 12, was not always sufficient for BN_mod_exp().
634 *) Make sure that shared libraries get the internal name engine with
635 the full version number and not just 0. This should mark the
636 shared libraries as not backward compatible. Of course, this should
637 be changed again when we can guarantee backward binary compatibility.
640 *) Fix typo in get_cert_by_subject() in by_dir.c
641 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
643 *) Rework the system to generate shared libraries:
645 - Make note of the expected extension for the shared libraries and
646 if there is a need for symbolic links from for example libcrypto.so.0
647 to libcrypto.so.0.9.7. There is extended info in Configure for
650 - Make as few rebuilds of the shared libraries as possible.
652 - Still avoid linking the OpenSSL programs with the shared libraries.
654 - When installing, install the shared libraries separately from the
658 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
660 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
661 and not in SSL_clear because the latter is also used by the
662 accept/connect functions; previously, the settings made by
663 SSL_set_read_ahead would be lost during the handshake.
664 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
666 *) Correct util/mkdef.pl to be selective about disabled algorithms.
667 Previously, it would create entries for disableed algorithms no
671 *) Added several new manual pages for SSL_* function.
674 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
676 *) In ssl23_get_client_hello, generate an error message when faced
677 with an initial SSL 3.0/TLS record that is too small to contain the
678 first two bytes of the ClientHello message, i.e. client_version.
679 (Note that this is a pathologic case that probably has never happened
680 in real life.) The previous approach was to use the version number
681 from the record header as a substitute; but our protocol choice
682 should not depend on that one because it is not authenticated
683 by the Finished messages.
686 *) More robust randomness gathering functions for Windows.
687 [Jeffrey Altman <jaltman@columbia.edu>]
689 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
690 not set then we don't setup the error code for issuer check errors
691 to avoid possibly overwriting other errors which the callback does
692 handle. If an application does set the flag then we assume it knows
693 what it is doing and can handle the new informational codes
697 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
698 a general "ANY" type, as such it should be able to decode anything
699 including tagged types. However it didn't check the class so it would
700 wrongly interpret tagged types in the same way as their universal
701 counterpart and unknown types were just rejected. Changed so that the
702 tagged and unknown types are handled in the same way as a SEQUENCE:
703 that is the encoding is stored intact. There is also a new type
704 "V_ASN1_OTHER" which is used when the class is not universal, in this
705 case we have no idea what the actual type is so we just lump them all
709 *) On VMS, stdout may very well lead to a file that is written to
710 in a record-oriented fashion. That means that every write() will
711 write a separate record, which will be read separately by the
712 programs trying to read from it. This can be very confusing.
714 The solution is to put a BIO filter in the way that will buffer
715 text until a linefeed is reached, and then write everything a
716 line at a time, so every record written will be an actual line,
717 not chunks of lines and not (usually doesn't happen, but I've
718 seen it once) several lines in one record. BIO_f_linebuffer() is
721 Currently, it's a VMS-only method, because that's where it has
722 been tested well enough.
725 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
726 it can return incorrect results.
727 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
728 but it was in 0.9.6-beta[12].)
731 *) Disable the check for content being present when verifying detached
732 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
733 include zero length content when signing messages.
736 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
737 BIO_ctrl (for BIO pairs).
740 *) Add DSO method for VMS.
743 *) Bug fix: Montgomery multiplication could produce results with the
747 *) Add RPM specification openssl.spec and modify it to build three
748 packages. The default package contains applications, application
749 documentation and run-time libraries. The devel package contains
750 include files, static libraries and function documentation. The
751 doc package contains the contents of the doc directory. The original
752 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
755 *) Add a large number of documentation files for many SSL routines.
756 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
758 *) Add a configuration entry for Sony News 4.
759 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
761 *) Don't set the two most significant bits to one when generating a
762 random number < q in the DSA library.
765 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
766 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
767 the underlying transport is blocking) if a handshake took place.
768 (The default behaviour is needed by applications such as s_client
769 and s_server that use select() to determine when to use SSL_read;
770 but for applications that know in advance when to expect data, it
771 just makes things more complicated.)
774 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
778 *) Add a few more EBCDIC conditionals that make `req' and `x509'
779 work better on such systems.
780 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
782 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
783 Update PKCS12_parse() so it copies the friendlyName and the
784 keyid to the certificates aux info.
787 *) Fix bug in PKCS7_verify() which caused an infinite loop
788 if there was more than one signature.
789 [Sven Uszpelkat <su@celocom.de>]
791 *) Major change in util/mkdef.pl to include extra information
792 about each symbol, as well as presentig variables as well
793 as functions. This change means that there's n more need
794 to rebuild the .num files when some algorithms are excluded.
797 *) Allow the verify time to be set by an application,
798 rather than always using the current time.
801 *) Phase 2 verify code reorganisation. The certificate
802 verify code now looks up an issuer certificate by a
803 number of criteria: subject name, authority key id
804 and key usage. It also verifies self signed certificates
805 by the same criteria. The main comparison function is
806 X509_check_issued() which performs these checks.
808 Lot of changes were necessary in order to support this
809 without completely rewriting the lookup code.
811 Authority and subject key identifier are now cached.
813 The LHASH 'certs' is X509_STORE has now been replaced
814 by a STACK_OF(X509_OBJECT). This is mainly because an
815 LHASH can't store or retrieve multiple objects with
818 As a result various functions (which were all internal
819 use only) have changed to handle the new X509_STORE
820 structure. This will break anything that messed round
821 with X509_STORE internally.
823 The functions X509_STORE_add_cert() now checks for an
824 exact match, rather than just subject name.
826 The X509_STORE API doesn't directly support the retrieval
827 of multiple certificates matching a given criteria, however
828 this can be worked round by performing a lookup first
829 (which will fill the cache with candidate certificates)
830 and then examining the cache for matches. This is probably
831 the best we can do without throwing out X509_LOOKUP
832 entirely (maybe later...).
834 The X509_VERIFY_CTX structure has been enhanced considerably.
836 All certificate lookup operations now go via a get_issuer()
837 callback. Although this currently uses an X509_STORE it
838 can be replaced by custom lookups. This is a simple way
839 to bypass the X509_STORE hackery necessary to make this
840 work and makes it possible to use more efficient techniques
841 in future. A very simple version which uses a simple
842 STACK for its trusted certificate store is also provided
843 using X509_STORE_CTX_trusted_stack().
845 The verify_cb() and verify() callbacks now have equivalents
846 in the X509_STORE_CTX structure.
848 X509_STORE_CTX also has a 'flags' field which can be used
849 to customise the verify behaviour.
852 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
853 excludes S/MIME capabilities.
856 *) When a certificate request is read in keep a copy of the
857 original encoding of the signed data and use it when outputing
858 again. Signatures then use the original encoding rather than
859 a decoded, encoded version which may cause problems if the
860 request is improperly encoded.
863 *) For consistency with other BIO_puts implementations, call
864 buffer_write(b, ...) directly in buffer_puts instead of calling
867 In BIO_puts, increment b->num_write as in BIO_write.
868 [Peter.Sylvester@EdelWeb.fr]
870 *) Fix BN_mul_word for the case where the word is 0. (We have to use
871 BN_zero, we may not return a BIGNUM with an array consisting of
875 *) Avoid calling abort() from within the library when problems are
876 detected, except if preprocessor symbols have been defined
877 (such as REF_CHECK, BN_DEBUG etc.).
880 *) New openssl application 'rsautl'. This utility can be
881 used for low level RSA operations. DER public key
882 BIO/fp routines also added.
885 *) New Configure entry and patches for compiling on QNX 4.
886 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
888 *) A demo state-machine implementation was sponsored by
889 Nuron (http://www.nuron.com/) and is now available in
893 *) New options added to the 'dgst' utility for signature
894 generation and verification.
897 *) Unrecognized PKCS#7 content types are now handled via a
898 catch all ASN1_TYPE structure. This allows unsupported
899 types to be stored as a "blob" and an application can
900 encode and decode it manually.
903 *) Fix various signed/unsigned issues to make a_strex.c
905 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
907 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
908 length if passed a buffer. ASN1_INTEGER_to_BN failed
909 if passed a NULL BN and its argument was negative.
910 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
912 *) Modification to PKCS#7 encoding routines to output definite
913 length encoding. Since currently the whole structures are in
914 memory there's not real point in using indefinite length
915 constructed encoding. However if OpenSSL is compiled with
916 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
919 *) Added BIO_vprintf() and BIO_vsnprintf().
922 *) Added more prefixes to parse for in the the strings written
923 through a logging bio, to cover all the levels that are available
924 through syslog. The prefixes are now:
926 PANIC, EMERG, EMR => LOG_EMERG
927 ALERT, ALR => LOG_ALERT
928 CRIT, CRI => LOG_CRIT
929 ERROR, ERR => LOG_ERR
930 WARNING, WARN, WAR => LOG_WARNING
931 NOTICE, NOTE, NOT => LOG_NOTICE
932 INFO, INF => LOG_INFO
933 DEBUG, DBG => LOG_DEBUG
935 and as before, if none of those prefixes are present at the
936 beginning of the string, LOG_ERR is chosen.
938 On Win32, the LOG_* levels are mapped according to this:
940 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
941 LOG_WARNING => EVENTLOG_WARNING_TYPE
942 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
946 *) Made it possible to reconfigure with just the configuration
947 argument "reconf" or "reconfigure". The command line arguments
948 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
949 and are retrieved from there when reconfiguring.
953 [Assar Westerlund <assar@sics.se>, Richard Levitte]
955 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
958 *) The obj_dat.pl script was messing up the sorting of object
959 names. The reason was that it compared the quoted version
960 of strings as a result "OCSP" > "OCSP Signing" because
961 " > SPACE. Changed script to store unquoted versions of
962 names and add quotes on output. It was also omitting some
963 names from the lookup table if they were given a default
964 value (that is if SN is missing it is given the same
965 value as LN and vice versa), these are now added on the
966 grounds that if an object has a name we should be able to
967 look it up. Finally added warning output when duplicate
968 short or long names are found.
971 *) Changes needed for Tandem NSK.
972 [Scott Uroff <scott@xypro.com>]
974 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
975 RSA_padding_check_SSLv23(), special padding was never detected
976 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
977 version rollback attacks was not effective.
979 In s23_clnt.c, don't use special rollback-attack detection padding
980 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
981 client; similarly, in s23_srvr.c, don't do the rollback check if
982 SSL 2.0 is the only protocol enabled in the server.
985 *) Make it possible to get hexdumps of unprintable data with 'openssl
986 asn1parse'. By implication, the functions ASN1_parse_dump() and
987 BIO_dump_indent() are added.
990 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
991 these print out strings and name structures based on various
992 flags including RFC2253 support and proper handling of
993 multibyte characters. Added options to the 'x509' utility
994 to allow the various flags to be set.
997 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
998 Also change the functions X509_cmp_current_time() and
999 X509_gmtime_adj() work with an ASN1_TIME structure,
1000 this will enable certificates using GeneralizedTime in validity
1001 dates to be checked.
1004 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
1005 negative public key encodings) on by default,
1006 NO_NEG_PUBKEY_BUG can be set to disable it.
1009 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
1010 content octets. An i2c_ASN1_OBJECT is unnecessary because
1011 the encoding can be trivially obtained from the structure.
1014 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
1015 not read locks (CRYPTO_r_[un]lock).
1018 *) A first attempt at creating official support for shared
1019 libraries through configuration. I've kept it so the
1020 default is static libraries only, and the OpenSSL programs
1021 are always statically linked for now, but there are
1022 preparations for dynamic linking in place.
1023 This has been tested on Linux and Tru64.
1026 *) Randomness polling function for Win9x, as described in:
1027 Peter Gutmann, Software Generation of Practically Strong
1031 *) Fix so PRNG is seeded in req if using an already existing
1035 *) New options to smime application. -inform and -outform
1036 allow alternative formats for the S/MIME message including
1037 PEM and DER. The -content option allows the content to be
1038 specified separately. This should allow things like Netscape
1039 form signing output easier to verify.
1042 *) Fix the ASN1 encoding of tags using the 'long form'.
1045 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
1046 STRING types. These convert content octets to and from the
1047 underlying type. The actual tag and length octets are
1048 already assumed to have been read in and checked. These
1049 are needed because all other string types have virtually
1050 identical handling apart from the tag. By having versions
1051 of the ASN1 functions that just operate on content octets
1052 IMPLICIT tagging can be handled properly. It also allows
1053 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
1054 and ASN1_INTEGER are identical apart from the tag.
1057 *) Change the handling of OID objects as follows:
1059 - New object identifiers are inserted in objects.txt, following
1060 the syntax given in objects.README.
1061 - objects.pl is used to process obj_mac.num and create a new
1063 - obj_dat.pl is used to create a new obj_dat.h, using the data in
1066 This is currently kind of a hack, and the perl code in objects.pl
1067 isn't very elegant, but it works as I intended. The simplest way
1068 to check that it worked correctly is to look in obj_dat.h and
1069 check the array nid_objs and make sure the objects haven't moved
1070 around (this is important!). Additions are OK, as well as
1071 consistent name changes.
1074 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
1077 *) Addition of the command line parameter '-rand file' to 'openssl req'.
1078 The given file adds to whatever has already been seeded into the
1079 random pool through the RANDFILE configuration file option or
1080 environment variable, or the default random state file.
1083 *) mkstack.pl now sorts each macro group into lexical order.
1084 Previously the output order depended on the order the files
1085 appeared in the directory, resulting in needless rewriting
1089 *) Patches to make OpenSSL compile under Win32 again. Mostly
1090 work arounds for the VC++ problem that it treats func() as
1091 func(void). Also stripped out the parts of mkdef.pl that
1092 added extra typesafe functions: these no longer exist.
1095 *) Reorganisation of the stack code. The macros are now all
1096 collected in safestack.h . Each macro is defined in terms of
1097 a "stack macro" of the form SKM_<name>(type, a, b). The
1098 DEBUG_SAFESTACK is now handled in terms of function casts,
1099 this has the advantage of retaining type safety without the
1100 use of additional functions. If DEBUG_SAFESTACK is not defined
1101 then the non typesafe macros are used instead. Also modified the
1102 mkstack.pl script to handle the new form. Needs testing to see
1103 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
1104 the default if no major problems. Similar behaviour for ASN1_SET_OF
1105 and PKCS12_STACK_OF.
1108 *) When some versions of IIS use the 'NET' form of private key the
1109 key derivation algorithm is different. Normally MD5(password) is
1110 used as a 128 bit RC4 key. In the modified case
1111 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
1112 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
1113 as the old Netscape_RSA functions except they have an additional
1114 'sgckey' parameter which uses the modified algorithm. Also added
1115 an -sgckey command line option to the rsa utility. Thanks to
1116 Adrian Peck <bertie@ncipher.com> for posting details of the modified
1117 algorithm to openssl-dev.
1120 *) The evp_local.h macros were using 'c.##kname' which resulted in
1121 invalid expansion on some systems (SCO 5.0.5 for example).
1122 Corrected to 'c.kname'.
1123 [Phillip Porch <root@theporch.com>]
1125 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
1126 a STACK of email addresses from a certificate or request, these look
1127 in the subject name and the subject alternative name extensions and
1128 omit any duplicate addresses.
1131 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
1132 This makes DSA verification about 2 % faster.
1135 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
1136 (meaning that now 2^5 values will be precomputed, which is only 4 KB
1137 plus overhead for 1024 bit moduli).
1138 This makes exponentiations about 0.5 % faster for 1024 bit
1139 exponents (as measured by "openssl speed rsa2048").
1142 *) Rename memory handling macros to avoid conflicts with other
1144 Malloc => OPENSSL_malloc
1145 Malloc_locked => OPENSSL_malloc_locked
1146 Realloc => OPENSSL_realloc
1147 Free => OPENSSL_free
1150 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
1151 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
1154 *) CygWin32 support.
1155 [John Jarvie <jjarvie@newsguy.com>]
1157 *) The type-safe stack code has been rejigged. It is now only compiled
1158 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
1159 by default all type-specific stack functions are "#define"d back to
1160 standard stack functions. This results in more streamlined output
1161 but retains the type-safety checking possibilities of the original
1165 *) The STACK code has been cleaned up, and certain type declarations
1166 that didn't make a lot of sense have been brought in line. This has
1167 also involved a cleanup of sorts in safestack.h to more correctly
1168 map type-safe stack functions onto their plain stack counterparts.
1169 This work has also resulted in a variety of "const"ifications of
1170 lots of the code, especially "_cmp" operations which should normally
1171 be prototyped with "const" parameters anyway.
1174 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
1175 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
1176 (The PRNG state consists of two parts, the large pool 'state' and 'md',
1177 where all of 'md' is used each time the PRNG is used, but 'state'
1178 is used only indexed by a cyclic counter. As entropy may not be
1179 well distributed from the beginning, 'md' is important as a
1180 chaining variable. However, the output function chains only half
1181 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
1182 all of 'md', and seeding with STATE_SIZE dummy bytes will result
1183 in all of 'state' being rewritten, with the new values depending
1184 on virtually all of 'md'. This overcomes the 80 bit limitation.)
1187 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
1188 the handshake is continued after ssl_verify_cert_chain();
1189 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
1190 can lead to 'unexplainable' connection aborts later.
1191 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
1193 *) Major EVP API cipher revision.
1194 Add hooks for extra EVP features. This allows various cipher
1195 parameters to be set in the EVP interface. Support added for variable
1196 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
1197 setting of RC2 and RC5 parameters.
1199 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
1202 Remove lots of duplicated code from the EVP library. For example *every*
1203 cipher init() function handles the 'iv' in the same way according to the
1204 cipher mode. They also all do nothing if the 'key' parameter is NULL and
1205 for CFB and OFB modes they zero ctx->num.
1207 New functionality allows removal of S/MIME code RC2 hack.
1209 Most of the routines have the same form and so can be declared in terms
1212 By shifting this to the top level EVP_CipherInit() it can be removed from
1213 all individual ciphers. If the cipher wants to handle IVs or keys
1214 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
1217 Change lots of functions like EVP_EncryptUpdate() to now return a
1218 value: although software versions of the algorithms cannot fail
1219 any installed hardware versions can.
1222 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
1223 this option is set, tolerate broken clients that send the negotiated
1224 protocol version number instead of the requested protocol version
1228 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
1229 i.e. non-zero for export ciphersuites, zero otherwise.
1230 Previous versions had this flag inverted, inconsistent with
1231 rsa_tmp_cb (..._TMP_RSA_CB).
1232 [Bodo Moeller; problem reported by Amit Chopra]
1234 *) Add missing DSA library text string. Work around for some IIS
1235 key files with invalid SEQUENCE encoding.
1238 *) Add a document (doc/standards.txt) that list all kinds of standards
1239 and so on that are implemented in OpenSSL.
1242 *) Enhance c_rehash script. Old version would mishandle certificates
1243 with the same subject name hash and wouldn't handle CRLs at all.
1244 Added -fingerprint option to crl utility, to support new c_rehash
1248 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
1251 *) Fix for SSL server purpose checking. Server checking was
1252 rejecting certificates which had extended key usage present
1253 but no ssl client purpose.
1254 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
1256 *) Make PKCS#12 code work with no password. The PKCS#12 spec
1257 is a little unclear about how a blank password is handled.
1258 Since the password in encoded as a BMPString with terminating
1259 double NULL a zero length password would end up as just the
1260 double NULL. However no password at all is different and is
1261 handled differently in the PKCS#12 key generation code. NS
1262 treats a blank password as zero length. MSIE treats it as no
1263 password on export: but it will try both on import. We now do
1264 the same: PKCS12_parse() tries zero length and no password if
1265 the password is set to "" or NULL (NULL is now a valid password:
1266 it wasn't before) as does the pkcs12 application.
1269 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
1270 perror when PEM_read_bio_X509_REQ fails, the error message must
1271 be obtained from the error queue.
1274 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
1275 it in ERR_remove_state if appropriate, and change ERR_get_state
1276 accordingly to avoid race conditions (this is necessary because
1277 thread_hash is no longer constant once set).
1280 *) Bugfix for linux-elf makefile.one.
1283 *) RSA_get_default_method() will now cause a default
1284 RSA_METHOD to be chosen if one doesn't exist already.
1285 Previously this was only set during a call to RSA_new()
1286 or RSA_new_method(NULL) meaning it was possible for
1287 RSA_get_default_method() to return NULL.
1290 *) Added native name translation to the existing DSO code
1291 that will convert (if the flag to do so is set) filenames
1292 that are sufficiently small and have no path information
1293 into a canonical native form. Eg. "blah" converted to
1294 "libblah.so" or "blah.dll" etc.
1297 *) New function ERR_error_string_n(e, buf, len) which is like
1298 ERR_error_string(e, buf), but writes at most 'len' bytes
1299 including the 0 terminator. For ERR_error_string_n, 'buf'
1301 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
1303 *) CONF library reworked to become more general. A new CONF
1304 configuration file reader "class" is implemented as well as a
1305 new functions (NCONF_*, for "New CONF") to handle it. The now
1306 old CONF_* functions are still there, but are reimplemented to
1307 work in terms of the new functions. Also, a set of functions
1308 to handle the internal storage of the configuration data is
1309 provided to make it easier to write new configuration file
1310 reader "classes" (I can definitely see something reading a
1311 configuration file in XML format, for example), called _CONF_*,
1312 or "the configuration storage API"...
1314 The new configuration file reading functions are:
1316 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
1317 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
1319 NCONF_default, NCONF_WIN32
1321 NCONF_dump_fp, NCONF_dump_bio
1323 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
1324 NCONF_new creates a new CONF object. This works in the same way
1325 as other interfaces in OpenSSL, like the BIO interface.
1326 NCONF_dump_* dump the internal storage of the configuration file,
1327 which is useful for debugging. All other functions take the same
1328 arguments as the old CONF_* functions wth the exception of the
1329 first that must be a `CONF *' instead of a `LHASH *'.
1331 To make it easer to use the new classes with the old CONF_* functions,
1332 the function CONF_set_default_method is provided.
1335 *) Add '-tls1' option to 'openssl ciphers', which was already
1336 mentioned in the documentation but had not been implemented.
1337 (This option is not yet really useful because even the additional
1338 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
1341 *) Initial DSO code added into libcrypto for letting OpenSSL (and
1342 OpenSSL-based applications) load shared libraries and bind to
1343 them in a portable way.
1344 [Geoff Thorpe, with contributions from Richard Levitte]
1346 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
1348 *) Make sure _lrotl and _lrotr are only used with MSVC.
1350 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
1351 (the default implementation of RAND_status).
1353 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
1354 to '-clrext' (= clear extensions), as intended and documented.
1355 [Bodo Moeller; inconsistency pointed out by Michael Attili
1358 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
1359 was larger than the MD block size.
1360 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
1362 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
1363 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
1364 using the passed key: if the passed key was a private key the result
1365 of X509_print(), for example, would be to print out all the private key
1369 *) des_quad_cksum() byte order bug fix.
1370 [Ulf Möller, using the problem description in krb4-0.9.7, where
1371 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
1373 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
1375 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
1377 *) For easily testing in shell scripts whether some command
1378 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
1379 returns with exit code 0 iff no command of the given name is available.
1380 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
1381 the output goes to stdout and nothing is printed to stderr.
1382 Additional arguments are always ignored.
1384 Since for each cipher there is a command of the same name,
1385 the 'no-cipher' compilation switches can be tested this way.
1387 ('openssl no-XXX' is not able to detect pseudo-commands such
1388 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
1391 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
1394 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
1395 is set; it will be thrown away anyway because each handshake creates
1397 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
1398 to parameters -- in previous versions (since OpenSSL 0.9.3) the
1399 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
1400 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
1403 *) New s_client option -ign_eof: EOF at stdin is ignored, and
1404 'Q' and 'R' lose their special meanings (quit/renegotiate).
1405 This is part of what -quiet does; unlike -quiet, -ign_eof
1406 does not suppress any output.
1409 *) Add compatibility options to the purpose and trust code. The
1410 purpose X509_PURPOSE_ANY is "any purpose" which automatically
1411 accepts a certificate or CA, this was the previous behaviour,
1412 with all the associated security issues.
1414 X509_TRUST_COMPAT is the old trust behaviour: only and
1415 automatically trust self signed roots in certificate store. A
1416 new trust setting X509_TRUST_DEFAULT is used to specify that
1417 a purpose has no associated trust setting and it should instead
1418 use the value in the default purpose.
1421 *) Fix the PKCS#8 DSA private key code so it decodes keys again
1422 and fix a memory leak.
1425 *) In util/mkerr.pl (which implements 'make errors'), preserve
1426 reason strings from the previous version of the .c file, as
1427 the default to have only downcase letters (and digits) in
1428 automatically generated reasons codes is not always appropriate.
1431 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
1432 using strerror. Previously, ERR_reason_error_string() returned
1433 library names as reason strings for SYSerr; but SYSerr is a special
1434 case where small numbers are errno values, not library numbers.
1437 *) Add '-dsaparam' option to 'openssl dhparam' application. This
1438 converts DSA parameters into DH parameters. (When creating parameters,
1439 DSA_generate_parameters is used.)
1442 *) Include 'length' (recommended exponent length) in C code generated
1443 by 'openssl dhparam -C'.
1446 *) The second argument to set_label in perlasm was already being used
1447 so couldn't be used as a "file scope" flag. Moved to third argument
1451 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
1452 instead of RAND_bytes for encryption IVs and salts.
1455 *) Include RAND_status() into RAND_METHOD instead of implementing
1456 it only for md_rand.c Otherwise replacing the PRNG by calling
1457 RAND_set_rand_method would be impossible.
1460 *) Don't let DSA_generate_key() enter an infinite loop if the random
1461 number generation fails.
1464 *) New 'rand' application for creating pseudo-random output.
1467 *) Added configuration support for Linux/IA64
1468 [Rolf Haberrecker <rolf@suse.de>]
1470 *) Assembler module support for Mingw32.
1473 *) Shared library support for HPUX (in shlib/).
1474 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
1476 *) Shared library support for Solaris gcc.
1477 [Lutz Behnke <behnke@trustcenter.de>]
1479 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
1481 *) PKCS7_encrypt() was adding text MIME headers twice because they
1482 were added manually and by SMIME_crlf_copy().
1485 *) In bntest.c don't call BN_rand with zero bits argument.
1486 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
1488 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
1489 case was implemented. This caused BN_div_recp() to fail occasionally.
1492 *) Add an optional second argument to the set_label() in the perl
1493 assembly language builder. If this argument exists and is set
1494 to 1 it signals that the assembler should use a symbol whose
1495 scope is the entire file, not just the current function. This
1496 is needed with MASM which uses the format label:: for this scope.
1497 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
1499 *) Change the ASN1 types so they are typedefs by default. Before
1500 almost all types were #define'd to ASN1_STRING which was causing
1501 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
1505 *) Change names of new functions to the new get1/get0 naming
1506 convention: After 'get1', the caller owns a reference count
1507 and has to call ..._free; 'get0' returns a pointer to some
1508 data structure without incrementing reference counters.
1509 (Some of the existing 'get' functions increment a reference
1510 counter, some don't.)
1511 Similarly, 'set1' and 'add1' functions increase reference
1512 counters or duplicate objects.
1515 *) Allow for the possibility of temp RSA key generation failure:
1516 the code used to assume it always worked and crashed on failure.
1519 *) Fix potential buffer overrun problem in BIO_printf().
1520 [Ulf Möller, using public domain code by Patrick Powell; problem
1521 pointed out by David Sacerdote <das33@cornell.edu>]
1523 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
1524 RAND_egd() and RAND_status(). In the command line application,
1525 the EGD socket can be specified like a seed file using RANDFILE
1529 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
1530 Some CAs (e.g. Verisign) distribute certificates in this form.
1533 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
1534 list to exclude them. This means that no special compilation option
1535 is needed to use anonymous DH: it just needs to be included in the
1539 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
1540 EVP_MD_type. The old functionality is available in a new macro called
1541 EVP_MD_md(). Change code that uses it and update docs.
1544 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
1545 where the 'void *' argument is replaced by a function pointer argument.
1546 Previously 'void *' was abused to point to functions, which works on
1547 many platforms, but is not correct. As these functions are usually
1548 called by macros defined in OpenSSL header files, most source code
1549 should work without changes.
1552 *) <openssl/opensslconf.h> (which is created by Configure) now contains
1553 sections with information on -D... compiler switches used for
1554 compiling the library so that applications can see them. To enable
1555 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
1556 must be defined. E.g.,
1557 #define OPENSSL_ALGORITHM_DEFINES
1558 #include <openssl/opensslconf.h>
1559 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
1560 [Richard Levitte, Ulf and Bodo Möller]
1562 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
1566 *) Change the 'other' type in certificate aux info to a STACK_OF
1567 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
1568 the required ASN1 format: arbitrary types determined by an OID.
1571 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
1572 argument to 'req'. This is not because the function is newer or
1573 better than others it just uses the work 'NEW' in the certificate
1574 request header lines. Some software needs this.
1577 *) Reorganise password command line arguments: now passwords can be
1578 obtained from various sources. Delete the PEM_cb function and make
1579 it the default behaviour: i.e. if the callback is NULL and the
1580 usrdata argument is not NULL interpret it as a null terminated pass
1581 phrase. If usrdata and the callback are NULL then the pass phrase
1582 is prompted for as usual.
1585 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
1586 the support is automatically enabled. The resulting binaries will
1587 autodetect the card and use it if present.
1588 [Ben Laurie and Compaq Inc.]
1590 *) Work around for Netscape hang bug. This sends certificate request
1591 and server done in one record. Since this is perfectly legal in the
1592 SSL/TLS protocol it isn't a "bug" option and is on by default. See
1593 the bugs/SSLv3 entry for more info.
1596 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
1599 *) Add -rand argument to smime and pkcs12 applications and read/write
1603 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
1606 *) Add command line password options to the remaining applications.
1609 *) Bug fix for BN_div_recp() for numerators with an even number of
1613 *) More tests in bntest.c, and changed test_bn output.
1616 *) ./config recognizes MacOS X now.
1619 *) Bug fix for BN_div() when the first words of num and divsor are
1620 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
1623 *) Add support for various broken PKCS#8 formats, and command line
1624 options to produce them.
1627 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
1628 get temporary BIGNUMs from a BN_CTX.
1631 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
1635 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
1636 include a #define from the old name to the new. The original intent
1637 was that statically linked binaries could for example just call
1638 SSLeay_add_all_ciphers() to just add ciphers to the table and not
1639 link with digests. This never worked becayse SSLeay_add_all_digests()
1640 and SSLeay_add_all_ciphers() were in the same source file so calling
1641 one would link with the other. They are now in separate source files.
1644 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
1647 *) Use a less unusual form of the Miller-Rabin primality test (it used
1648 a binary algorithm for exponentiation integrated into the Miller-Rabin
1649 loop, our standard modexp algorithms are faster).
1652 *) Support for the EBCDIC character set completed.
1653 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
1655 *) Source code cleanups: use const where appropriate, eliminate casts,
1656 use void * instead of char * in lhash.
1659 *) Bugfix: ssl3_send_server_key_exchange was not restartable
1660 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
1661 this the server could overwrite ephemeral keys that the client
1665 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
1666 using 50 iterations of the Rabin-Miller test.
1668 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
1669 iterations of the Rabin-Miller test as required by the appendix
1670 to FIPS PUB 186[-1]) instead of DSA_is_prime.
1671 As BN_is_prime_fasttest includes trial division, DSA parameter
1672 generation becomes much faster.
1674 This implies a change for the callback functions in DSA_is_prime
1675 and DSA_generate_parameters: The callback function is called once
1676 for each positive witness in the Rabin-Miller test, not just
1677 occasionally in the inner loop; and the parameters to the
1678 callback function now provide an iteration count for the outer
1679 loop rather than for the current invocation of the inner loop.
1680 DSA_generate_parameters additionally can call the callback
1681 function with an 'iteration count' of -1, meaning that a
1682 candidate has passed the trial division test (when q is generated
1683 from an application-provided seed, trial division is skipped).
1686 *) New function BN_is_prime_fasttest that optionally does trial
1687 division before starting the Rabin-Miller test and has
1688 an additional BN_CTX * argument (whereas BN_is_prime always
1689 has to allocate at least one BN_CTX).
1690 'callback(1, -1, cb_arg)' is called when a number has passed the
1691 trial division stage.
1694 *) Fix for bug in CRL encoding. The validity dates weren't being handled
1698 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
1701 *) New function BN_pseudo_rand().
1704 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
1705 bignum version of BN_from_montgomery() with the working code from
1706 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
1710 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
1711 made it impossible to use the same SSL_SESSION data structure in
1712 SSL2 clients in multiple threads.
1715 *) The return value of RAND_load_file() no longer counts bytes obtained
1716 by stat(). RAND_load_file(..., -1) is new and uses the complete file
1717 to seed the PRNG (previously an explicit byte count was required).
1718 [Ulf Möller, Bodo Möller]
1720 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
1721 used (char *) instead of (void *) and had casts all over the place.
1724 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
1727 *) Retain source code compatibility for BN_prime_checks macro:
1728 BN_is_prime(..., BN_prime_checks, ...) now uses
1729 BN_prime_checks_for_size to determine the appropriate number of
1730 Rabin-Miller iterations.
1733 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
1734 DH_CHECK_P_NOT_SAFE_PRIME.
1735 (Check if this is true? OpenPGP calls them "strong".)
1738 *) Merge the functionality of "dh" and "gendh" programs into a new program
1739 "dhparam". The old programs are retained for now but will handle DH keys
1740 (instead of parameters) in future.
1743 *) Make the ciphers, s_server and s_client programs check the return values
1744 when a new cipher list is set.
1747 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
1748 ciphers. Before when the 56bit ciphers were enabled the sorting was
1751 The syntax for the cipher sorting has been extended to support sorting by
1752 cipher-strength (using the strength_bits hard coded in the tables).
1753 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
1755 Fix a bug in the cipher-command parser: when supplying a cipher command
1756 string with an "undefined" symbol (neither command nor alphanumeric
1757 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
1758 an error is flagged.
1760 Due to the strength-sorting extension, the code of the
1761 ssl_create_cipher_list() function was completely rearranged. I hope that
1762 the readability was also increased :-)
1763 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
1765 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
1766 for the first serial number and places 2 in the serial number file. This
1767 avoids problems when the root CA is created with serial number zero and
1768 the first user certificate has the same issuer name and serial number
1772 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
1773 the new code. Add documentation for this stuff.
1776 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
1777 X509_*() to X509at_*() on the grounds that they don't handle X509
1778 structures and behave in an analagous way to the X509v3 functions:
1779 they shouldn't be called directly but wrapper functions should be used
1782 So we also now have some wrapper functions that call the X509at functions
1783 when passed certificate requests. (TO DO: similar things can be done with
1784 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
1785 things. Some of these need some d2i or i2d and print functionality
1786 because they handle more complex structures.)
1789 *) Add missing #ifndefs that caused missing symbols when building libssl
1790 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
1791 NO_RSA in ssl/s2*.c.
1792 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
1794 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
1795 has a return value which indicates the quality of the random data
1796 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
1797 error queue. New function RAND_pseudo_bytes() generates output that is
1798 guaranteed to be unique but not unpredictable. RAND_add is like
1799 RAND_seed, but takes an extra argument for an entropy estimate
1800 (RAND_seed always assumes full entropy).
1803 *) Do more iterations of Rabin-Miller probable prime test (specifically,
1804 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
1805 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
1806 in crypto/bn/bn_prime.c for the complete table). This guarantees a
1807 false-positive rate of at most 2^-80 for random input.
1810 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
1813 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
1814 in the 0.9.5 release), this returns the chain
1815 from an X509_CTX structure with a dup of the stack and all
1816 the X509 reference counts upped: so the stack will exist
1817 after X509_CTX_cleanup() has been called. Modify pkcs12.c
1820 Also make SSL_SESSION_print() print out the verify return
1824 *) Add manpage for the pkcs12 command. Also change the default
1825 behaviour so MAC iteration counts are used unless the new
1826 -nomaciter option is used. This improves file security and
1827 only older versions of MSIE (4.0 for example) need it.
1830 *) Honor the no-xxx Configure options when creating .DEF files.
1833 *) Add PKCS#10 attributes to field table: challengePassword,
1834 unstructuredName and unstructuredAddress. These are taken from
1835 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
1836 international characters are used.
1838 More changes to X509_ATTRIBUTE code: allow the setting of types
1839 based on strings. Remove the 'loc' parameter when adding
1840 attributes because these will be a SET OF encoding which is sorted
1844 *) Initial changes to the 'req' utility to allow request generation
1845 automation. This will allow an application to just generate a template
1846 file containing all the field values and have req construct the
1849 Initial support for X509_ATTRIBUTE handling. Stacks of these are
1850 used all over the place including certificate requests and PKCS#7
1851 structures. They are currently handled manually where necessary with
1852 some primitive wrappers for PKCS#7. The new functions behave in a
1853 manner analogous to the X509 extension functions: they allow
1854 attributes to be looked up by NID and added.
1856 Later something similar to the X509V3 code would be desirable to
1857 automatically handle the encoding, decoding and printing of the
1858 more complex types. The string types like challengePassword can
1859 be handled by the string table functions.
1861 Also modified the multi byte string table handling. Now there is
1862 a 'global mask' which masks out certain types. The table itself
1863 can use the flag STABLE_NO_MASK to ignore the mask setting: this
1864 is useful when for example there is only one permissible type
1865 (as in countryName) and using the mask might result in no valid
1869 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
1870 SSL_get_peer_finished to allow applications to obtain the latest
1871 Finished messages sent to the peer or expected from the peer,
1872 respectively. (SSL_get_peer_finished is usually the Finished message
1873 actually received from the peer, otherwise the protocol will be aborted.)
1875 As the Finished message are message digests of the complete handshake
1876 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
1877 be used for external authentication procedures when the authentication
1878 provided by SSL/TLS is not desired or is not enough.
1881 *) Enhanced support for Alpha Linux is added. Now ./config checks if
1882 the host supports BWX extension and if Compaq C is present on the
1883 $PATH. Just exploiting of the BWX extension results in 20-30%
1884 performance kick for some algorithms, e.g. DES and RC4 to mention
1885 a couple. Compaq C in turn generates ~20% faster code for MD5 and
1889 *) Add support for MS "fast SGC". This is arguably a violation of the
1890 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
1891 weak crypto and after checking the certificate is SGC a second one
1892 with strong crypto. MS SGC stops the first handshake after receiving
1893 the server certificate message and sends a second client hello. Since
1894 a server will typically do all the time consuming operations before
1895 expecting any further messages from the client (server key exchange
1896 is the most expensive) there is little difference between the two.
1898 To get OpenSSL to support MS SGC we have to permit a second client
1899 hello message after we have sent server done. In addition we have to
1900 reset the MAC if we do get this second client hello.
1903 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
1904 if a DER encoded private key is RSA or DSA traditional format. Changed
1905 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
1906 format DER encoded private key. Newer code should use PKCS#8 format which
1907 has the key type encoded in the ASN1 structure. Added DER private key
1908 support to pkcs8 application.
1911 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
1912 ciphersuites has been selected (as required by the SSL 3/TLS 1
1913 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
1914 is set, we interpret this as a request to violate the specification
1915 (the worst that can happen is a handshake failure, and 'correct'
1916 behaviour would result in a handshake failure anyway).
1919 *) In SSL_CTX_add_session, take into account that there might be multiple
1920 SSL_SESSION structures with the same session ID (e.g. when two threads
1921 concurrently obtain them from an external cache).
1922 The internal cache can handle only one SSL_SESSION with a given ID,
1923 so if there's a conflict, we now throw out the old one to achieve
1927 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
1928 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
1929 some routines that use cipher OIDs: some ciphers do not have OIDs
1930 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
1934 *) Simplify the trust setting structure and code. Now we just have
1935 two sequences of OIDs for trusted and rejected settings. These will
1936 typically have values the same as the extended key usage extension
1937 and any application specific purposes.
1939 The trust checking code now has a default behaviour: it will just
1940 check for an object with the same NID as the passed id. Functions can
1941 be provided to override either the default behaviour or the behaviour
1942 for a given id. SSL client, server and email already have functions
1943 in place for compatibility: they check the NID and also return "trusted"
1944 if the certificate is self signed.
1947 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
1948 traditional format into an EVP_PKEY structure.
1951 *) Add a password callback function PEM_cb() which either prompts for
1952 a password if usr_data is NULL or otherwise assumes it is a null
1953 terminated password. Allow passwords to be passed on command line
1954 environment or config files in a few more utilities.
1957 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
1958 keys. Add some short names for PKCS#8 PBE algorithms and allow them
1959 to be specified on the command line for the pkcs8 and pkcs12 utilities.
1960 Update documentation.
1963 *) Support for ASN1 "NULL" type. This could be handled before by using
1964 ASN1_TYPE but there wasn't any function that would try to read a NULL
1965 and produce an error if it couldn't. For compatibility we also have
1966 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
1967 don't allocate anything because they don't need to.
1970 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
1972 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
1974 *) Rebuild of the memory allocation routines used by OpenSSL code and
1975 possibly others as well. The purpose is to make an interface that
1976 provide hooks so anyone can build a separate set of allocation and
1977 deallocation routines to be used by OpenSSL, for example memory
1978 pool implementations, or something else, which was previously hard
1979 since Malloc(), Realloc() and Free() were defined as macros having
1980 the values malloc, realloc and free, respectively (except for Win32
1981 compilations). The same is provided for memory debugging code.
1982 OpenSSL already comes with functionality to find memory leaks, but
1983 this gives people a chance to debug other memory problems.
1985 With these changes, a new set of functions and macros have appeared:
1987 CRYPTO_set_mem_debug_functions() [F]
1988 CRYPTO_get_mem_debug_functions() [F]
1989 CRYPTO_dbg_set_options() [F]
1990 CRYPTO_dbg_get_options() [F]
1991 CRYPTO_malloc_debug_init() [M]
1993 The memory debug functions are NULL by default, unless the library
1994 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
1995 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
1996 gives the standard debugging functions that come with OpenSSL) or
1997 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
1998 provided by the library user) must be used. When the standard
1999 debugging functions are used, CRYPTO_dbg_set_options can be used to
2000 request additional information:
2001 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
2002 the CRYPTO_MDEBUG_xxx macro when compiling the library.
2004 Also, things like CRYPTO_set_mem_functions will always give the
2005 expected result (the new set of functions is used for allocation
2006 and deallocation) at all times, regardless of platform and compiler
2009 To finish it up, some functions that were never use in any other
2010 way than through macros have a new API and new semantic:
2013 CRYPTO_dbg_realloc()
2016 All macros of value have retained their old syntax.
2017 [Richard Levitte and Bodo Moeller]
2019 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
2020 ordering of SMIMECapabilities wasn't in "strength order" and there
2021 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
2025 *) Some ASN1 types with illegal zero length encoding (INTEGER,
2026 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
2027 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
2029 *) Merge in my S/MIME library for OpenSSL. This provides a simple
2030 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
2031 functionality to handle multipart/signed properly) and a utility
2032 called 'smime' to call all this stuff. This is based on code I
2033 originally wrote for Celo who have kindly allowed it to be
2034 included in OpenSSL.
2037 *) Add variants des_set_key_checked and des_set_key_unchecked of
2038 des_set_key (aka des_key_sched). Global variable des_check_key
2039 decides which of these is called by des_set_key; this way
2040 des_check_key behaves as it always did, but applications and
2041 the library itself, which was buggy for des_check_key == 1,
2042 have a cleaner way to pick the version they need.
2045 *) New function PKCS12_newpass() which changes the password of a
2049 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
2050 dynamic mix. In both cases the ids can be used as an index into the
2051 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
2052 functions so they accept a list of the field values and the
2053 application doesn't need to directly manipulate the X509_TRUST
2057 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
2061 *) Modify the way the V3 extension code looks up extensions. This now
2062 works in a similar way to the object code: we have some "standard"
2063 extensions in a static table which is searched with OBJ_bsearch()
2064 and the application can add dynamic ones if needed. The file
2065 crypto/x509v3/ext_dat.h now has the info: this file needs to be
2066 updated whenever a new extension is added to the core code and kept
2067 in ext_nid order. There is a simple program 'tabtest.c' which checks
2068 this. New extensions are not added too often so this file can readily
2069 be maintained manually.
2071 There are two big advantages in doing things this way. The extensions
2072 can be looked up immediately and no longer need to be "added" using
2073 X509V3_add_standard_extensions(): this function now does nothing.
2074 [Side note: I get *lots* of email saying the extension code doesn't
2075 work because people forget to call this function]
2076 Also no dynamic allocation is done unless new extensions are added:
2077 so if we don't add custom extensions there is no need to call
2078 X509V3_EXT_cleanup().
2081 *) Modify enc utility's salting as follows: make salting the default. Add a
2082 magic header, so unsalted files fail gracefully instead of just decrypting
2083 to garbage. This is because not salting is a big security hole, so people
2084 should be discouraged from doing it.
2087 *) Fixes and enhancements to the 'x509' utility. It allowed a message
2088 digest to be passed on the command line but it only used this
2089 parameter when signing a certificate. Modified so all relevant
2090 operations are affected by the digest parameter including the
2091 -fingerprint and -x509toreq options. Also -x509toreq choked if a
2092 DSA key was used because it didn't fix the digest.
2095 *) Initial certificate chain verify code. Currently tests the untrusted
2096 certificates for consistency with the verify purpose (which is set
2097 when the X509_STORE_CTX structure is set up) and checks the pathlength.
2099 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
2100 this is because it will reject chains with invalid extensions whereas
2101 every previous version of OpenSSL and SSLeay made no checks at all.
2103 Trust code: checks the root CA for the relevant trust settings. Trust
2104 settings have an initial value consistent with the verify purpose: e.g.
2105 if the verify purpose is for SSL client use it expects the CA to be
2106 trusted for SSL client use. However the default value can be changed to
2107 permit custom trust settings: one example of this would be to only trust
2108 certificates from a specific "secure" set of CAs.
2110 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
2111 which should be used for version portability: especially since the
2112 verify structure is likely to change more often now.
2114 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
2115 to set them. If not set then assume SSL clients will verify SSL servers
2118 Two new options to the verify program: -untrusted allows a set of
2119 untrusted certificates to be passed in and -purpose which sets the
2120 intended purpose of the certificate. If a purpose is set then the
2121 new chain verify code is used to check extension consistency.
2124 *) Support for the authority information access extension.
2127 *) Modify RSA and DSA PEM read routines to transparently handle
2128 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
2129 public keys in a format compatible with certificate
2130 SubjectPublicKeyInfo structures. Unfortunately there were already
2131 functions called *_PublicKey_* which used various odd formats so
2132 these are retained for compatibility: however the DSA variants were
2133 never in a public release so they have been deleted. Changed dsa/rsa
2134 utilities to handle the new format: note no releases ever handled public
2135 keys so we should be OK.
2137 The primary motivation for this change is to avoid the same fiasco
2138 that dogs private keys: there are several incompatible private key
2139 formats some of which are standard and some OpenSSL specific and
2140 require various evil hacks to allow partial transparent handling and
2141 even then it doesn't work with DER formats. Given the option anything
2142 other than PKCS#8 should be dumped: but the other formats have to
2143 stay in the name of compatibility.
2145 With public keys and the benefit of hindsight one standard format
2146 is used which works with EVP_PKEY, RSA or DSA structures: though
2147 it clearly returns an error if you try to read the wrong kind of key.
2149 Added a -pubkey option to the 'x509' utility to output the public key.
2150 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
2151 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
2152 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
2153 that do the same as the EVP_PKEY_assign_*() except they up the
2154 reference count of the added key (they don't "swallow" the
2158 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
2159 CRLs would fail if the file contained no certificates or no CRLs:
2160 added a new function to read in both types and return the number
2161 read: this means that if none are read it will be an error. The
2162 DER versions of the certificate and CRL reader would always fail
2163 because it isn't possible to mix certificates and CRLs in DER format
2164 without choking one or the other routine. Changed this to just read
2165 a certificate: this is the best we can do. Also modified the code
2166 in apps/verify.c to take notice of return codes: it was previously
2167 attempting to read in certificates from NULL pointers and ignoring
2168 any errors: this is one reason why the cert and CRL reader seemed
2169 to work. It doesn't check return codes from the default certificate
2170 routines: these may well fail if the certificates aren't installed.
2173 *) Code to support otherName option in GeneralName.
2176 *) First update to verify code. Change the verify utility
2177 so it warns if it is passed a self signed certificate:
2178 for consistency with the normal behaviour. X509_verify
2179 has been modified to it will now verify a self signed
2180 certificate if *exactly* the same certificate appears
2181 in the store: it was previously impossible to trust a
2182 single self signed certificate. This means that:
2183 openssl verify ss.pem
2184 now gives a warning about a self signed certificate but
2185 openssl verify -CAfile ss.pem ss.pem
2189 *) For servers, store verify_result in SSL_SESSION data structure
2190 (and add it to external session representation).
2191 This is needed when client certificate verifications fails,
2192 but an application-provided verification callback (set by
2193 SSL_CTX_set_cert_verify_callback) allows accepting the session
2194 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
2195 but returns 1): When the session is reused, we have to set
2196 ssl->verify_result to the appropriate error code to avoid
2198 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
2200 *) Fix a bug in the new PKCS#7 code: it didn't consider the
2201 case in PKCS7_dataInit() where the signed PKCS7 structure
2202 didn't contain any existing data because it was being created.
2203 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
2205 *) Add a salt to the key derivation routines in enc.c. This
2206 forms the first 8 bytes of the encrypted file. Also add a
2207 -S option to allow a salt to be input on the command line.
2210 *) New function X509_cmp(). Oddly enough there wasn't a function
2211 to compare two certificates. We do this by working out the SHA1
2212 hash and comparing that. X509_cmp() will be needed by the trust
2216 *) SSL_get1_session() is like SSL_get_session(), but increments
2217 the reference count in the SSL_SESSION returned.
2218 [Geoff Thorpe <geoff@eu.c2.net>]
2220 *) Fix for 'req': it was adding a null to request attributes.
2221 Also change the X509_LOOKUP and X509_INFO code to handle
2222 certificate auxiliary information.
2225 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
2229 *) Add the possibility to add extra information to the memory leak
2230 detecting output, to form tracebacks, showing from where each
2231 allocation was originated: CRYPTO_push_info("constant string") adds
2232 the string plus current file name and line number to a per-thread
2233 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
2234 is like calling CYRPTO_pop_info() until the stack is empty.
2235 Also updated memory leak detection code to be multi-thread-safe.
2238 *) Add options -text and -noout to pkcs7 utility and delete the
2239 encryption options which never did anything. Update docs.
2242 *) Add options to some of the utilities to allow the pass phrase
2243 to be included on either the command line (not recommended on
2244 OSes like Unix) or read from the environment. Update the
2245 manpages and fix a few bugs.
2248 *) Add a few manpages for some of the openssl commands.
2251 *) Fix the -revoke option in ca. It was freeing up memory twice,
2252 leaking and not finding already revoked certificates.
2255 *) Extensive changes to support certificate auxiliary information.
2256 This involves the use of X509_CERT_AUX structure and X509_AUX
2257 functions. An X509_AUX function such as PEM_read_X509_AUX()
2258 can still read in a certificate file in the usual way but it
2259 will also read in any additional "auxiliary information". By
2260 doing things this way a fair degree of compatibility can be
2261 retained: existing certificates can have this information added
2262 using the new 'x509' options.
2264 Current auxiliary information includes an "alias" and some trust
2265 settings. The trust settings will ultimately be used in enhanced
2266 certificate chain verification routines: currently a certificate
2267 can only be trusted if it is self signed and then it is trusted
2271 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
2272 The problem was that one of the replacement routines had not been working
2273 since SSLeay releases. For now the offending routine has been replaced
2274 with non-optimised assembler. Even so, this now gives around 95%
2275 performance improvement for 1024 bit RSA signs.
2278 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
2279 handling. Most clients have the effective key size in bits equal to
2280 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
2281 A few however don't do this and instead use the size of the decrypted key
2282 to determine the RC2 key length and the AlgorithmIdentifier to determine
2283 the effective key length. In this case the effective key length can still
2284 be 40 bits but the key length can be 168 bits for example. This is fixed
2285 by manually forcing an RC2 key into the EVP_PKEY structure because the
2286 EVP code can't currently handle unusual RC2 key sizes: it always assumes
2287 the key length and effective key length are equal.
2290 *) Add a bunch of functions that should simplify the creation of
2291 X509_NAME structures. Now you should be able to do:
2292 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
2293 and have it automatically work out the correct field type and fill in
2294 the structures. The more adventurous can try:
2295 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
2296 and it will (hopefully) work out the correct multibyte encoding.
2299 *) Change the 'req' utility to use the new field handling and multibyte
2300 copy routines. Before the DN field creation was handled in an ad hoc
2301 way in req, ca, and x509 which was rather broken and didn't support
2302 BMPStrings or UTF8Strings. Since some software doesn't implement
2303 BMPStrings or UTF8Strings yet, they can be enabled using the config file
2304 using the dirstring_type option. See the new comment in the default
2305 openssl.cnf for more info.
2308 *) Make crypto/rand/md_rand.c more robust:
2309 - Assure unique random numbers after fork().
2310 - Make sure that concurrent threads access the global counter and
2311 md serializably so that we never lose entropy in them
2312 or use exactly the same state in multiple threads.
2313 Access to the large state is not always serializable because
2314 the additional locking could be a performance killer, and
2315 md should be large enough anyway.
2318 *) New file apps/app_rand.c with commonly needed functionality
2319 for handling the random seed file.
2321 Use the random seed file in some applications that previously did not:
2323 dsaparam -genkey (which also ignored its '-rand' option),
2326 x509 (when signing).
2327 Except on systems with /dev/urandom, it is crucial to have a random
2328 seed file at least for key creation, DSA signing, and for DH exchanges;
2329 for RSA signatures we could do without one.
2331 gendh and gendsa (unlike genrsa) used to read only the first byte
2332 of each file listed in the '-rand' option. The function as previously
2333 found in genrsa is now in app_rand.c and is used by all programs
2334 that support '-rand'.
2337 *) In RAND_write_file, use mode 0600 for creating files;
2338 don't just chmod when it may be too late.
2341 *) Report an error from X509_STORE_load_locations
2342 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
2345 *) New function ASN1_mbstring_copy() this copies a string in either
2346 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
2347 into an ASN1_STRING type. A mask of permissible types is passed
2348 and it chooses the "minimal" type to use or an error if not type
2352 *) Add function equivalents to the various macros in asn1.h. The old
2353 macros are retained with an M_ prefix. Code inside the library can
2354 use the M_ macros. External code (including the openssl utility)
2355 should *NOT* in order to be "shared library friendly".
2358 *) Add various functions that can check a certificate's extensions
2359 to see if it usable for various purposes such as SSL client,
2360 server or S/MIME and CAs of these types. This is currently
2361 VERY EXPERIMENTAL but will ultimately be used for certificate chain
2362 verification. Also added a -purpose flag to x509 utility to
2363 print out all the purposes.
2366 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
2370 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
2371 for, obtain and decode and extension and obtain its critical flag.
2372 This allows all the necessary extension code to be handled in a
2373 single function call.
2376 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
2377 platforms. See crypto/rc4/rc4_enc.c for further details.
2380 *) New -noout option to asn1parse. This causes no output to be produced
2381 its main use is when combined with -strparse and -out to extract data
2382 from a file (which may not be in ASN.1 format).
2385 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
2386 when producing the local key id.
2387 [Richard Levitte <levitte@stacken.kth.se>]
2389 *) New option -dhparam in s_server. This allows a DH parameter file to be
2390 stated explicitly. If it is not stated then it tries the first server
2391 certificate file. The previous behaviour hard coded the filename
2395 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
2396 a public key to be input or output. For example:
2397 openssl rsa -in key.pem -pubout -out pubkey.pem
2398 Also added necessary DSA public key functions to handle this.
2401 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
2402 in the message. This was handled by allowing
2403 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
2404 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
2406 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
2407 to the end of the strings whereas this didn't. This would cause problems
2408 if strings read with d2i_ASN1_bytes() were later modified.
2409 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
2411 *) Fix for base64 decode bug. When a base64 bio reads only one line of
2412 data and it contains EOF it will end up returning an error. This is
2413 caused by input 46 bytes long. The cause is due to the way base64
2414 BIOs find the start of base64 encoded data. They do this by trying a
2415 trial decode on each line until they find one that works. When they
2416 do a flag is set and it starts again knowing it can pass all the
2417 data directly through the decoder. Unfortunately it doesn't reset
2418 the context it uses. This means that if EOF is reached an attempt
2419 is made to pass two EOFs through the context and this causes the
2420 resulting error. This can also cause other problems as well. As is
2421 usual with these problems it takes *ages* to find and the fix is
2422 trivial: move one line.
2423 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
2425 *) Ugly workaround to get s_client and s_server working under Windows. The
2426 old code wouldn't work because it needed to select() on sockets and the
2427 tty (for keypresses and to see if data could be written). Win32 only
2428 supports select() on sockets so we select() with a 1s timeout on the
2429 sockets and then see if any characters are waiting to be read, if none
2430 are present then we retry, we also assume we can always write data to
2431 the tty. This isn't nice because the code then blocks until we've
2432 received a complete line of data and it is effectively polling the
2433 keyboard at 1s intervals: however it's quite a bit better than not
2434 working at all :-) A dedicated Windows application might handle this
2435 with an event loop for example.
2438 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
2439 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
2440 will be called when RSA_sign() and RSA_verify() are used. This is useful
2441 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
2442 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
2443 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
2444 This necessitated the support of an extra signature type NID_md5_sha1
2445 for SSL signatures and modifications to the SSL library to use it instead
2446 of calling RSA_public_decrypt() and RSA_private_encrypt().
2449 *) Add new -verify -CAfile and -CApath options to the crl program, these
2450 will lookup a CRL issuers certificate and verify the signature in a
2451 similar way to the verify program. Tidy up the crl program so it
2452 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
2453 less strict. It will now permit CRL extensions even if it is not
2454 a V2 CRL: this will allow it to tolerate some broken CRLs.
2457 *) Initialize all non-automatic variables each time one of the openssl
2458 sub-programs is started (this is necessary as they may be started
2459 multiple times from the "OpenSSL>" prompt).
2460 [Lennart Bang, Bodo Moeller]
2462 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
2463 removing all other RSA functionality (this is what NO_RSA does). This
2464 is so (for example) those in the US can disable those operations covered
2465 by the RSA patent while allowing storage and parsing of RSA keys and RSA
2469 *) Non-copying interface to BIO pairs.
2470 (still largely untested)
2473 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
2474 ASCII string. This was handled independently in various places before.
2477 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
2478 UTF8 strings a character at a time.
2481 *) Use client_version from client hello to select the protocol
2482 (s23_srvr.c) and for RSA client key exchange verification
2483 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
2486 *) Add various utility functions to handle SPKACs, these were previously
2487 handled by poking round in the structure internals. Added new function
2488 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
2489 print, verify and generate SPKACs. Based on an original idea from
2490 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
2493 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
2496 *) Allow the config file extension section to be overwritten on the
2497 command line. Based on an original idea from Massimiliano Pala
2498 <madwolf@comune.modena.it>. The new option is called -extensions
2499 and can be applied to ca, req and x509. Also -reqexts to override
2500 the request extensions in req and -crlexts to override the crl extensions
2504 *) Add new feature to the SPKAC handling in ca. Now you can include
2505 the same field multiple times by preceding it by "XXXX." for example:
2508 this is the same syntax as used in the req config file.
2511 *) Allow certificate extensions to be added to certificate requests. These
2512 are specified in a 'req_extensions' option of the req section of the
2513 config file. They can be printed out with the -text option to req but
2514 are otherwise ignored at present.
2517 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
2518 data read consists of only the final block it would not decrypted because
2519 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
2520 A misplaced 'break' also meant the decrypted final block might not be
2521 copied until the next read.
2524 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
2525 a few extra parameters to the DH structure: these will be useful if
2526 for example we want the value of 'q' or implement X9.42 DH.
2529 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
2530 provides hooks that allow the default DSA functions or functions on a
2531 "per key" basis to be replaced. This allows hardware acceleration and
2532 hardware key storage to be handled without major modification to the
2533 library. Also added low level modexp hooks and CRYPTO_EX structure and
2534 associated functions.
2537 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
2538 as "read only": it can't be written to and the buffer it points to will
2539 not be freed. Reading from a read only BIO is much more efficient than
2540 a normal memory BIO. This was added because there are several times when
2541 an area of memory needs to be read from a BIO. The previous method was
2542 to create a memory BIO and write the data to it, this results in two
2543 copies of the data and an O(n^2) reading algorithm. There is a new
2544 function BIO_new_mem_buf() which creates a read only memory BIO from
2545 an area of memory. Also modified the PKCS#7 routines to use read only
2549 *) Bugfix: ssl23_get_client_hello did not work properly when called in
2550 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
2551 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
2552 but a retry condition occured while trying to read the rest.
2555 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
2556 NID_pkcs7_encrypted by default: this was wrong since this should almost
2557 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
2558 the encrypted data type: this is a more sensible place to put it and it
2559 allows the PKCS#12 code to be tidied up that duplicated this
2563 *) Changed obj_dat.pl script so it takes its input and output files on
2564 the command line. This should avoid shell escape redirection problems
2568 *) Initial support for certificate extension requests, these are included
2569 in things like Xenroll certificate requests. Included functions to allow
2570 extensions to be obtained and added.
2573 *) -crlf option to s_client and s_server for sending newlines as
2574 CRLF (as required by many protocols).
2577 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
2579 *) Install libRSAglue.a when OpenSSL is built with RSAref.
2580 [Ralf S. Engelschall]
2582 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
2583 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
2585 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
2589 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
2590 DH parameters/keys (q is lost during that conversion, but the resulting
2591 DH parameters contain its length).
2593 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
2594 much faster than DH_generate_parameters (which creates parameters
2595 where p = 2*q + 1), and also the smaller q makes DH computations
2596 much more efficient (160-bit exponentiation instead of 1024-bit
2597 exponentiation); so this provides a convenient way to support DHE
2598 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
2599 utter importance to use
2600 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2602 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
2603 when such DH parameters are used, because otherwise small subgroup
2604 attacks may become possible!
2607 *) Avoid memory leak in i2d_DHparams.
2610 *) Allow the -k option to be used more than once in the enc program:
2611 this allows the same encrypted message to be read by multiple recipients.
2614 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
2615 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
2616 it will always use the numerical form of the OID, even if it has a short
2620 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
2621 method only got called if p,q,dmp1,dmq1,iqmp components were present,
2622 otherwise bn_mod_exp was called. In the case of hardware keys for example
2623 no private key components need be present and it might store extra data
2624 in the RSA structure, which cannot be accessed from bn_mod_exp.
2625 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
2626 private key operations.
2629 *) Added support for SPARC Linux.
2632 *) pem_password_cb function type incompatibly changed from
2633 typedef int pem_password_cb(char *buf, int size, int rwflag);
2635 ....(char *buf, int size, int rwflag, void *userdata);
2636 so that applications can pass data to their callbacks:
2637 The PEM[_ASN1]_{read,write}... functions and macros now take an
2638 additional void * argument, which is just handed through whenever
2639 the password callback is called.
2640 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
2642 New function SSL_CTX_set_default_passwd_cb_userdata.
2644 Compatibility note: As many C implementations push function arguments
2645 onto the stack in reverse order, the new library version is likely to
2646 interoperate with programs that have been compiled with the old
2647 pem_password_cb definition (PEM_whatever takes some data that
2648 happens to be on the stack as its last argument, and the callback
2649 just ignores this garbage); but there is no guarantee whatsoever that
2652 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
2653 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
2654 problems not only on Windows, but also on some Unix platforms.
2655 To avoid problematic command lines, these definitions are now in an
2656 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
2657 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
2660 *) MIPS III/IV assembler module is reimplemented.
2663 *) More DES library cleanups: remove references to srand/rand and
2664 delete an unused file.
2667 *) Add support for the the free Netwide assembler (NASM) under Win32,
2668 since not many people have MASM (ml) and it can be hard to obtain.
2669 This is currently experimental but it seems to work OK and pass all
2670 the tests. Check out INSTALL.W32 for info.
2673 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
2674 without temporary keys kept an extra copy of the server key,
2675 and connections with temporary keys did not free everything in case
2679 *) New function RSA_check_key and new openssl rsa option -check
2680 for verifying the consistency of RSA keys.
2681 [Ulf Moeller, Bodo Moeller]
2683 *) Various changes to make Win32 compile work:
2684 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
2685 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
2686 comparison" warnings.
2687 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
2690 *) Add a debugging option to PKCS#5 v2 key generation function: when
2691 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
2692 derived keys are printed to stderr.
2695 *) Copy the flags in ASN1_STRING_dup().
2696 [Roman E. Pavlov <pre@mo.msk.ru>]
2698 *) The x509 application mishandled signing requests containing DSA
2699 keys when the signing key was also DSA and the parameters didn't match.
2701 It was supposed to omit the parameters when they matched the signing key:
2702 the verifying software was then supposed to automatically use the CA's
2703 parameters if they were absent from the end user certificate.
2705 Omitting parameters is no longer recommended. The test was also
2706 the wrong way round! This was probably due to unusual behaviour in
2707 EVP_cmp_parameters() which returns 1 if the parameters match.
2708 This meant that parameters were omitted when they *didn't* match and
2709 the certificate was useless. Certificates signed with 'ca' didn't have
2711 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
2713 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
2714 The interface is as follows:
2715 Applications can use
2716 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
2717 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
2718 "off" is now the default.
2719 The library internally uses
2720 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
2721 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
2722 to disable memory-checking temporarily.
2724 Some inconsistent states that previously were possible (and were
2725 even the default) are now avoided.
2727 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
2728 with each memory chunk allocated; this is occasionally more helpful
2729 than just having a counter.
2731 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
2733 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
2737 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
2738 which largely parallels "options", but is for changing API behaviour,
2739 whereas "options" are about protocol behaviour.
2740 Initial "mode" flags are:
2742 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
2743 a single record has been written.
2744 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
2745 retries use the same buffer location.
2746 (But all of the contents must be
2750 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
2753 *) Fix problems with no-hmac etc.
2754 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
2756 *) New functions RSA_get_default_method(), RSA_set_method() and
2757 RSA_get_method(). These allows replacement of RSA_METHODs without having
2758 to mess around with the internals of an RSA structure.
2761 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
2762 Also really enable memory leak checks in openssl.c and in some
2764 [Chad C. Mulligan, Bodo Moeller]
2766 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
2767 up the length of negative integers. This has now been simplified to just
2768 store the length when it is first determined and use it later, rather
2769 than trying to keep track of where data is copied and updating it to
2771 [Steve Henson, reported by Brien Wheeler
2772 <bwheeler@authentica-security.com>]
2774 *) Add a new function PKCS7_signatureVerify. This allows the verification
2775 of a PKCS#7 signature but with the signing certificate passed to the
2776 function itself. This contrasts with PKCS7_dataVerify which assumes the
2777 certificate is present in the PKCS#7 structure. This isn't always the
2778 case: certificates can be omitted from a PKCS#7 structure and be
2779 distributed by "out of band" means (such as a certificate database).
2782 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
2783 function prototypes in pem.h, also change util/mkdef.pl to add the
2784 necessary function names.
2787 *) mk1mf.pl (used by Windows builds) did not properly read the
2788 options set by Configure in the top level Makefile, and Configure
2789 was not even able to write more than one option correctly.
2790 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
2793 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
2794 file to be loaded from a BIO or FILE pointer. The BIO version will
2795 for example allow memory BIOs to contain config info.
2798 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
2799 Whoever hopes to achieve shared-library compatibility across versions
2800 must use this, not the compile-time macro.
2801 (Exercise 0.9.4: Which is the minimum library version required by
2803 Note: All this applies only to multi-threaded programs, others don't
2807 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
2808 through a BIO pair triggered the default case, i.e.
2809 SSLerr(...,SSL_R_UNKNOWN_STATE).
2812 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
2813 can use the SSL library even if none of the specific BIOs is
2817 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
2818 for the encoded length.
2819 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
2821 *) Add initial documentation of the X509V3 functions.
2824 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
2825 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
2826 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
2827 secure PKCS#8 private key format with a high iteration count.
2830 *) Fix determination of Perl interpreter: A perl or perl5
2831 _directory_ in $PATH was also accepted as the interpreter.
2832 [Ralf S. Engelschall]
2834 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
2835 wrong with it but it was very old and did things like calling
2836 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
2840 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
2841 to use the new extension code.
2844 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
2845 with macros. This should make it easier to change their form, add extra
2846 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
2850 *) Add to configuration table a new entry that can specify an alternative
2851 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
2852 according to Mark Crispin <MRC@Panda.COM>.
2856 *) DES CBC did not update the IV. Weird.
2859 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
2860 Changing the behaviour of the former might break existing programs --
2861 where IV updating is needed, des_ncbc_encrypt can be used.
2864 *) When bntest is run from "make test" it drives bc to check its
2865 calculations, as well as internally checking them. If an internal check
2866 fails, it needs to cause bc to give a non-zero result or make test carries
2867 on without noticing the failure. Fixed.
2870 *) DES library cleanups.
2873 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
2874 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
2875 ciphers. NOTE: although the key derivation function has been verified
2876 against some published test vectors it has not been extensively tested
2877 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
2881 *) Instead of "mkdir -p", which is not fully portable, use new
2882 Perl script "util/mkdir-p.pl".
2885 *) Rewrite the way password based encryption (PBE) is handled. It used to
2886 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
2887 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
2888 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
2889 the 'parameter' field of the AlgorithmIdentifier is passed to the
2890 underlying key generation function so it must do its own ASN1 parsing.
2891 This has also changed the EVP_PBE_CipherInit() function which now has a
2892 'parameter' argument instead of literal salt and iteration count values
2893 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
2896 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
2897 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
2898 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
2899 KEY" because this clashed with PKCS#8 unencrypted string. Since this
2900 value was just used as a "magic string" and not used directly its
2901 value doesn't matter.
2904 *) Introduce some semblance of const correctness to BN. Shame C doesn't
2908 *) "linux-sparc64" configuration (ultrapenguin).
2909 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
2910 "linux-sparc" configuration.
2911 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
2913 *) config now generates no-xxx options for missing ciphers.
2916 *) Support the EBCDIC character set (work in progress).
2917 File ebcdic.c not yet included because it has a different license.
2918 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2920 *) Support BS2000/OSD-POSIX.
2921 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2923 *) Make callbacks for key generation use void * instead of char *.
2926 *) Make S/MIME samples compile (not yet tested).
2929 *) Additional typesafe stacks.
2932 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
2936 Changes between 0.9.3 and 0.9.3a [29 May 1999]
2938 *) New configuration variant "sco5-gcc".
2940 *) Updated some demos.
2941 [Sean O Riordain, Wade Scholine]
2943 *) Add missing BIO_free at exit of pkcs12 application.
2946 *) Fix memory leak in conf.c.
2949 *) Updates for Win32 to assembler version of MD5.
2952 *) Set #! path to perl in apps/der_chop to where we found it
2953 instead of using a fixed path.
2956 *) SHA library changes for irix64-mips4-cc.
2959 *) Improvements for VMS support.
2963 Changes between 0.9.2b and 0.9.3 [24 May 1999]
2965 *) Bignum library bug fix. IRIX 6 passes "make test" now!
2966 This also avoids the problems with SC4.2 and unpatched SC5.
2967 [Andy Polyakov <appro@fy.chalmers.se>]
2969 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
2970 These are required because of the typesafe stack would otherwise break
2971 existing code. If old code used a structure member which used to be STACK
2972 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
2973 sk_num or sk_value it would produce an error because the num, data members
2974 are not present in STACK_OF. Now it just produces a warning. sk_set
2975 replaces the old method of assigning a value to sk_value
2976 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
2977 that does this will no longer work (and should use sk_set instead) but
2978 this could be regarded as a "questionable" behaviour anyway.
2981 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
2982 correctly handle encrypted S/MIME data.
2985 *) Change type of various DES function arguments from des_cblock
2986 (which means, in function argument declarations, pointer to char)
2987 to des_cblock * (meaning pointer to array with 8 char elements),
2988 which allows the compiler to do more typechecking; it was like
2989 that back in SSLeay, but with lots of ugly casts.
2991 Introduce new type const_des_cblock.
2994 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
2995 problems: find RecipientInfo structure that matches recipient certificate
2996 and initialise the ASN1 structures properly based on passed cipher.
2999 *) Belatedly make the BN tests actually check the results.
3002 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
3003 to and from BNs: it was completely broken. New compilation option
3004 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
3005 key elements as negative integers.
3008 *) Reorganize and speed up MD5.
3009 [Andy Polyakov <appro@fy.chalmers.se>]
3012 [Richard Levitte <richard@levitte.org>]
3014 *) New option -out to asn1parse to allow the parsed structure to be
3015 output to a file. This is most useful when combined with the -strparse
3016 option to examine the output of things like OCTET STRINGS.
3019 *) Make SSL library a little more fool-proof by not requiring any longer
3020 that SSL_set_{accept,connect}_state be called before
3021 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
3022 in many applications because usually everything *appeared* to work as
3023 intended anyway -- now it really works as intended).
3026 *) Move openssl.cnf out of lib/.
3029 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
3030 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
3031 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
3032 [Ralf S. Engelschall]
3034 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
3035 handle PKCS#7 enveloped data properly.
3036 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
3038 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
3039 copying pointers. The cert_st handling is changed by this in
3040 various ways (and thus what used to be known as ctx->default_cert
3041 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
3042 any longer when s->cert does not give us what we need).
3043 ssl_cert_instantiate becomes obsolete by this change.
3044 As soon as we've got the new code right (possibly it already is?),
3045 we have solved a couple of bugs of the earlier code where s->cert
3046 was used as if it could not have been shared with other SSL structures.
3048 Note that using the SSL API in certain dirty ways now will result
3049 in different behaviour than observed with earlier library versions:
3050 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
3051 does not influence s as it used to.
3053 In order to clean up things more thoroughly, inside SSL_SESSION
3054 we don't use CERT any longer, but a new structure SESS_CERT
3055 that holds per-session data (if available); currently, this is
3056 the peer's certificate chain and, for clients, the server's certificate
3057 and temporary key. CERT holds only those values that can have
3058 meaningful defaults in an SSL_CTX.
3061 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
3062 from the internal representation. Various PKCS#7 fixes: remove some
3063 evil casts and set the enc_dig_alg field properly based on the signing
3067 *) Allow PKCS#12 password to be set from the command line or the
3068 environment. Let 'ca' get its config file name from the environment
3069 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
3073 *) Allow certificate policies extension to use an IA5STRING for the
3074 organization field. This is contrary to the PKIX definition but
3075 VeriSign uses it and IE5 only recognises this form. Document 'x509'
3079 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
3080 without disallowing inline assembler and the like for non-pedantic builds.
3083 *) Support Borland C++ builder.
3084 [Janez Jere <jj@void.si>, modified by Ulf Möller]
3089 *) SHA-1 cleanups and performance enhancements.
3090 [Andy Polyakov <appro@fy.chalmers.se>]
3092 *) Sparc v8plus assembler for the bignum library.
3093 [Andy Polyakov <appro@fy.chalmers.se>]
3095 *) Accept any -xxx and +xxx compiler options in Configure.
3098 *) Update HPUX configuration.
3101 *) Add missing sk_<type>_unshift() function to safestack.h
3102 [Ralf S. Engelschall]
3104 *) New function SSL_CTX_use_certificate_chain_file that sets the
3105 "extra_cert"s in addition to the certificate. (This makes sense
3106 only for "PEM" format files, as chains as a whole are not
3110 *) Support verify_depth from the SSL API.
3111 x509_vfy.c had what can be considered an off-by-one-error:
3112 Its depth (which was not part of the external interface)
3113 was actually counting the number of certificates in a chain;
3114 now it really counts the depth.
3117 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
3118 instead of X509err, which often resulted in confusing error
3119 messages since the error codes are not globally unique
3120 (e.g. an alleged error in ssl3_accept when a certificate
3121 didn't match the private key).
3123 *) New function SSL_CTX_set_session_id_context that allows to set a default
3124 value (so that you don't need SSL_set_session_id_context for each
3125 connection using the SSL_CTX).
3128 *) OAEP decoding bug fix.
3131 *) Support INSTALL_PREFIX for package builders, as proposed by
3135 *) New Configure options "threads" and "no-threads". For systems
3136 where the proper compiler options are known (currently Solaris
3137 and Linux), "threads" is the default.
3140 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
3143 *) Install various scripts to $(OPENSSLDIR)/misc, not to
3144 $(INSTALLTOP)/bin -- they shouldn't clutter directories
3145 such as /usr/local/bin.
3148 *) "make linux-shared" to build shared libraries.
3149 [Niels Poppe <niels@netbox.org>]
3151 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
3154 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
3155 extension adding in x509 utility.
3158 *) Remove NOPROTO sections and error code comments.
3161 *) Partial rewrite of the DEF file generator to now parse the ANSI
3165 *) New Configure options --prefix=DIR and --openssldir=DIR.
3168 *) Complete rewrite of the error code script(s). It is all now handled
3169 by one script at the top level which handles error code gathering,
3170 header rewriting and C source file generation. It should be much better
3171 than the old method: it now uses a modified version of Ulf's parser to
3172 read the ANSI prototypes in all header files (thus the old K&R definitions
3173 aren't needed for error creation any more) and do a better job of
3174 translating function codes into names. The old 'ASN1 error code imbedded
3175 in a comment' is no longer necessary and it doesn't use .err files which
3176 have now been deleted. Also the error code call doesn't have to appear all
3177 on one line (which resulted in some large lines...).
3180 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
3183 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
3184 0 (which usually indicates a closed connection), but continue reading.
3187 *) Fix some race conditions.
3190 *) Add support for CRL distribution points extension. Add Certificate
3191 Policies and CRL distribution points documentation.
3194 *) Move the autogenerated header file parts to crypto/opensslconf.h.
3197 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
3198 8 of keying material. Merlin has also confirmed interop with this fix
3199 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
3200 [Merlin Hughes <merlin@baltimore.ie>]
3202 *) Fix lots of warnings.
3203 [Richard Levitte <levitte@stacken.kth.se>]
3205 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
3206 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
3207 [Richard Levitte <levitte@stacken.kth.se>]
3209 *) Fix problems with sizeof(long) == 8.
3210 [Andy Polyakov <appro@fy.chalmers.se>]
3212 *) Change functions to ANSI C.
3215 *) Fix typos in error codes.
3216 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
3218 *) Remove defunct assembler files from Configure.
3221 *) SPARC v8 assembler BIGNUM implementation.
3222 [Andy Polyakov <appro@fy.chalmers.se>]
3224 *) Support for Certificate Policies extension: both print and set.
3225 Various additions to support the r2i method this uses.
3228 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
3229 return a const string when you are expecting an allocated buffer.
3232 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
3233 types DirectoryString and DisplayText.
3236 *) Add code to allow r2i extensions to access the configuration database,
3237 add an LHASH database driver and add several ctx helper functions.
3240 *) Fix an evil bug in bn_expand2() which caused various BN functions to
3241 fail when they extended the size of a BIGNUM.
3244 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
3245 support typesafe stack.
3248 *) Fix typo in SSL_[gs]et_options().
3249 [Nils Frostberg <nils@medcom.se>]
3251 *) Delete various functions and files that belonged to the (now obsolete)
3252 old X509V3 handling code.
3255 *) New Configure option "rsaref".
3258 *) Don't auto-generate pem.h.
3261 *) Introduce type-safe ASN.1 SETs.
3264 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
3265 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
3267 *) Introduce type-safe STACKs. This will almost certainly break lots of code
3268 that links with OpenSSL (well at least cause lots of warnings), but fear
3269 not: the conversion is trivial, and it eliminates loads of evil casts. A
3270 few STACKed things have been converted already. Feel free to convert more.
3271 In the fullness of time, I'll do away with the STACK type altogether.
3274 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
3275 specified in <certfile> by updating the entry in the index.txt file.
3276 This way one no longer has to edit the index.txt file manually for
3277 revoking a certificate. The -revoke option does the gory details now.
3278 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
3280 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
3281 `-text' option at all and this way the `-noout -text' combination was
3282 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
3283 [Ralf S. Engelschall]
3285 *) Make sure a corresponding plain text error message exists for the
3286 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
3287 verify callback function determined that a certificate was revoked.
3288 [Ralf S. Engelschall]
3290 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
3291 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
3292 all available cipers including rc5, which was forgotten until now.
3293 In order to let the testing shell script know which algorithms
3294 are available, a new (up to now undocumented) command
3295 "openssl list-cipher-commands" is used.
3298 *) Bugfix: s_client occasionally would sleep in select() when
3299 it should have checked SSL_pending() first.
3302 *) New functions DSA_do_sign and DSA_do_verify to provide access to
3303 the raw DSA values prior to ASN.1 encoding.
3306 *) Tweaks to Configure
3307 [Niels Poppe <niels@netbox.org>]
3309 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
3313 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
3316 *) New config option to avoid instructions that are illegal on the 80386.
3317 The default code is faster, but requires at least a 486.
3320 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
3321 SSL2_SERVER_VERSION (not used at all) macros, which are now the
3322 same as SSL2_VERSION anyway.
3325 *) New "-showcerts" option for s_client.
3328 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
3329 application. Various cleanups and fixes.
3332 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
3333 modify error routines to work internally. Add error codes and PBE init
3334 to library startup routines.
3337 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
3338 packing functions to asn1 and evp. Changed function names and error
3339 codes along the way.
3342 *) PKCS12 integration: and so it begins... First of several patches to
3343 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
3344 objects to objects.h
3347 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
3348 and display support for Thawte strong extranet extension.
3351 *) Add LinuxPPC support.
3352 [Jeff Dubrule <igor@pobox.org>]
3354 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
3355 bn_div_words in alpha.s.
3356 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
3358 *) Make sure the RSA OAEP test is skipped under -DRSAref because
3359 OAEP isn't supported when OpenSSL is built with RSAref.
3360 [Ulf Moeller <ulf@fitug.de>]
3362 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
3363 so they no longer are missing under -DNOPROTO.
3364 [Soren S. Jorvang <soren@t.dk>]
3367 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
3369 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
3370 doesn't work when the session is reused. Coming soon!
3373 *) Fix a security hole, that allows sessions to be reused in the wrong
3374 context thus bypassing client cert protection! All software that uses
3375 client certs and session caches in multiple contexts NEEDS PATCHING to
3376 allow session reuse! A fuller solution is in the works.
3377 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
3379 *) Some more source tree cleanups (removed obsolete files
3380 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
3381 permission on "config" script to be executable) and a fix for the INSTALL
3383 [Ulf Moeller <ulf@fitug.de>]
3385 *) Remove some legacy and erroneous uses of malloc, free instead of
3387 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
3389 *) Make rsa_oaep_test return non-zero on error.
3390 [Ulf Moeller <ulf@fitug.de>]
3392 *) Add support for native Solaris shared libraries. Configure
3393 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
3394 if someone would make that last step automatic.
3395 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
3397 *) ctx_size was not built with the right compiler during "make links". Fixed.
3400 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
3401 except NULL ciphers". This means the default cipher list will no longer
3402 enable NULL ciphers. They need to be specifically enabled e.g. with
3403 the string "DEFAULT:eNULL".
3406 *) Fix to RSA private encryption routines: if p < q then it would
3407 occasionally produce an invalid result. This will only happen with
3408 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
3411 *) Be less restrictive and allow also `perl util/perlpath.pl
3412 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
3413 because this way one can also use an interpreter named `perl5' (which is
3414 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
3415 installed as `perl').
3416 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3418 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
3419 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3421 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
3422 advapi32.lib to Win32 build and change the pem test comparision
3423 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
3424 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
3425 and crypto/des/ede_cbcm_enc.c.
3428 *) DES quad checksum was broken on big-endian architectures. Fixed.
3431 *) Comment out two functions in bio.h that aren't implemented. Fix up the
3432 Win32 test batch file so it (might) work again. The Win32 test batch file
3433 is horrible: I feel ill....
3436 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
3437 in e_os.h. Audit of header files to check ANSI and non ANSI
3438 sections: 10 functions were absent from non ANSI section and not exported
3439 from Windows DLLs. Fixed up libeay.num for new functions.
3442 *) Make `openssl version' output lines consistent.
3443 [Ralf S. Engelschall]
3445 *) Fix Win32 symbol export lists for BIO functions: Added
3446 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
3447 to ms/libeay{16,32}.def.
3448 [Ralf S. Engelschall]
3450 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
3451 fine under Unix and passes some trivial tests I've now added. But the
3452 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
3453 added to make sure no one expects that this stuff really works in the
3454 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
3455 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
3457 [Ralf S. Engelschall]
3459 *) Fix the generation of two part addresses in perl.
3460 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
3462 *) Add config entry for Linux on MIPS.
3463 [John Tobey <jtobey@channel1.com>]
3465 *) Make links whenever Configure is run, unless we are on Windoze.
3468 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
3469 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
3473 *) Add a useful kludge to allow package maintainers to specify compiler and
3474 other platforms details on the command line without having to patch the
3475 Configure script everytime: One now can use ``perl Configure
3476 <id>:<details>'', i.e. platform ids are allowed to have details appended
3477 to them (seperated by colons). This is treated as there would be a static
3478 pre-configured entry in Configure's %table under key <id> with value
3479 <details> and ``perl Configure <id>'' is called. So, when you want to
3480 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
3481 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
3482 now, which overrides the FreeBSD-elf entry on-the-fly.
3483 [Ralf S. Engelschall]
3485 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
3488 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
3489 on the `perl Configure ...' command line. This way one can compile
3490 OpenSSL libraries with Position Independent Code (PIC) which is needed
3491 for linking it into DSOs.
3492 [Ralf S. Engelschall]
3494 *) Remarkably, export ciphers were totally broken and no-one had noticed!
3498 *) Cleaned up the LICENSE document: The official contact for any license
3499 questions now is the OpenSSL core team under openssl-core@openssl.org.
3500 And add a paragraph about the dual-license situation to make sure people
3501 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
3502 to the OpenSSL toolkit.
3503 [Ralf S. Engelschall]
3505 *) General source tree makefile cleanups: Made `making xxx in yyy...'
3506 display consistent in the source tree and replaced `/bin/rm' by `rm'.
3507 Additonally cleaned up the `make links' target: Remove unnecessary
3508 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
3509 to speed processing and no longer clutter the display with confusing
3510 stuff. Instead only the actually done links are displayed.
3511 [Ralf S. Engelschall]
3513 *) Permit null encryption ciphersuites, used for authentication only. It used
3514 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
3515 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
3519 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
3520 signed attributes when verifying signatures (this would break them),
3521 the detached data encoding was wrong and public keys obtained using
3522 X509_get_pubkey() weren't freed.
3525 *) Add text documentation for the BUFFER functions. Also added a work around
3526 to a Win95 console bug. This was triggered by the password read stuff: the
3527 last character typed gets carried over to the next fread(). If you were
3528 generating a new cert request using 'req' for example then the last
3529 character of the passphrase would be CR which would then enter the first
3533 *) Added the new `Includes OpenSSL Cryptography Software' button as
3534 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
3535 button and can be used by applications based on OpenSSL to show the
3536 relationship to the OpenSSL project.
3537 [Ralf S. Engelschall]
3539 *) Remove confusing variables in function signatures in files
3540 ssl/ssl_lib.c and ssl/ssl.h.
3541 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3543 *) Don't install bss_file.c under PREFIX/include/
3544 [Lennart Bong <lob@kulthea.stacken.kth.se>]
3546 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
3547 functions that return function pointers and has support for NT specific
3548 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
3549 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
3550 unsigned to signed types: this was killing the Win32 compile.
3553 *) Add new certificate file to stack functions,
3554 SSL_add_dir_cert_subjects_to_stack() and
3555 SSL_add_file_cert_subjects_to_stack(). These largely supplant
3556 SSL_load_client_CA_file(), and can be used to add multiple certs easily
3557 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
3558 This means that Apache-SSL and similar packages don't have to mess around
3559 to add as many CAs as they want to the preferred list.
3562 *) Experiment with doxygen documentation. Currently only partially applied to
3564 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
3565 openssl.doxy as the configuration file.
3568 *) Get rid of remaining C++-style comments which strict C compilers hate.
3569 [Ralf S. Engelschall, pointed out by Carlos Amengual]
3571 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
3572 compiled in by default: it has problems with large keys.
3575 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
3576 DH private keys and/or callback functions which directly correspond to
3577 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
3578 is needed for applications which have to configure certificates on a
3579 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
3581 For the RSA certificate situation is makes no difference, but
3582 for the DSA certificate situation this fixes the "no shared cipher"
3583 problem where the OpenSSL cipher selection procedure failed because the
3584 temporary keys were not overtaken from the context and the API provided
3585 no way to reconfigure them.
3586 The new functions now let applications reconfigure the stuff and they
3587 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
3588 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
3589 non-public-API function ssl_cert_instantiate() is used as a helper
3590 function and also to reduce code redundancy inside ssl_rsa.c.
3591 [Ralf S. Engelschall]
3593 *) Move s_server -dcert and -dkey options out of the undocumented feature
3594 area because they are useful for the DSA situation and should be
3595 recognized by the users.
3596 [Ralf S. Engelschall]
3598 *) Fix the cipher decision scheme for export ciphers: the export bits are
3599 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
3600 SSL_EXP_MASK. So, the original variable has to be used instead of the
3601 already masked variable.
3602 [Richard Levitte <levitte@stacken.kth.se>]
3604 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
3605 [Richard Levitte <levitte@stacken.kth.se>]
3607 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
3608 from `int' to `unsigned int' because it's a length and initialized by
3609 EVP_DigestFinal() which expects an `unsigned int *'.
3610 [Richard Levitte <levitte@stacken.kth.se>]
3612 *) Don't hard-code path to Perl interpreter on shebang line of Configure
3613 script. Instead use the usual Shell->Perl transition trick.
3614 [Ralf S. Engelschall]
3616 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
3617 (in addition to RSA certificates) to match the behaviour of `openssl dsa
3618 -noout -modulus' as it's already the case for `openssl rsa -noout
3619 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
3620 currently the public key is printed (a decision which was already done by
3621 `openssl dsa -modulus' in the past) which serves a similar purpose.
3622 Additionally the NO_RSA no longer completely removes the whole -modulus
3623 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
3625 [Ralf S. Engelschall]
3627 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
3628 BIO. See the source (crypto/evp/bio_ok.c) for more info.
3629 [Arne Ansper <arne@ats.cyber.ee>]
3631 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
3632 to be added. Now both 'req' and 'ca' can use new objects defined in the
3636 *) Add cool BIO that does syslog (or event log on NT).
3637 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
3639 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
3640 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
3641 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
3642 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
3645 *) Add preliminary config info for new extension code.
3648 *) Make RSA_NO_PADDING really use no padding.
3649 [Ulf Moeller <ulf@fitug.de>]
3651 *) Generate errors when private/public key check is done.
3654 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
3655 for some CRL extensions and new objects added.
3658 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
3659 key usage extension and fuller support for authority key id.
3662 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
3663 padding method for RSA, which is recommended for new applications in PKCS
3664 #1 v2.0 (RFC 2437, October 1998).
3665 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
3666 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
3667 against Bleichbacher's attack on RSA.
3668 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
3671 *) Updates to the new SSL compression code
3672 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3674 *) Fix so that the version number in the master secret, when passed
3675 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
3676 (because the server will not accept higher), that the version number
3677 is 0x03,0x01, not 0x03,0x00
3678 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3680 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
3681 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
3682 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
3685 *) Support for RAW extensions where an arbitrary extension can be
3686 created by including its DER encoding. See apps/openssl.cnf for
3690 *) Make sure latest Perl versions don't interpret some generated C array
3691 code as Perl array code in the crypto/err/err_genc.pl script.
3692 [Lars Weber <3weber@informatik.uni-hamburg.de>]
3694 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
3695 not many people have the assembler. Various Win32 compilation fixes and
3696 update to the INSTALL.W32 file with (hopefully) more accurate Win32
3700 *) Modify configure script 'Configure' to automatically create crypto/date.h
3701 file under Win32 and also build pem.h from pem.org. New script
3702 util/mkfiles.pl to create the MINFO file on environments that can't do a
3703 'make files': perl util/mkfiles.pl >MINFO should work.
3706 *) Major rework of DES function declarations, in the pursuit of correctness
3707 and purity. As a result, many evil casts evaporated, and some weirdness,
3708 too. You may find this causes warnings in your code. Zapping your evil
3709 casts will probably fix them. Mostly.
3712 *) Fix for a typo in asn1.h. Bug fix to object creation script
3713 obj_dat.pl. It considered a zero in an object definition to mean
3714 "end of object": none of the objects in objects.h have any zeros
3715 so it wasn't spotted.
3716 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
3718 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
3719 Masking (CBCM). In the absence of test vectors, the best I have been able
3720 to do is check that the decrypt undoes the encrypt, so far. Send me test
3721 vectors if you have them.
3724 *) Correct calculation of key length for export ciphers (too much space was
3725 allocated for null ciphers). This has not been tested!
3728 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
3729 message is now correct (it understands "crypto" and "ssl" on its
3730 command line). There is also now an "update" option. This will update
3731 the util/ssleay.num and util/libeay.num files with any new functions.
3733 perl util/mkdef.pl crypto ssl update
3734 it will update them.
3737 *) Overhauled the Perl interface (perl/*):
3738 - ported BN stuff to OpenSSL's different BN library
3739 - made the perl/ source tree CVS-aware
3740 - renamed the package from SSLeay to OpenSSL (the files still contain
3741 their history because I've copied them in the repository)
3742 - removed obsolete files (the test scripts will be replaced
3743 by better Test::Harness variants in the future)
3744 [Ralf S. Engelschall]
3746 *) First cut for a very conservative source tree cleanup:
3747 1. merge various obsolete readme texts into doc/ssleay.txt
3748 where we collect the old documents and readme texts.
3749 2. remove the first part of files where I'm already sure that we no
3750 longer need them because of three reasons: either they are just temporary
3751 files which were left by Eric or they are preserved original files where
3752 I've verified that the diff is also available in the CVS via "cvs diff
3753 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
3754 the crypto/md/ stuff).
3755 [Ralf S. Engelschall]
3757 *) More extension code. Incomplete support for subject and issuer alt
3758 name, issuer and authority key id. Change the i2v function parameters
3759 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
3760 what that's for :-) Fix to ASN1 macro which messed up
3761 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
3764 *) Preliminary support for ENUMERATED type. This is largely copied from the
3768 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
3769 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3771 *) Make sure `make rehash' target really finds the `openssl' program.
3772 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
3774 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
3775 like to hear about it if this slows down other processors.
3778 *) Add CygWin32 platform information to Configure script.
3779 [Alan Batie <batie@aahz.jf.intel.com>]
3781 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
3782 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
3784 *) New program nseq to manipulate netscape certificate sequences
3787 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
3791 *) Fixes to BN code. Previously the default was to define BN_RECURSION
3792 but the BN code had some problems that would cause failures when
3793 doing certificate verification and some other functions.
3794 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
3796 *) Add ASN1 and PEM code to support netscape certificate sequences.
3799 *) Add ASN1 and PEM code to support netscape certificate sequences.
3802 *) Add several PKIX and private extended key usage OIDs.
3805 *) Modify the 'ca' program to handle the new extension code. Modify
3806 openssl.cnf for new extension format, add comments.
3809 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
3810 and add a sample to openssl.cnf so req -x509 now adds appropriate
3814 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
3815 error code, add initial support to X509_print() and x509 application.
3818 *) Takes a deep breath and start addding X509 V3 extension support code. Add
3819 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
3820 stuff is currently isolated and isn't even compiled yet.
3823 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
3824 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
3825 Removed the versions check from X509 routines when loading extensions:
3826 this allows certain broken certificates that don't set the version
3827 properly to be processed.
3830 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
3831 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
3832 can still be regenerated with "make depend".
3835 *) Spelling mistake in C version of CAST-128.
3836 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
3838 *) Changes to the error generation code. The perl script err-code.pl
3839 now reads in the old error codes and retains the old numbers, only
3840 adding new ones if necessary. It also only changes the .err files if new
3841 codes are added. The makefiles have been modified to only insert errors
3842 when needed (to avoid needlessly modifying header files). This is done
3843 by only inserting errors if the .err file is newer than the auto generated
3844 C file. To rebuild all the error codes from scratch (the old behaviour)
3845 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
3846 or delete all the .err files.
3849 *) CAST-128 was incorrectly implemented for short keys. The C version has
3850 been fixed, but is untested. The assembler versions are also fixed, but
3851 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
3852 to regenerate it if needed.
3853 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
3854 Hagino <itojun@kame.net>]
3856 *) File was opened incorrectly in randfile.c.
3857 [Ulf Möller <ulf@fitug.de>]
3859 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
3860 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
3861 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
3862 al: it's just almost always a UTCTime. Note this patch adds new error
3863 codes so do a "make errors" if there are problems.
3866 *) Correct Linux 1 recognition in config.
3867 [Ulf Möller <ulf@fitug.de>]
3869 *) Remove pointless MD5 hash when using DSA keys in ca.
3870 [Anonymous <nobody@replay.com>]
3872 *) Generate an error if given an empty string as a cert directory. Also
3873 generate an error if handed NULL (previously returned 0 to indicate an
3874 error, but didn't set one).
3875 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
3877 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
3880 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
3881 parameters. This was causing a warning which killed off the Win32 compile.
3884 *) Remove C++ style comments from crypto/bn/bn_local.h.
3885 [Neil Costigan <neil.costigan@celocom.com>]
3887 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
3888 based on a text string, looking up short and long names and finally
3889 "dot" format. The "dot" format stuff didn't work. Added new function
3890 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
3891 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
3892 OID is not part of the table.
3895 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
3896 X509_LOOKUP_by_alias().
3899 *) Sort openssl functions by name.
3902 *) Get the gendsa program working (hopefully) and add it to app list. Remove
3903 encryption from sample DSA keys (in case anyone is interested the password
3907 *) Make _all_ *_free functions accept a NULL pointer.
3908 [Frans Heymans <fheymans@isaserver.be>]
3910 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
3912 [Anonymous <nobody@replay.com>]
3914 *) s_server should send the CAfile as acceptable CAs, not its own cert.
3915 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3917 *) Don't blow it for numeric -newkey arguments to apps/req.
3918 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
3920 *) Temp key "for export" tests were wrong in s3_srvr.c.
3921 [Anonymous <nobody@replay.com>]
3923 *) Add prototype for temp key callback functions
3924 SSL_CTX_set_tmp_{rsa,dh}_callback().
3927 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
3928 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
3931 *) X509_name_add_entry() freed the wrong thing after an error.
3932 [Arne Ansper <arne@ats.cyber.ee>]
3934 *) rsa_eay.c would attempt to free a NULL context.
3935 [Arne Ansper <arne@ats.cyber.ee>]
3937 *) BIO_s_socket() had a broken should_retry() on Windoze.
3938 [Arne Ansper <arne@ats.cyber.ee>]
3940 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
3941 [Arne Ansper <arne@ats.cyber.ee>]
3943 *) Make sure the already existing X509_STORE->depth variable is initialized
3944 in X509_STORE_new(), but document the fact that this variable is still
3945 unused in the certificate verification process.
3946 [Ralf S. Engelschall]
3948 *) Fix the various library and apps files to free up pkeys obtained from
3949 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
3952 *) Fix reference counting in X509_PUBKEY_get(). This makes
3953 demos/maurice/example2.c work, amongst others, probably.
3954 [Steve Henson and Ben Laurie]
3956 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
3957 `openssl' and second, the shortcut symlinks for the `openssl <command>'
3958 are no longer created. This way we have a single and consistent command
3959 line interface `openssl <command>', similar to `cvs <command>'.
3960 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
3962 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
3963 BIT STRING wrapper always have zero unused bits.
3966 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
3969 *) Make the top-level INSTALL documentation easier to understand.
3972 *) Makefiles updated to exit if an error occurs in a sub-directory
3973 make (including if user presses ^C) [Paul Sutton]
3975 *) Make Montgomery context stuff explicit in RSA data structure.
3978 *) Fix build order of pem and err to allow for generated pem.h.
3981 *) Fix renumbering bug in X509_NAME_delete_entry().
3984 *) Enhanced the err-ins.pl script so it makes the error library number
3985 global and can add a library name. This is needed for external ASN1 and
3986 other error libraries.
3989 *) Fixed sk_insert which never worked properly.
3992 *) Fix ASN1 macros so they can handle indefinite length construted
3993 EXPLICIT tags. Some non standard certificates use these: they can now
3997 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
3998 into a single doc/ssleay.txt bundle. This way the information is still
3999 preserved but no longer messes up this directory. Now it's new room for
4000 the new set of documenation files.
4001 [Ralf S. Engelschall]
4003 *) SETs were incorrectly DER encoded. This was a major pain, because they
4004 shared code with SEQUENCEs, which aren't coded the same. This means that
4005 almost everything to do with SETs or SEQUENCEs has either changed name or
4006 number of arguments.
4007 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
4009 *) Fix test data to work with the above.
4012 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
4013 was already fixed by Eric for 0.9.1 it seems.
4014 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
4016 *) Autodetect FreeBSD3.
4019 *) Fix various bugs in Configure. This affects the following platforms:
4023 unixware-2.0-pentium
4027 *) Eliminate generated files from CVS. Reorder tests to regenerate files
4028 before they are needed.
4031 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
4035 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
4037 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
4038 changed SSLeay to OpenSSL in version strings.
4039 [Ralf S. Engelschall]
4041 *) Some fixups to the top-level documents.
4044 *) Fixed the nasty bug where rsaref.h was not found under compile-time
4045 because the symlink to include/ was missing.
4046 [Ralf S. Engelschall]
4048 *) Incorporated the popular no-RSA/DSA-only patches
4049 which allow to compile a RSA-free SSLeay.
4050 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
4052 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
4053 when "ssleay" is still not found.
4054 [Ralf S. Engelschall]
4056 *) Added more platforms to Configure: Cray T3E, HPUX 11,
4057 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
4059 *) Updated the README file.
4060 [Ralf S. Engelschall]
4062 *) Added various .cvsignore files in the CVS repository subdirs
4063 to make a "cvs update" really silent.
4064 [Ralf S. Engelschall]
4066 *) Recompiled the error-definition header files and added
4067 missing symbols to the Win32 linker tables.
4068 [Ralf S. Engelschall]
4070 *) Cleaned up the top-level documents;
4071 o new files: CHANGES and LICENSE
4072 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
4073 o merged COPYRIGHT into LICENSE
4074 o removed obsolete TODO file
4075 o renamed MICROSOFT to INSTALL.W32
4076 [Ralf S. Engelschall]
4078 *) Removed dummy files from the 0.9.1b source tree:
4079 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
4080 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
4081 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
4082 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
4083 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
4084 [Ralf S. Engelschall]
4086 *) Added various platform portability fixes.
4089 *) The Genesis of the OpenSSL rpject:
4090 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
4091 Young and Tim J. Hudson created while they were working for C2Net until
4093 [The OpenSSL Project]
4096 Changes between 0.9.0b and 0.9.1b [not released]
4098 *) Updated a few CA certificates under certs/
4101 *) Changed some BIGNUM api stuff.
4104 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
4105 DGUX x86, Linux Alpha, etc.
4108 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
4109 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
4113 *) Add -strparse option to asn1pars program which parses nested
4115 [Dr Stephen Henson <shenson@bigfoot.com>]
4117 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
4120 *) DSA fix for "ca" program.
4123 *) Added "-genkey" option to "dsaparam" program.
4126 *) Added RIPE MD160 (rmd160) message digest.
4129 *) Added -a (all) option to "ssleay version" command.
4132 *) Added PLATFORM define which is the id given to Configure.
4135 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
4138 *) Extended the ASN.1 parser routines.
4141 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
4144 *) Added a BN_CTX to the BN library.
4147 *) Fixed the weak key values in DES library
4150 *) Changed API in EVP library for cipher aliases.
4153 *) Added support for RC2/64bit cipher.
4156 *) Converted the lhash library to the crypto/mem.c functions.
4159 *) Added more recognized ASN.1 object ids.
4162 *) Added more RSA padding checks for SSL/TLS.
4165 *) Added BIO proxy/filter functionality.
4168 *) Added extra_certs to SSL_CTX which can be used
4169 send extra CA certificates to the client in the CA cert chain sending
4170 process. It can be configured with SSL_CTX_add_extra_chain_cert().
4173 *) Now Fortezza is denied in the authentication phase because
4174 this is key exchange mechanism is not supported by SSLeay at all.
4177 *) Additional PKCS1 checks.
4180 *) Support the string "TLSv1" for all TLS v1 ciphers.
4183 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
4184 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
4187 *) Fixed a few memory leaks.
4190 *) Fixed various code and comment typos.
4193 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
4194 bytes sent in the client random.
4195 [Edward Bishop <ebishop@spyglass.com>]