5 Changes between 0.9.6 and 0.9.7 [xx XXX 2001]
7 OpenSSL 0.9.6a/0.9.6b (bugfix releases, 5 Apr 2001 and 9 July 2001)
8 and OpenSSL 0.9.7 were developed in parallel, based on OpenSSL 0.9.6.
10 Change log entries are tagged as follows:
11 -) applies to 0.9.6a/0.9.6b/0.9.6c only
12 *) applies to 0.9.6a/0.9.6b/0.9.6c and 0.9.7
13 +) applies to 0.9.7 only
15 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
16 faced with a pathologically small ClientHello fragment that does
17 not contain client_version: Instead of aborting with an error,
18 simply choose the highest available protocol version (i.e.,
19 TLS 1.0 unless it is disabled). In practice, ClientHello
20 messages are never sent like this, but this change gives us
21 strictly correct behaviour at least for TLS.
24 +) Change all functions with names starting with des_ to be starting
25 with DES_ instead. This because there are increasing clashes with
26 libdes and other des libraries that are currently used by other
27 projects. The old libdes interface is provided, as well as crypt(),
28 if openssl/des_old.h is included. Note that crypt() is no longer
29 declared in openssl/des.h.
31 NOTE: This is a major break of an old API into a new one. Software
32 authors are encouraged to switch to the DES_ style functions. Some
33 time in the future, des_old.h and the libdes compatibility functions
34 will be completely removed.
37 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
38 never resets s->method to s->ctx->method when called from within
39 one of the SSL handshake functions.
40 [Bodo Moeller; problem pointed out by Niko Baric]
42 +) Test for certificates which contain unsupported critical extensions.
43 If such a certificate is found during a verify operation it is
44 rejected by default: this behaviour can be overridden by either
45 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
46 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
47 X509_supported_extension() has also been added which returns 1 if a
48 particular extension is supported.
51 +) New functions/macros
53 SSL_CTX_set_msg_callback(ctx, cb)
54 SSL_CTX_set_msg_callback_arg(ctx, arg)
55 SSL_set_msg_callback(ssl, cb)
56 SSL_set_msg_callback_arg(ssl, arg)
58 to request calling a callback function
60 void cb(int write_p, int version, int content_type,
61 const void *buf, size_t len, SSL *ssl, void *arg)
63 whenever a protocol message has been completely received
64 (write_p == 0) or sent (write_p == 1). Here 'version' is the
65 protocol version according to which the SSL library interprets
66 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
67 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
68 the content type as defined in the SSL 3.0/TLS 1.0 protocol
69 specification (change_cipher_spec(20), alert(21), handshake(22)).
70 'buf' and 'len' point to the actual message, 'ssl' to the
71 SSL object, and 'arg' is the application-defined value set by
72 SSL[_CTX]_set_msg_callback_arg().
74 'openssl s_client' and 'openssl s_server' have new '-msg' options
75 to enable a callback that displays all protocol messages.
77 TODO: SSL 2.0, doc/ssl/, doc/apps/
80 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
81 (sent using the client's version number) if client_version is
82 smaller than the protocol version in use. Also change
83 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
84 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
85 the client will at least see that alert.
88 +) Modify the behaviour of EVP cipher functions in similar way to digests
89 to retain compatibility with existing code.
92 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
93 compatibility with existing code. In particular the 'ctx' parameter does
94 not have to be to be initialized before the call to EVP_DigestInit() and
95 it is tidied up after a call to EVP_DigestFinal(). New function
96 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
97 EVP_MD_CTX_copy() changed to not require the destination to be
98 initialized valid and new function EVP_MD_CTX_copy_ex() added which
99 requires the destination to be valid.
101 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
102 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
105 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
106 so that complete 'Handshake' protocol structures are kept in memory
107 instead of overwriting 'msg_type' and 'length' with 'body' data.
110 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
114 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
115 [Massimo Santin via Richard Levitte]
117 +) Major restructuring to the underlying ENGINE code. This includes
118 reduction of linker bloat, separation of pure "ENGINE" manipulation
119 (initialisation, etc) from functionality dealing with implementations
120 of specific crypto iterfaces. This change also introduces integrated
121 support for symmetric ciphers and digest implementations - so ENGINEs
122 can now accelerate these by providing EVP_CIPHER and EVP_MD
123 implementations of their own. This is detailed in crypto/engine/README
124 as it couldn't be adequately described here. However, there are a few
125 API changes worth noting - some RSA, DSA, DH, and RAND functions that
126 were changed in the original introduction of ENGINE code have now
127 reverted back - the hooking from this code to ENGINE is now a good
128 deal more passive and at run-time, operations deal directly with
129 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
130 dereferencing through an ENGINE pointer any more. Also, the ENGINE
131 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
132 they were not being used by the framework as there is no concept of a
133 BIGNUM_METHOD and they could not be generalised to the new
134 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
135 ENGINE_cpy() has been removed as it cannot be consistently defined in
139 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
142 +) Change mkdef.pl to sort symbols that get the same entry number,
143 and make sure the automatically generated functions ERR_load_*
144 become part of libeay.num as well.
147 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
148 client receives HelloRequest while in a handshake.
149 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
151 +) New function SSL_renegotiate_pending(). This returns true once
152 renegotiation has been requested (either SSL_renegotiate() call
153 or HelloRequest/ClientHello receveived from the peer) and becomes
154 false once a handshake has been completed.
155 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
156 sends a HelloRequest, but does not ensure that a handshake takes
157 place. SSL_renegotiate_pending() is useful for checking if the
158 client has followed the request.)
161 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
162 By default, clients may request session resumption even during
163 renegotiation (if session ID contexts permit); with this option,
164 session resumption is possible only in the first handshake.
167 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
168 should end in 'break', not 'goto end' which circuments various
169 cleanups done in state SSL_ST_OK. But session related stuff
170 must be disabled for SSL_ST_OK in the case that we just sent a
173 Also avoid some overhead by not calling ssl_init_wbio_buffer()
174 before just sending a HelloRequest.
175 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
177 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
178 reveal whether illegal block cipher padding was found or a MAC
179 verification error occured. (Neither SSLerr() codes nor alerts
180 are directly visible to potential attackers, but the information
181 may leak via logfiles.)
183 Similar changes are not required for the SSL 2.0 implementation
184 because the number of padding bytes is sent in clear for SSL 2.0,
185 and the extra bytes are just ignored. However ssl/s2_pkt.c
186 failed to verify that the purported number of padding bytes is in
190 +) Add some demos for certificate and certificate request creation.
193 +) Make maximum certificate chain size accepted from the peer application
194 settable (SSL*_get/set_max_cert_list()), as proposed by
195 "Douglas E. Engert" <deengert@anl.gov>.
198 +) Add support for shared libraries for Unixware-7 and support including
199 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
202 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
203 'wristwatch attack' using huge encoding parameters (cf.
204 James H. Manger's CRYPTO 2001 paper). Note that the
205 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
206 encoding parameters and hence was not vulnerable.
209 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
210 be done prior to destruction. Use this to unload error strings from
211 ENGINEs that load their own error strings. NB: This adds two new API
212 functions to "get" and "set" this destroy handler in an ENGINE.
215 +) Alter all existing ENGINE implementations (except "openssl" and
216 "openbsd") to dynamically instantiate their own error strings. This
217 makes them more flexible to be built both as statically-linked ENGINEs
218 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
219 Also, add stub code to each that makes building them as self-contained
220 shared-libraries easier (see README.ENGINE).
223 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
224 implementations into applications that are completely implemented in
225 self-contained shared-libraries. The "dynamic" ENGINE exposes control
226 commands that can be used to configure what shared-library to load and
227 to control aspects of the way it is handled. Also, made an update to
228 the README.ENGINE file that brings its information up-to-date and
229 provides some information and instructions on the "dynamic" ENGINE
230 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
234 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
236 *) Make it possible to unload ranges of ERR strings with a new
237 "ERR_unload_strings" function.
240 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
241 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
242 followed by modular reduction.
243 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
245 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
246 equivalent based on BN_pseudo_rand() instead of BN_rand().
249 +) Add a copy() function to EVP_MD.
252 +) Make EVP_MD routines take a context pointer instead of just the
253 md_data void pointer.
256 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
257 that the digest can only process a single chunk of data
258 (typically because it is provided by a piece of
259 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
260 is only going to provide a single chunk of data, and hence the
261 framework needn't accumulate the data for oneshot drivers.
264 +) As with "ERR", make it possible to replace the underlying "ex_data"
265 functions. This change also alters the storage and management of global
266 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
267 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
268 index counters. The API functions that use this state have been changed
269 to take a "class_index" rather than pointers to the class's local STACK
270 and counter, and there is now an API function to dynamically create new
271 classes. This centralisation allows us to (a) plug a lot of the
272 thread-safety problems that existed, and (b) makes it possible to clean
273 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
274 such data would previously have always leaked in application code and
275 workarounds were in place to make the memory debugging turn a blind eye
276 to it. Application code that doesn't use this new function will still
277 leak as before, but their memory debugging output will announce it now
278 rather than letting it slide.
280 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
281 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
282 has a return value to indicate success or failure.
285 +) Make it possible to replace the underlying "ERR" functions such that the
286 global state (2 LHASH tables and 2 locks) is only used by the "default"
287 implementation. This change also adds two functions to "get" and "set"
288 the implementation prior to it being automatically set the first time
289 any other ERR function takes place. Ie. an application can call "get",
290 pass the return value to a module it has just loaded, and that module
291 can call its own "set" function using that value. This means the
292 module's "ERR" operations will use (and modify) the error state in the
293 application and not in its own statically linked copy of OpenSSL code.
296 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
297 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
298 the operation, and provides a more encapsulated way for external code
299 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
300 to use these functions rather than manually incrementing the counts.
302 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
305 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
306 This function was broken, as the check for a new client hello message
307 to handle SGC did not allow these large messages.
308 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
311 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
314 +) Add EVP test program.
317 +) Add symmetric cipher support to ENGINE. Expect the API to change!
320 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
321 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
322 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
323 These allow a CRL to be built without having to access X509_CRL fields
324 directly. Modify 'ca' application to use new functions.
327 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
328 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
331 *) Rework the configuration and shared library support for Tru64 Unix.
332 The configuration part makes use of modern compiler features and
333 still retains old compiler behavior for those that run older versions
334 of the OS. The shared library support part includes a variant that
335 uses the RPATH feature, and is available through the special
336 configuration target "alpha-cc-rpath", which will never be selected
338 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
340 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
341 with the same message size as in ssl3_get_certificate_request().
342 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
343 messages might inadvertently be reject as too long.
344 [Petr Lampa <lampa@fee.vutbr.cz>]
346 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
347 bug workarounds. Rollback attack detection is a security feature.
348 The problem will only arise on OpenSSL servers when TLSv1 is not
349 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
350 Software authors not wanting to support TLSv1 will have special reasons
351 for their choice and can explicitly enable this option.
352 [Bodo Moeller, Lutz Jaenicke]
354 +) Rationalise EVP so it can be extended: don't include a union of
355 cipher/digest structures, add init/cleanup functions. This also reduces
356 the number of header dependencies.
359 +) Make DES key schedule conform to the usual scheme, as well as
360 correcting its structure. This means that calls to DES functions
361 now have to pass a pointer to a des_key_schedule instead of a
362 plain des_key_schedule (which was actually always a pointer
366 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
369 *) Modified SSL library such that the verify_callback that has been set
370 specificly for an SSL object with SSL_set_verify() is actually being
371 used. Before the change, a verify_callback set with this function was
372 ignored and the verify_callback() set in the SSL_CTX at the time of
373 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
374 to allow the necessary settings.
377 +) Initial reduction of linker bloat: the use of some functions, such as
378 PEM causes large amounts of unused functions to be linked in due to
379 poor organisation. For example pem_all.c contains every PEM function
380 which has a knock on effect of linking in large amounts of (unused)
381 ASN1 code. Grouping together similar functions and splitting unrelated
382 functions prevents this.
385 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
386 explicitly to NULL, as at least on Solaris 8 this seems not always to be
387 done automatically (in contradiction to the requirements of the C
388 standard). This made problems when used from OpenSSH.
391 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
392 dh->length and always used
394 BN_rand_range(priv_key, dh->p).
396 BN_rand_range() is not necessary for Diffie-Hellman, and this
397 specific range makes Diffie-Hellman unnecessarily inefficient if
398 dh->length (recommended exponent length) is much smaller than the
399 length of dh->p. We could use BN_rand_range() if the order of
400 the subgroup was stored in the DH structure, but we only have
405 BN_rand(priv_key, l, ...)
407 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
413 RSA_eay_public_encrypt
414 RSA_eay_private_decrypt
415 RSA_eay_private_encrypt (signing)
416 RSA_eay_public_decrypt (signature verification)
418 (default implementations for RSA_public_encrypt,
419 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
420 always reject numbers >= n.
423 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
424 to synchronize access to 'locking_thread'. This is necessary on
425 systems where access to 'locking_thread' (an 'unsigned long'
426 variable) is not atomic.
429 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
430 *before* setting the 'crypto_lock_rand' flag. The previous code had
431 a race condition if 0 is a valid thread ID.
432 [Travis Vitek <vitek@roguewave.com>]
434 +) Cleanup of EVP macros.
437 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
441 +) Add initial OCSP responder support to ocsp application. The
442 revocation information is handled using the text based index
443 use by the ca application. The responder can either handle
444 requests generated internally, supplied in files (for example
445 via a CGI script) or using an internal minimal server.
448 +) Add configuration choices to get zlib compression for TLS.
451 +) Changes to Kerberos SSL for RFC 2712 compliance:
452 1. Implemented real KerberosWrapper, instead of just using
453 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
454 2. Implemented optional authenticator field of KerberosWrapper.
456 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
457 and authenticator structs; see crypto/krb5/.
459 Generalized Kerberos calls to support multiple Kerberos libraries.
460 [Vern Staats <staatsvr@asc.hpc.mil>,
461 Jeffrey Altman <jaltman@columbia.edu>
464 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
465 already does with RSA. testdsa.h now has 'priv_key/pub_key'
466 values for each of the key sizes rather than having just
467 parameters (and 'speed' generating keys each time).
470 -) OpenSSL 0.9.6b released [9 July 2001]
472 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
473 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
474 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
475 PRNG state recovery was possible based on the output of
476 one PRNG request appropriately sized to gain knowledge on
477 'md' followed by enough consecutive 1-byte PRNG requests
478 to traverse all of 'state'.
480 1. When updating 'md_local' (the current thread's copy of 'md')
481 during PRNG output generation, hash all of the previous
482 'md_local' value, not just the half used for PRNG output.
484 2. Make the number of bytes from 'state' included into the hash
485 independent from the number of PRNG bytes requested.
487 The first measure alone would be sufficient to avoid
488 Markku-Juhani's attack. (Actually it had never occurred
489 to me that the half of 'md_local' used for chaining was the
490 half from which PRNG output bytes were taken -- I had always
491 assumed that the secret half would be used.) The second
492 measure makes sure that additional data from 'state' is never
493 mixed into 'md_local' in small portions; this heuristically
494 further strengthens the PRNG.
497 +) Speed up EVP routines.
500 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
501 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
502 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
503 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
505 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
506 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
507 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
510 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
512 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
515 *) Fix crypto/bn/asm/mips3.s.
518 *) When only the key is given to "enc", the IV is undefined. Print out
519 an error message in this case.
522 +) Added the OS2-EMX target.
523 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
525 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
526 to support NCONF routines in extension code. New function CONF_set_nconf()
527 to allow functions which take an NCONF to also handle the old LHASH
528 structure: this means that the old CONF compatible routines can be
529 retained (in particular wrt extensions) without having to duplicate the
530 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
533 *) Handle special case when X509_NAME is empty in X509 printing routines.
536 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
537 positive and less than q.
540 +) Enhance the general user interface with mechanisms for inner control
541 and with possibilities to have yes/no kind of prompts.
544 +) Change all calls to low level digest routines in the library and
545 applications to use EVP. Add missing calls to HMAC_cleanup() and
546 don't assume HMAC_CTX can be copied using memcpy().
547 [Verdon Walker <VWalker@novell.com>, Steve Henson]
549 +) Add the possibility to control engines through control names but with
550 arbitrary arguments instead of just a string.
551 Change the key loaders to take a UI_METHOD instead of a callback
552 function pointer. NOTE: this breaks binary compatibility with earlier
553 versions of OpenSSL [engine].
554 Adapt the nCipher code for these new conditions and add a card insertion
558 +) Enhance the general user interface with mechanisms to better support
559 dialog box interfaces, application-defined prompts, the possibility
560 to use defaults (for example default passwords from somewhere else)
561 and interrupts/cancellations.
564 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
565 used: it isn't thread safe and the add_lock_callback should handle
567 [Paul Rose <Paul.Rose@bridge.com>]
569 *) Verify that incoming data obeys the block size in
570 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
573 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
574 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
578 [Ulf Möller, Bodo Möller]
580 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
581 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
582 when fixing the server behaviour for backwards-compatible 'client
583 hello' messages. (Note that the attack is impractical against
584 SSL 3.0 and TLS 1.0 anyway because length and version checking
585 means that the probability of guessing a valid ciphertext is
586 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
589 Before 0.9.5, the countermeasure (hide the error by generating a
590 random 'decryption result') did not work properly because
591 ERR_clear_error() was missing, meaning that SSL_get_error() would
592 detect the supposedly ignored error.
594 Both problems are now fixed.
597 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
598 (previously it was 1024).
601 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
602 tidy up some unnecessarily weird code in 'sk_new()').
603 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
605 +) Change the key loading routines for ENGINEs to use the same kind
606 callback (pem_password_cb) as all other routines that need this
610 *) Fix for compatibility mode trust settings: ignore trust settings
611 unless some valid trust or reject settings are present.
614 *) Fix for blowfish EVP: its a variable length cipher.
617 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
618 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
619 than this minimum value is recommended.
622 +) New random seeder for OpenVMS, using the system process statistics
623 that are easily reachable.
626 +) Windows apparently can't transparently handle global
627 variables defined in DLLs. Initialisations such as:
629 const ASN1_ITEM *it = &ASN1_INTEGER_it;
631 wont compile. This is used by the any applications that need to
632 declare their own ASN1 modules. This was fixed by adding the option
633 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
634 needed for static libraries under Win32.
637 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
638 setting of purpose and trust fields. New X509_STORE trust and
639 purpose functions and tidy up setting in other SSL functions.
642 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
643 structure. These are inherited by X509_STORE_CTX when it is
644 initialised. This allows various defaults to be set in the
645 X509_STORE structure (such as flags for CRL checking and custom
646 purpose or trust settings) for functions which only use X509_STORE_CTX
647 internally such as S/MIME.
649 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
650 trust settings if they are not set in X509_STORE. This allows X509_STORE
651 purposes and trust (in S/MIME for example) to override any set by default.
653 Add command line options for CRL checking to smime, s_client and s_server
657 +) Initial CRL based revocation checking. If the CRL checking flag(s)
658 are set then the CRL is looked up in the X509_STORE structure and
659 its validity and signature checked, then if the certificate is found
660 in the CRL the verify fails with a revoked error.
662 Various new CRL related callbacks added to X509_STORE_CTX structure.
664 Command line options added to 'verify' application to support this.
666 This needs some additional work, such as being able to handle multiple
667 CRLs with different times, extension based lookup (rather than just
668 by subject name) and ultimately more complete V2 CRL extension
672 +) Add a general user interface API (crypto/ui/). This is designed
673 to replace things like des_read_password and friends (backward
674 compatibility functions using this new API are provided).
675 The purpose is to remove prompting functions from the DES code
676 section as well as provide for prompting through dialog boxes in
677 a window system and the like.
680 *) Fix various bugs related to DSA S/MIME verification. Handle missing
681 parameters in DSA public key structures and return an error in the
682 DSA routines if parameters are absent.
685 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
686 in the current directory if neither $RANDFILE nor $HOME was set.
687 RAND_file_name() in 0.9.6a returned NULL in this case. This has
688 caused some confusion to Windows users who haven't defined $HOME.
689 Thus RAND_file_name() is changed again: e_os.h can define a
690 DEFAULT_HOME, which will be used if $HOME is not set.
691 For Windows, we use "C:"; on other platforms, we still require
692 environment variables.
694 +) Add "ex_data" support to ENGINE so implementations can add state at a
695 per-structure level rather than having to store it globally.
698 +) Make it possible for ENGINE structures to be copied when retrieved by
699 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
700 This causes the "original" ENGINE structure to act like a template,
701 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
702 operational state can be localised to each ENGINE structure, despite the
703 fact they all share the same "methods". New ENGINE structures returned in
704 this case have no functional references and the return value is the single
705 structural reference. This matches the single structural reference returned
706 by ENGINE_by_id() normally, when it is incremented on the pre-existing
710 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
711 needs to match any other type at all we need to manually clear the
715 +) Changes to the "openssl engine" utility to include;
716 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
717 about an ENGINE's available control commands.
718 - executing control commands from command line arguments using the
719 '-pre' and '-post' switches. '-post' is only used if '-t' is
720 specified and the ENGINE is successfully initialised. The syntax for
721 the individual commands are colon-separated, for example;
722 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
725 +) New dynamic control command support for ENGINEs. ENGINEs can now
726 declare their own commands (numbers), names (strings), descriptions,
727 and input types for run-time discovery by calling applications. A
728 subset of these commands are implicitly classed as "executable"
729 depending on their input type, and only these can be invoked through
730 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
731 can be based on user input, config files, etc). The distinction is
732 that "executable" commands cannot return anything other than a boolean
733 result and can only support numeric or string input, whereas some
734 discoverable commands may only be for direct use through
735 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
736 pointers, or other custom uses. The "executable" commands are to
737 support parameterisations of ENGINE behaviour that can be
738 unambiguously defined by ENGINEs and used consistently across any
739 OpenSSL-based application. Commands have been added to all the
740 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
741 control over shared-library paths without source code alterations.
744 +) Changed all ENGINE implementations to dynamically allocate their
745 ENGINEs rather than declaring them statically. Apart from this being
746 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
747 this also allows the implementations to compile without using the
748 internal engine_int.h header.
751 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
752 'const' value. Any code that should be able to modify a RAND_METHOD
753 should already have non-const pointers to it (ie. they should only
754 modify their own ones).
757 +) Made a variety of little tweaks to the ENGINE code.
758 - "atalla" and "ubsec" string definitions were moved from header files
759 to C code. "nuron" string definitions were placed in variables
760 rather than hard-coded - allowing parameterisation of these values
761 later on via ctrl() commands.
762 - Removed unused "#if 0"'d code.
763 - Fixed engine list iteration code so it uses ENGINE_free() to release
764 structural references.
765 - Constified the RAND_METHOD element of ENGINE structures.
766 - Constified various get/set functions as appropriate and added
767 missing functions (including a catch-all ENGINE_cpy that duplicates
768 all ENGINE values onto a new ENGINE except reference counts/state).
769 - Removed NULL parameter checks in get/set functions. Setting a method
770 or function to NULL is a way of cancelling out a previously set
771 value. Passing a NULL ENGINE parameter is just plain stupid anyway
772 and doesn't justify the extra error symbols and code.
773 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
774 flags from engine_int.h to engine.h.
775 - Changed prototypes for ENGINE handler functions (init(), finish(),
776 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
779 *) Move 'if (!initialized) RAND_poll()' into regions protected by
780 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
781 having multiple threads call RAND_poll() concurrently.
784 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
785 combination of a flag and a thread ID variable.
786 Otherwise while one thread is in ssleay_rand_bytes (which sets the
787 flag), *other* threads can enter ssleay_add_bytes without obeying
788 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
789 that they do not hold after the first thread unsets add_do_not_lock).
792 +) Implement binary inversion algorithm for BN_mod_inverse in addition
793 to the algorithm using long division. The binary algorithm can be
794 used only if the modulus is odd. On 32-bit systems, it is faster
795 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
796 roughly 5-15% for 256-bit moduli), so we use it only for moduli
797 up to 450 bits. In 64-bit environments, the binary algorithm
798 appears to be advantageous for much longer moduli; here we use it
799 for moduli up to 2048 bits.
802 *) Change bctest again: '-x' expressions are not available in all
806 -) OpenSSL 0.9.6a released [5 Apr 2001]
808 *) Fix a couple of memory leaks in PKCS7_dataDecode()
809 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
811 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
812 the default extension for executables, if any. Also, make the perl
813 scripts that use symlink() to test if it really exists and use "cp"
814 if it doesn't. All this made OpenSSL compilable and installable in
818 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
819 could not support the combine flag in choice fields.
822 -) Fix for asn1_GetSequence() for indefinite length constructed data.
823 If SEQUENCE is length is indefinite just set c->slen to the total
824 amount of data available.
825 [Steve Henson, reported by shige@FreeBSD.org]
826 [This change does not apply to 0.9.7.]
828 *) Change bctest to avoid here-documents inside command substitution
829 (workaround for FreeBSD /bin/sh bug).
830 For compatibility with Ultrix, avoid shell functions (introduced
831 in the bctest version that searches along $PATH).
834 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
835 with des_encrypt() defined on some operating systems, like Solaris
839 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
840 On the Importance of Eliminating Errors in Cryptographic
841 Computations, J. Cryptology 14 (2001) 2, 101-119,
842 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
845 *) MIPS assembler BIGNUM division bug fix.
848 *) Disabled incorrect Alpha assembler code.
851 -) Fix PKCS#7 decode routines so they correctly update the length
852 after reading an EOC for the EXPLICIT tag.
854 [This change does not apply to 0.9.7.]
856 *) Fix bug in PKCS#12 key generation routines. This was triggered
857 if a 3DES key was generated with a 0 initial byte. Include
858 PKCS12_BROKEN_KEYGEN compilation option to retain the old
859 (but broken) behaviour.
862 *) Enhance bctest to search for a working bc along $PATH and print
864 [Tim Rice <tim@multitalents.net> via Richard Levitte]
866 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
867 extensions from a certificate request to the certificate.
870 +) Allow multiple 'certopt' and 'nameopt' options to be separated
871 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
872 file: this allows the display of the certificate about to be
873 signed to be customised, to allow certain fields to be included
874 or excluded and extension details. The old system didn't display
875 multicharacter strings properly, omitted fields not in the policy
876 and couldn't display additional details such as extensions.
879 *) Fix memory leaks in err.c: free err_data string if necessary;
880 don't write to the wrong index in ERR_set_error_data.
883 +) Function EC_POINTs_mul for simultaneous scalar multiplication
884 of an arbitrary number of elliptic curve points, optionally
885 including the generator defined for the EC_GROUP.
886 EC_POINT_mul is a simple wrapper function for the typical case
887 that the point list has just one item (besides the optional
891 +) First EC_METHODs for curves over GF(p):
893 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
894 operations and provides various method functions that can also
895 operate with faster implementations of modular arithmetic.
897 EC_GFp_mont_method() reuses most functions that are part of
898 EC_GFp_simple_method, but uses Montgomery arithmetic.
900 [Bodo Moeller; point addition and point doubling
901 implementation directly derived from source code provided by
902 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
904 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
907 Curves are EC_GROUP objects (with an optional group generator)
908 based on EC_METHODs that are built into the library.
910 Points are EC_POINT objects based on EC_GROUP objects.
912 Most of the framework would be able to handle curves over arbitrary
913 finite fields, but as there are no obvious types for fields other
914 than GF(p), some functions are limited to that for now.
917 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
918 that the file contains a complete HTTP response.
921 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
922 change the def and num file printf format specifier from "%-40sXXX"
923 to "%-39s XXX". The latter will always guarantee a space after the
924 field while the former will cause them to run together if the field
925 is 40 of more characters long.
928 +) Constify the cipher and digest 'method' functions and structures
929 and modify related functions to take constant EVP_MD and EVP_CIPHER
933 *) Implement ssl23_peek (analogous to ssl23_read), which previously
937 *) Replace rdtsc with _emit statements for VC++ version 5.
938 [Jeremy Cooper <jeremy@baymoo.org>]
940 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
941 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
944 +) Modify EVP_Digest*() routines so they now return values. Although the
945 internal software routines can never fail additional hardware versions
949 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
951 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
952 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
955 ERR_R_NESTED_ASN1_ERROR
957 ERR_R_MISSING_ASN1_EOS
958 were 4 .. 9, conflicting with
959 ERR_LIB_RSA (= ERR_R_RSA_LIB)
961 ERR_LIB_PEM (= ERR_R_PEM_LIB).
962 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
964 Add new error code 'ERR_R_INTERNAL_ERROR'.
967 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
971 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
972 sets the subject name for a new request or supersedes the
973 subject name in a given request. Formats that can be parsed are
974 'CN=Some Name, OU=myOU, C=IT'
976 'CN=Some Name/OU=myOU/C=IT'.
978 Add options '-batch' and '-verbose' to 'openssl req'.
979 [Massimiliano Pala <madwolf@hackmasters.net>]
981 *) Make it possible to reuse SSLv2 sessions.
984 +) Introduce the possibility to access global variables through
985 functions on platform were that's the best way to handle exporting
986 global variables in shared libraries. To enable this functionality,
987 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
988 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
989 is normally done by Configure or something similar).
991 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
992 in the source file (foo.c) like this:
994 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
995 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
997 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
998 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1000 OPENSSL_DECLARE_GLOBAL(int,foo);
1001 #define foo OPENSSL_GLOBAL_REF(foo)
1002 OPENSSL_DECLARE_GLOBAL(double,bar);
1003 #define bar OPENSSL_GLOBAL_REF(bar)
1005 The #defines are very important, and therefore so is including the
1006 header file everywhere where the defined globals are used.
1008 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1009 of ASN.1 items, but that structure is a bit different.
1011 The largest change is in util/mkdef.pl which has been enhanced with
1012 better and easier to understand logic to choose which symbols should
1013 go into the Windows .def files as well as a number of fixes and code
1014 cleanup (among others, algorithm keywords are now sorted
1015 lexicographically to avoid constant rewrites).
1018 *) In copy_email() check for >= 0 as a return value for
1019 X509_NAME_get_index_by_NID() since 0 is a valid index.
1020 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1022 +) In BN_div() keep a copy of the sign of 'num' before writing the
1023 result to 'rm' because if rm==num the value will be overwritten
1024 and produce the wrong result if 'num' is negative: this caused
1025 problems with BN_mod() and BN_nnmod().
1028 +) Function OCSP_request_verify(). This checks the signature on an
1029 OCSP request and verifies the signer certificate. The signer
1030 certificate is just checked for a generic purpose and OCSP request
1034 +) Add OCSP_check_validity() function to check the validity of OCSP
1035 responses. OCSP responses are prepared in real time and may only
1036 be a few seconds old. Simply checking that the current time lies
1037 between thisUpdate and nextUpdate max reject otherwise valid responses
1038 caused by either OCSP responder or client clock inaccuracy. Instead
1039 we allow thisUpdate and nextUpdate to fall within a certain period of
1040 the current time. The age of the response can also optionally be
1041 checked. Two new options -validity_period and -status_age added to
1045 +) If signature or public key algorithm is unrecognized print out its
1046 OID rather that just UNKNOWN.
1049 *) Avoid coredump with unsupported or invalid public keys by checking if
1050 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1051 PKCS7_verify() fails with non detached data.
1054 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1055 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1056 ID to be generated from the issuer certificate alone which can then be
1057 passed to OCSP_id_issuer_cmp().
1060 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1061 ASN1 modules to export functions returning ASN1_ITEM pointers
1062 instead of the ASN1_ITEM structures themselves. This adds several
1063 new macros which allow the underlying ASN1 function/structure to
1064 be accessed transparently. As a result code should not use ASN1_ITEM
1065 references directly (such as &X509_it) but instead use the relevant
1066 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1067 use of the new ASN1 code on platforms where exporting structures
1068 is problematical (for example in shared libraries) but exporting
1069 functions returning pointers to structures is not.
1072 +) Add support for overriding the generation of SSL/TLS session IDs.
1073 These callbacks can be registered either in an SSL_CTX or per SSL.
1074 The purpose of this is to allow applications to control, if they wish,
1075 the arbitrary values chosen for use as session IDs, particularly as it
1076 can be useful for session caching in multiple-server environments. A
1077 command-line switch for testing this (and any client code that wishes
1078 to use such a feature) has been added to "s_server".
1079 [Geoff Thorpe, Lutz Jaenicke]
1081 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1082 of the form '#if defined(...) || defined(...) || ...' and
1083 '#if !defined(...) && !defined(...) && ...'. This also avoids
1084 the growing number of special cases it was previously handling.
1087 *) Don't use getenv in library functions when run as setuid/setgid.
1088 New function OPENSSL_issetugid().
1091 +) Make all configuration macros available for application by making
1092 sure they are available in opensslconf.h, by giving them names starting
1093 with "OPENSSL_" to avoid conflicts with other packages and by making
1094 sure e_os2.h will cover all platform-specific cases together with
1096 Additionally, it is now possible to define configuration/platform-
1097 specific names (called "system identities"). In the C code, these
1098 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1099 macro with the name beginning with "OPENSSL_SYS_", which is determined
1100 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1104 +) New option -set_serial to 'req' and 'x509' this allows the serial
1105 number to use to be specified on the command line. Previously self
1106 signed certificates were hard coded with serial number 0 and the
1107 CA options of 'x509' had to use a serial number in a file which was
1111 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1112 due to incorrect handling of multi-threading:
1114 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1116 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1118 3. Count how many times MemCheck_off() has been called so that
1119 nested use can be treated correctly. This also avoids
1120 inband-signalling in the previous code (which relied on the
1121 assumption that thread ID 0 is impossible).
1124 +) New options to 'ca' utility to support V2 CRL entry extensions.
1125 Currently CRL reason, invalidity date and hold instruction are
1126 supported. Add new CRL extensions to V3 code and some new objects.
1129 *) Add "-rand" option also to s_client and s_server.
1132 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1133 disable standard block padding (aka PKCS#5 padding) in the EVP
1134 API, which was previously mandatory. This means that the data is
1135 not padded in any way and so the total length much be a multiple
1136 of the block size, otherwise an error occurs.
1139 +) Initial (incomplete) OCSP SSL support.
1142 *) Fix CPU detection on Irix 6.x.
1143 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1144 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1146 +) New function OCSP_parse_url(). This splits up a URL into its host,
1147 port and path components: primarily to parse OCSP URLs. New -url
1148 option to ocsp utility.
1151 +) New nonce behavior. The return value of OCSP_check_nonce() now
1152 reflects the various checks performed. Applications can decide
1153 whether to tolerate certain situations such as an absent nonce
1154 in a response when one was present in a request: the ocsp application
1155 just prints out a warning. New function OCSP_add1_basic_nonce()
1156 this is to allow responders to include a nonce in a response even if
1157 the request is nonce-less.
1160 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1163 [This change does not apply to 0.9.7.]
1165 *) Use the cached encoding of an X509_NAME structure rather than
1166 copying it. This is apparently the reason for the libsafe "errors"
1167 but the code is actually correct.
1170 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1171 skipped when using openssl x509 multiple times on a single input file,
1172 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1175 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1176 set string type: to handle setting ASN1_TIME structures. Fix ca
1177 utility to correctly initialize revocation date of CRLs.
1180 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1181 the clients preferred ciphersuites and rather use its own preferences.
1182 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1183 Internet Explorer by ensuring unchanged hash method during stepup.
1184 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1187 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1188 to aes and add a new 'exist' option to print out symbols that don't
1192 +) Additional options to ocsp utility to allow flags to be set and
1193 additional certificates supplied.
1196 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1197 OCSP client a number of certificate to only verify the response
1201 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1202 Bleichenbacher's DSA attack.
1203 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1204 to be set and top=0 forces the highest bit to be set; top=-1 is new
1205 and leaves the highest bit random.
1206 [Ulf Moeller, Bodo Moeller]
1208 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1209 handle the new API. Currently only ECB, CBC modes supported. Add new
1210 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1211 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1212 [Ben Laurie, Steve Henson]
1214 *) In the NCONF_...-based implementations for CONF_... queries
1215 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1216 a temporary CONF structure with the data component set to NULL
1217 (which gives segmentation faults in lh_retrieve).
1218 Instead, use NULL for the CONF pointer in CONF_get_string and
1219 CONF_get_number (which may use environment variables) and directly
1220 return NULL from CONF_get_section.
1223 *) Fix potential buffer overrun for EBCDIC.
1226 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1227 request to response.
1230 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1231 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1232 extract information from a certificate request. OCSP_response_create()
1233 creates a response and optionally adds a basic response structure.
1234 OCSP_basic_add1_status() adds a complete single response to a basic
1235 response and returns the OCSP_SINGLERESP structure just added (to allow
1236 extensions to be included for example). OCSP_basic_add1_cert() adds a
1237 certificate to a basic response and OCSP_basic_sign() signs a basic
1238 response with various flags. New helper functions ASN1_TIME_check()
1239 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1240 (converts ASN1_TIME to GeneralizedTime).
1243 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1244 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1245 structure from a certificate. X509_pubkey_digest() digests the public_key
1246 contents: this is used in various key identifiers.
1249 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1250 keyUsage if basicConstraints absent for a CA.
1253 *) Make SMIME_write_PKCS7() write mail header values with a format that
1254 is more generally accepted (no spaces before the semicolon), since
1255 some programs can't parse those values properly otherwise. Also make
1256 sure BIO's that break lines after each write do not create invalid
1260 +) Make sk_sort() tolerate a NULL argument.
1261 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1263 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1264 passed by the function are trusted implicitly. If any of them signed the
1265 response then it is assumed to be valid and is not verified.
1268 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1269 macros previously used would not encode an empty SEQUENCE OF
1270 and break the signature.
1272 [This change does not apply to 0.9.7.]
1274 *) Zero the premaster secret after deriving the master secret in
1278 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1279 to data. This was previously part of the PKCS7 ASN1 code. This
1280 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1281 [Steve Henson, reported by Kenneth R. Robinette
1282 <support@securenetterm.com>]
1284 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1285 routines: without these tracing memory leaks is very painful.
1286 Fix leaks in PKCS12 and PKCS7 routines.
1289 *) Add some EVP_add_digest_alias registrations (as found in
1290 OpenSSL_add_all_digests()) to SSL_library_init()
1291 aka OpenSSL_add_ssl_algorithms(). This provides improved
1292 compatibility with peers using X.509 certificates
1293 with unconventional AlgorithmIdentifier OIDs.
1296 *) Fix for Irix with NO_ASM.
1297 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1299 *) ./config script fixes.
1300 [Ulf Moeller, Richard Levitte]
1302 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1303 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1304 effectively meant GeneralizedTime would never be used. Now it
1305 is initialised to -1 but X509_time_adj() now has to check the value
1306 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1307 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1308 [Steve Henson, reported by Kenneth R. Robinette
1309 <support@securenetterm.com>]
1311 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1312 result in a zero length in the ASN1_INTEGER structure which was
1313 not consistent with the structure when d2i_ASN1_INTEGER() was used
1314 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1315 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1316 where it did not print out a minus for negative ASN1_INTEGER.
1319 *) Fix 'openssl passwd -1'.
1322 +) Add summary printout to ocsp utility. The various functions which
1323 convert status values to strings have been renamed to:
1324 OCSP_response_status_str(), OCSP_cert_status_str() and
1325 OCSP_crl_reason_str() and are no longer static. New options
1326 to verify nonce values and to disable verification. OCSP response
1327 printout format cleaned up.
1330 +) Add additional OCSP certificate checks. These are those specified
1331 in RFC2560. This consists of two separate checks: the CA of the
1332 certificate being checked must either be the OCSP signer certificate
1333 or the issuer of the OCSP signer certificate. In the latter case the
1334 OCSP signer certificate must contain the OCSP signing extended key
1335 usage. This check is performed by attempting to match the OCSP
1336 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1337 in the OCSP_CERTID structures of the response.
1340 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1341 and related routines. This uses the standard OpenSSL certificate
1342 verify routines to perform initial checks (just CA validity) and
1343 to obtain the certificate chain. Then additional checks will be
1344 performed on the chain. Currently the root CA is checked to see
1345 if it is explicitly trusted for OCSP signing. This is used to set
1346 a root CA as a global signing root: that is any certificate that
1347 chains to that CA is an acceptable OCSP signing certificate.
1350 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1351 extensions from a separate configuration file.
1352 As when reading extensions from the main configuration file,
1353 the '-extensions ...' option may be used for specifying the
1355 [Massimiliano Pala <madwolf@comune.modena.it>]
1357 *) Change PKCS12_key_gen_asc() so it can cope with non null
1358 terminated strings whose length is passed in the passlen
1359 parameter, for example from PEM callbacks. This was done
1360 by adding an extra length parameter to asc2uni().
1361 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1363 +) New OCSP utility. Allows OCSP requests to be generated or
1364 read. The request can be sent to a responder and the output
1365 parsed, outputed or printed in text form. Not complete yet:
1366 still needs to check the OCSP response validity.
1369 +) New subcommands for 'openssl ca':
1370 'openssl ca -status <serial>' prints the status of the cert with
1371 the given serial number (according to the index file).
1372 'openssl ca -updatedb' updates the expiry status of certificates
1374 [Massimiliano Pala <madwolf@comune.modena.it>]
1376 +) New '-newreq-nodes' command option to CA.pl. This is like
1377 '-newreq', but calls 'openssl req' with the '-nodes' option
1378 so that the resulting key is not encrypted.
1379 [Damien Miller <djm@mindrot.org>]
1381 +) New configuration for the GNU Hurd.
1382 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1384 +) Initial code to implement OCSP basic response verify. This
1385 is currently incomplete. Currently just finds the signer's
1386 certificate and verifies the signature on the response.
1389 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1390 value of OPENSSLDIR. This is available via the new '-d' option
1391 to 'openssl version', and is also included in 'openssl version -a'.
1394 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1395 call failed, free the DSA structure.
1398 +) Allowing defining memory allocation callbacks that will be given
1399 file name and line number information in additional arguments
1400 (a const char* and an int). The basic functionality remains, as
1401 well as the original possibility to just replace malloc(),
1402 realloc() and free() by functions that do not know about these
1403 additional arguments. To register and find out the current
1404 settings for extended allocation functions, the following
1405 functions are provided:
1407 CRYPTO_set_mem_ex_functions
1408 CRYPTO_set_locked_mem_ex_functions
1409 CRYPTO_get_mem_ex_functions
1410 CRYPTO_get_locked_mem_ex_functions
1412 These work the same way as CRYPTO_set_mem_functions and friends.
1413 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1414 extended allocation function is enabled.
1415 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1416 a conventional allocation function is enabled.
1417 [Richard Levitte, Bodo Moeller]
1419 *) Fix to uni2asc() to cope with zero length Unicode strings.
1420 These are present in some PKCS#12 files.
1423 +) Finish off removing the remaining LHASH function pointer casts.
1424 There should no longer be any prototype-casting required when using
1425 the LHASH abstraction, and any casts that remain are "bugs". See
1426 the callback types and macros at the head of lhash.h for details
1427 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1430 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1431 If /dev/[u]random devices are not available or do not return enough
1432 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1434 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1435 /etc/entropy will be queried once each in this sequence, quering stops
1436 when enough entropy was collected without querying more sockets.
1439 +) Change the Unix RAND_poll() variant to be able to poll several
1440 random devices, as specified by DEVRANDOM, until a sufficient amount
1441 of data has been collected. We spend at most 10 ms on each file
1442 (select timeout) and read in non-blocking mode. DEVRANDOM now
1443 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1444 (previously it was just the string "/dev/urandom"), so on typical
1445 platforms the 10 ms delay will never occur.
1446 Also separate out the Unix variant to its own file, rand_unix.c.
1447 For VMS, there's a currently-empty rand_vms.c.
1450 +) Move OCSP client related routines to ocsp_cl.c. These
1451 provide utility functions which an application needing
1452 to issue a request to an OCSP responder and analyse the
1453 response will typically need: as opposed to those which an
1454 OCSP responder itself would need which will be added later.
1456 OCSP_request_sign() signs an OCSP request with an API similar
1457 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1458 response. OCSP_response_get1_basic() extracts basic response
1459 from response. OCSP_resp_find_status(): finds and extracts status
1460 information from an OCSP_CERTID structure (which will be created
1461 when the request structure is built). These are built from lower
1462 level functions which work on OCSP_SINGLERESP structures but
1463 wont normally be used unless the application wishes to examine
1464 extensions in the OCSP response for example.
1466 Replace nonce routines with a pair of functions.
1467 OCSP_request_add1_nonce() adds a nonce value and optionally
1468 generates a random value. OCSP_check_nonce() checks the
1469 validity of the nonce in an OCSP response.
1472 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1473 This doesn't copy the supplied OCSP_CERTID and avoids the
1474 need to free up the newly created id. Change return type
1475 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1476 This can then be used to add extensions to the request.
1477 Deleted OCSP_request_new(), since most of its functionality
1478 is now in OCSP_REQUEST_new() (and the case insensitive name
1479 clash) apart from the ability to set the request name which
1480 will be added elsewhere.
1483 +) Update OCSP API. Remove obsolete extensions argument from
1484 various functions. Extensions are now handled using the new
1485 OCSP extension code. New simple OCSP HTTP function which
1486 can be used to send requests and parse the response.
1489 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1490 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1491 uses the special reorder version of SET OF to sort the attributes
1492 and reorder them to match the encoded order. This resolves a long
1493 standing problem: a verify on a PKCS7 structure just after signing
1494 it used to fail because the attribute order did not match the
1495 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1496 it uses the received order. This is necessary to tolerate some broken
1497 software that does not order SET OF. This is handled by encoding
1498 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1499 to produce the required SET OF.
1502 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1503 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1504 files to get correct declarations of the ASN.1 item variables.
1507 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1508 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1509 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1510 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1511 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1512 ASN1_ITEM and no wrapper functions.
1515 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1516 replace the old function pointer based I/O routines. Change most of
1517 the *_d2i_bio() and *_d2i_fp() functions to use these.
1520 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1521 lines, recognice more "algorithms" that can be deselected, and make
1522 it complain about algorithm deselection that isn't recognised.
1525 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1526 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1527 to use new functions. Add NO_ASN1_OLD which can be set to remove
1528 some old style ASN1 functions: this can be used to determine if old
1529 code will still work when these eventually go away.
1532 +) New extension functions for OCSP structures, these follow the
1533 same conventions as certificates and CRLs.
1536 +) New function X509V3_add1_i2d(). This automatically encodes and
1537 adds an extension. Its behaviour can be customised with various
1538 flags to append, replace or delete. Various wrappers added for
1539 certifcates and CRLs.
1542 +) Fix to avoid calling the underlying ASN1 print routine when
1543 an extension cannot be parsed. Correct a typo in the
1544 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1547 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1548 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1549 when writing a 32767 byte record.
1550 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1552 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1553 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1555 (RSA objects have a reference count access to which is protected
1556 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1557 so they are meant to be shared between threads.)
1558 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1559 "Reddie, Steven" <Steven.Reddie@ca.com>]
1561 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1562 entries for variables.
1565 *) Fix a deadlock in CRYPTO_mem_leaks().
1568 +) Add functionality to apps/openssl.c for detecting locking
1569 problems: As the program is single-threaded, all we have
1570 to do is register a locking callback using an array for
1571 storing which locks are currently held by the program.
1574 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1575 SSL_get_ex_data_X509_STORE_idx(), which is used in
1576 ssl_verify_cert_chain() and thus can be called at any time
1577 during TLS/SSL handshakes so that thread-safety is essential.
1578 Unfortunately, the ex_data design is not at all suited
1579 for multi-threaded use, so it probably should be abolished.
1582 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1583 [Broadcom, tweaked and integrated by Geoff Thorpe]
1585 +) Move common extension printing code to new function
1586 X509V3_print_extensions(). Reorganise OCSP print routines and
1587 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1590 +) New function X509_signature_print() to remove duplication in some
1594 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1595 set (this was treated exactly the same as SET OF previously). This
1596 is used to reorder the STACK representing the structure to match the
1597 encoding. This will be used to get round a problem where a PKCS7
1598 structure which was signed could not be verified because the STACK
1599 order did not reflect the encoded order.
1602 +) Reimplement the OCSP ASN1 module using the new code.
1605 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1606 for its ASN1 operations. The old style function pointers still exist
1607 for now but they will eventually go away.
1610 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1611 completely replaces the old ASN1 functionality with a table driven
1612 encoder and decoder which interprets an ASN1_ITEM structure describing
1613 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1614 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1615 has also been converted to the new form.
1618 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1619 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1620 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1621 for negative moduli.
1624 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1625 of not touching the result's sign bit.
1628 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1632 +) Changed the LHASH code to use prototypes for callbacks, and created
1633 macros to declare and implement thin (optionally static) functions
1634 that provide type-safety and avoid function pointer casting for the
1635 type-specific callbacks.
1638 *) Use better test patterns in bntest.
1641 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1643 [Veers Staats <staatsvr@asc.hpc.mil>,
1644 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1646 *) rand_win.c fix for Borland C.
1649 *) BN_rshift bugfix for n == 0.
1652 +) Reformat the FAQ so the different questions and answers can be divided
1653 in sections depending on the subject.
1656 +) Have the zlib compression code load ZLIB.DLL dynamically under
1660 +) New function BN_mod_sqrt for computing square roots modulo a prime
1661 (using the probabilistic Tonelli-Shanks algorithm unless
1662 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1663 be handled deterministically).
1664 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1666 *) Add a 'bctest' script that checks for some known 'bc' bugs
1667 so that 'make test' does not abort just because 'bc' is broken.
1670 *) Store verify_result within SSL_SESSION also for client side to
1671 avoid potential security hole. (Re-used sessions on the client side
1672 always resulted in verify_result==X509_V_OK, not using the original
1673 result of the server certificate verification.)
1676 +) Make BN_mod_inverse faster by explicitly handling small quotients
1677 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1678 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1681 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1682 SSL3_RT_APPLICATION_DATA, return 0.
1683 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1687 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1688 releases, have been re-implemented by renaming the previous
1689 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1690 and ssl3_read_internal, respectively, and adding 'peek' parameters
1691 to them. The new ssl[23]_{read,peek} functions are calls to
1692 ssl[23]_read_internal with the 'peek' flag set appropriately.
1693 A 'peek' parameter has also been added to ssl3_read_bytes, which
1694 does the actual work for ssl3_read_internal.
1697 +) New function BN_kronecker.
1700 +) Fix BN_gcd so that it works on negative inputs; the result is
1701 positive unless both parameters are zero.
1702 Previously something reasonably close to an infinite loop was
1703 possible because numbers could be growing instead of shrinking
1704 in the implementation of Euclid's algorithm.
1707 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1708 sign of the number in question.
1710 Fix BN_is_word(a,w) to work correctly for w == 0.
1712 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1713 because its test if the absolute value of 'a' equals 'w'.
1714 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1715 it exists mostly for use in the implementations of BN_is_zero(),
1716 BN_is_one(), and BN_is_word().
1719 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1720 the method-specific "init()" handler. Also clean up ex_data after
1721 calling the method-specific "finish()" handler. Previously, this was
1722 happening the other way round.
1725 +) New function BN_swap.
1728 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1729 the exponentiation functions are more likely to produce reasonable
1730 results on negative inputs.
1733 +) Change BN_mod_mul so that the result is always non-negative.
1734 Previously, it could be negative if one of the factors was negative;
1735 I don't think anyone really wanted that behaviour.
1738 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1739 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1740 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1741 and add new functions:
1750 BN_mod_lshift1_quick
1754 These functions always generate non-negative results.
1756 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1757 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1759 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1760 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1761 be reduced modulo m.
1762 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1764 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1765 was actually never needed) and in BN_mul(). The removal in BN_mul()
1766 required a small change in bn_mul_part_recursive() and the addition
1767 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1768 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1769 bn_sub_words() and bn_add_words() except they take arrays with
1773 +) In 'openssl passwd', verify passwords read from the terminal
1774 unless the '-salt' option is used (which usually means that
1775 verification would just waste user's time since the resulting
1776 hash is going to be compared with some given password hash)
1777 or the new '-noverify' option is used.
1779 This is an incompatible change, but it does not affect
1780 non-interactive use of 'openssl passwd' (passwords on the command
1781 line, '-stdin' option, '-in ...' option) and thus should not
1785 +) Remove all references to RSAref, since there's no more need for it.
1788 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1789 The previous value, 12, was not always sufficient for BN_mod_exp().
1792 +) Make DSO load along a path given through an environment variable
1793 (SHLIB_PATH) with shl_load().
1796 +) Constify the ENGINE code as a result of BIGNUM constification.
1797 Also constify the RSA code and most things related to it. In a
1798 few places, most notable in the depth of the ASN.1 code, ugly
1799 casts back to non-const were required (to be solved at a later
1803 +) Make it so the openssl application has all engines loaded by default.
1806 +) Constify the BIGNUM routines a little more.
1809 *) Make sure that shared libraries get the internal name engine with
1810 the full version number and not just 0. This should mark the
1811 shared libraries as not backward compatible. Of course, this should
1812 be changed again when we can guarantee backward binary compatibility.
1815 +) Add the following functions:
1817 ENGINE_load_cswift()
1819 ENGINE_load_atalla()
1821 ENGINE_load_builtin_engines()
1823 That way, an application can itself choose if external engines that
1824 are built-in in OpenSSL shall ever be used or not. The benefit is
1825 that applications won't have to be linked with libdl or other dso
1826 libraries unless it's really needed.
1828 Changed 'openssl engine' to load all engines on demand.
1829 Changed the engine header files to avoid the duplication of some
1830 declarations (they differed!).
1833 +) 'openssl engine' can now list capabilities.
1836 +) Better error reporting in 'openssl engine'.
1839 +) Never call load_dh_param(NULL) in s_server.
1842 +) Add engine application. It can currently list engines by name and
1843 identity, and test if they are actually available.
1846 +) Add support for shared libraries under Irix.
1847 [Albert Chin-A-Young <china@thewrittenword.com>]
1849 +) Improve RPM specification file by forcing symbolic linking and making
1850 sure the installed documentation is also owned by root.root.
1851 [Damien Miller <djm@mindrot.org>]
1853 +) Add configuration option to build on Linux on both big-endian and
1855 [Ralf Baechle <ralf@uni-koblenz.de>]
1857 +) Give the OpenSSL applications more possibilities to make use of
1858 keys (public as well as private) handled by engines.
1861 +) Add OCSP code that comes from CertCo.
1864 +) Add VMS support for the Rijndael code.
1867 +) Added untested support for Nuron crypto accelerator.
1870 +) Add support for external cryptographic devices. This code was
1871 previously distributed separately as the "engine" branch.
1872 [Geoff Thorpe, Richard Levitte]
1874 +) Rework the filename-translation in the DSO code. It is now possible to
1875 have far greater control over how a "name" is turned into a filename
1876 depending on the operating environment and any oddities about the
1877 different shared library filenames on each system.
1880 +) Support threads on FreeBSD-elf in Configure.
1883 +) Add the possibility to create shared libraries on HP-UX
1886 +) Fix for SHA1 assembly problem with MASM: it produces
1887 warnings about corrupt line number information when assembling
1888 with debugging information. This is caused by the overlapping
1890 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1893 NCONF_get_number() has no error checking at all. As a replacement,
1894 NCONF_get_number_e() is defined (_e for "error checking") and is
1895 promoted strongly. The old NCONF_get_number is kept around for
1896 binary backward compatibility.
1897 Make it possible for methods to load from something other than a BIO,
1898 by providing a function pointer that is given a name instead of a BIO.
1899 For example, this could be used to load configuration data from an
1903 *) Fix typo in get_cert_by_subject() in by_dir.c
1904 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1906 *) Rework the system to generate shared libraries:
1908 - Make note of the expected extension for the shared libraries and
1909 if there is a need for symbolic links from for example libcrypto.so.0
1910 to libcrypto.so.0.9.7. There is extended info in Configure for
1913 - Make as few rebuilds of the shared libraries as possible.
1915 - Still avoid linking the OpenSSL programs with the shared libraries.
1917 - When installing, install the shared libraries separately from the
1921 +) Fix for non blocking accept BIOs. Added new I/O special reason
1922 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1923 with non blocking I/O was not possible because no retry code was
1924 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1928 +) Added the beginnings of Rijndael support.
1931 +) Fix for bug in DirectoryString mask setting. Add support for
1932 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1933 to allow certificate printing to more controllable, additional
1934 'certopt' option to 'x509' to allow new printing options to be
1938 +) Clean old EAY MD5 hack from e_os.h.
1941 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1943 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1944 and not in SSL_clear because the latter is also used by the
1945 accept/connect functions; previously, the settings made by
1946 SSL_set_read_ahead would be lost during the handshake.
1947 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1949 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1950 Previously, it would create entries for disableed algorithms no
1954 *) Added several new manual pages for SSL_* function.
1957 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
1959 *) In ssl23_get_client_hello, generate an error message when faced
1960 with an initial SSL 3.0/TLS record that is too small to contain the
1961 first two bytes of the ClientHello message, i.e. client_version.
1962 (Note that this is a pathologic case that probably has never happened
1963 in real life.) The previous approach was to use the version number
1964 from the record header as a substitute; but our protocol choice
1965 should not depend on that one because it is not authenticated
1966 by the Finished messages.
1969 *) More robust randomness gathering functions for Windows.
1970 [Jeffrey Altman <jaltman@columbia.edu>]
1972 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
1973 not set then we don't setup the error code for issuer check errors
1974 to avoid possibly overwriting other errors which the callback does
1975 handle. If an application does set the flag then we assume it knows
1976 what it is doing and can handle the new informational codes
1980 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
1981 a general "ANY" type, as such it should be able to decode anything
1982 including tagged types. However it didn't check the class so it would
1983 wrongly interpret tagged types in the same way as their universal
1984 counterpart and unknown types were just rejected. Changed so that the
1985 tagged and unknown types are handled in the same way as a SEQUENCE:
1986 that is the encoding is stored intact. There is also a new type
1987 "V_ASN1_OTHER" which is used when the class is not universal, in this
1988 case we have no idea what the actual type is so we just lump them all
1992 *) On VMS, stdout may very well lead to a file that is written to
1993 in a record-oriented fashion. That means that every write() will
1994 write a separate record, which will be read separately by the
1995 programs trying to read from it. This can be very confusing.
1997 The solution is to put a BIO filter in the way that will buffer
1998 text until a linefeed is reached, and then write everything a
1999 line at a time, so every record written will be an actual line,
2000 not chunks of lines and not (usually doesn't happen, but I've
2001 seen it once) several lines in one record. BIO_f_linebuffer() is
2004 Currently, it's a VMS-only method, because that's where it has
2005 been tested well enough.
2008 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2009 it can return incorrect results.
2010 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2011 but it was in 0.9.6-beta[12].)
2014 *) Disable the check for content being present when verifying detached
2015 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2016 include zero length content when signing messages.
2019 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2020 BIO_ctrl (for BIO pairs).
2023 *) Add DSO method for VMS.
2026 *) Bug fix: Montgomery multiplication could produce results with the
2030 *) Add RPM specification openssl.spec and modify it to build three
2031 packages. The default package contains applications, application
2032 documentation and run-time libraries. The devel package contains
2033 include files, static libraries and function documentation. The
2034 doc package contains the contents of the doc directory. The original
2035 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2038 *) Add a large number of documentation files for many SSL routines.
2039 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2041 *) Add a configuration entry for Sony News 4.
2042 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2044 *) Don't set the two most significant bits to one when generating a
2045 random number < q in the DSA library.
2048 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2049 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2050 the underlying transport is blocking) if a handshake took place.
2051 (The default behaviour is needed by applications such as s_client
2052 and s_server that use select() to determine when to use SSL_read;
2053 but for applications that know in advance when to expect data, it
2054 just makes things more complicated.)
2057 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2061 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2062 work better on such systems.
2063 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2065 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2066 Update PKCS12_parse() so it copies the friendlyName and the
2067 keyid to the certificates aux info.
2070 *) Fix bug in PKCS7_verify() which caused an infinite loop
2071 if there was more than one signature.
2072 [Sven Uszpelkat <su@celocom.de>]
2074 *) Major change in util/mkdef.pl to include extra information
2075 about each symbol, as well as presentig variables as well
2076 as functions. This change means that there's n more need
2077 to rebuild the .num files when some algorithms are excluded.
2080 *) Allow the verify time to be set by an application,
2081 rather than always using the current time.
2084 *) Phase 2 verify code reorganisation. The certificate
2085 verify code now looks up an issuer certificate by a
2086 number of criteria: subject name, authority key id
2087 and key usage. It also verifies self signed certificates
2088 by the same criteria. The main comparison function is
2089 X509_check_issued() which performs these checks.
2091 Lot of changes were necessary in order to support this
2092 without completely rewriting the lookup code.
2094 Authority and subject key identifier are now cached.
2096 The LHASH 'certs' is X509_STORE has now been replaced
2097 by a STACK_OF(X509_OBJECT). This is mainly because an
2098 LHASH can't store or retrieve multiple objects with
2099 the same hash value.
2101 As a result various functions (which were all internal
2102 use only) have changed to handle the new X509_STORE
2103 structure. This will break anything that messed round
2104 with X509_STORE internally.
2106 The functions X509_STORE_add_cert() now checks for an
2107 exact match, rather than just subject name.
2109 The X509_STORE API doesn't directly support the retrieval
2110 of multiple certificates matching a given criteria, however
2111 this can be worked round by performing a lookup first
2112 (which will fill the cache with candidate certificates)
2113 and then examining the cache for matches. This is probably
2114 the best we can do without throwing out X509_LOOKUP
2115 entirely (maybe later...).
2117 The X509_VERIFY_CTX structure has been enhanced considerably.
2119 All certificate lookup operations now go via a get_issuer()
2120 callback. Although this currently uses an X509_STORE it
2121 can be replaced by custom lookups. This is a simple way
2122 to bypass the X509_STORE hackery necessary to make this
2123 work and makes it possible to use more efficient techniques
2124 in future. A very simple version which uses a simple
2125 STACK for its trusted certificate store is also provided
2126 using X509_STORE_CTX_trusted_stack().
2128 The verify_cb() and verify() callbacks now have equivalents
2129 in the X509_STORE_CTX structure.
2131 X509_STORE_CTX also has a 'flags' field which can be used
2132 to customise the verify behaviour.
2135 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2136 excludes S/MIME capabilities.
2139 *) When a certificate request is read in keep a copy of the
2140 original encoding of the signed data and use it when outputing
2141 again. Signatures then use the original encoding rather than
2142 a decoded, encoded version which may cause problems if the
2143 request is improperly encoded.
2146 *) For consistency with other BIO_puts implementations, call
2147 buffer_write(b, ...) directly in buffer_puts instead of calling
2150 In BIO_puts, increment b->num_write as in BIO_write.
2151 [Peter.Sylvester@EdelWeb.fr]
2153 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2154 BN_zero, we may not return a BIGNUM with an array consisting of
2158 *) Avoid calling abort() from within the library when problems are
2159 detected, except if preprocessor symbols have been defined
2160 (such as REF_CHECK, BN_DEBUG etc.).
2163 *) New openssl application 'rsautl'. This utility can be
2164 used for low level RSA operations. DER public key
2165 BIO/fp routines also added.
2168 *) New Configure entry and patches for compiling on QNX 4.
2169 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2171 *) A demo state-machine implementation was sponsored by
2172 Nuron (http://www.nuron.com/) and is now available in
2173 demos/state_machine.
2176 *) New options added to the 'dgst' utility for signature
2177 generation and verification.
2180 *) Unrecognized PKCS#7 content types are now handled via a
2181 catch all ASN1_TYPE structure. This allows unsupported
2182 types to be stored as a "blob" and an application can
2183 encode and decode it manually.
2186 *) Fix various signed/unsigned issues to make a_strex.c
2188 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2190 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2191 length if passed a buffer. ASN1_INTEGER_to_BN failed
2192 if passed a NULL BN and its argument was negative.
2193 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2195 *) Modification to PKCS#7 encoding routines to output definite
2196 length encoding. Since currently the whole structures are in
2197 memory there's not real point in using indefinite length
2198 constructed encoding. However if OpenSSL is compiled with
2199 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2202 *) Added BIO_vprintf() and BIO_vsnprintf().
2205 *) Added more prefixes to parse for in the the strings written
2206 through a logging bio, to cover all the levels that are available
2207 through syslog. The prefixes are now:
2209 PANIC, EMERG, EMR => LOG_EMERG
2210 ALERT, ALR => LOG_ALERT
2211 CRIT, CRI => LOG_CRIT
2212 ERROR, ERR => LOG_ERR
2213 WARNING, WARN, WAR => LOG_WARNING
2214 NOTICE, NOTE, NOT => LOG_NOTICE
2215 INFO, INF => LOG_INFO
2216 DEBUG, DBG => LOG_DEBUG
2218 and as before, if none of those prefixes are present at the
2219 beginning of the string, LOG_ERR is chosen.
2221 On Win32, the LOG_* levels are mapped according to this:
2223 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2224 LOG_WARNING => EVENTLOG_WARNING_TYPE
2225 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2229 *) Made it possible to reconfigure with just the configuration
2230 argument "reconf" or "reconfigure". The command line arguments
2231 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2232 and are retrieved from there when reconfiguring.
2236 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2238 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2241 *) The obj_dat.pl script was messing up the sorting of object
2242 names. The reason was that it compared the quoted version
2243 of strings as a result "OCSP" > "OCSP Signing" because
2244 " > SPACE. Changed script to store unquoted versions of
2245 names and add quotes on output. It was also omitting some
2246 names from the lookup table if they were given a default
2247 value (that is if SN is missing it is given the same
2248 value as LN and vice versa), these are now added on the
2249 grounds that if an object has a name we should be able to
2250 look it up. Finally added warning output when duplicate
2251 short or long names are found.
2254 *) Changes needed for Tandem NSK.
2255 [Scott Uroff <scott@xypro.com>]
2257 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2258 RSA_padding_check_SSLv23(), special padding was never detected
2259 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2260 version rollback attacks was not effective.
2262 In s23_clnt.c, don't use special rollback-attack detection padding
2263 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2264 client; similarly, in s23_srvr.c, don't do the rollback check if
2265 SSL 2.0 is the only protocol enabled in the server.
2268 *) Make it possible to get hexdumps of unprintable data with 'openssl
2269 asn1parse'. By implication, the functions ASN1_parse_dump() and
2270 BIO_dump_indent() are added.
2273 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2274 these print out strings and name structures based on various
2275 flags including RFC2253 support and proper handling of
2276 multibyte characters. Added options to the 'x509' utility
2277 to allow the various flags to be set.
2280 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2281 Also change the functions X509_cmp_current_time() and
2282 X509_gmtime_adj() work with an ASN1_TIME structure,
2283 this will enable certificates using GeneralizedTime in validity
2284 dates to be checked.
2287 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2288 negative public key encodings) on by default,
2289 NO_NEG_PUBKEY_BUG can be set to disable it.
2292 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2293 content octets. An i2c_ASN1_OBJECT is unnecessary because
2294 the encoding can be trivially obtained from the structure.
2297 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2298 not read locks (CRYPTO_r_[un]lock).
2301 *) A first attempt at creating official support for shared
2302 libraries through configuration. I've kept it so the
2303 default is static libraries only, and the OpenSSL programs
2304 are always statically linked for now, but there are
2305 preparations for dynamic linking in place.
2306 This has been tested on Linux and Tru64.
2309 *) Randomness polling function for Win9x, as described in:
2310 Peter Gutmann, Software Generation of Practically Strong
2314 *) Fix so PRNG is seeded in req if using an already existing
2318 *) New options to smime application. -inform and -outform
2319 allow alternative formats for the S/MIME message including
2320 PEM and DER. The -content option allows the content to be
2321 specified separately. This should allow things like Netscape
2322 form signing output easier to verify.
2325 *) Fix the ASN1 encoding of tags using the 'long form'.
2328 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2329 STRING types. These convert content octets to and from the
2330 underlying type. The actual tag and length octets are
2331 already assumed to have been read in and checked. These
2332 are needed because all other string types have virtually
2333 identical handling apart from the tag. By having versions
2334 of the ASN1 functions that just operate on content octets
2335 IMPLICIT tagging can be handled properly. It also allows
2336 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2337 and ASN1_INTEGER are identical apart from the tag.
2340 *) Change the handling of OID objects as follows:
2342 - New object identifiers are inserted in objects.txt, following
2343 the syntax given in objects.README.
2344 - objects.pl is used to process obj_mac.num and create a new
2346 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2349 This is currently kind of a hack, and the perl code in objects.pl
2350 isn't very elegant, but it works as I intended. The simplest way
2351 to check that it worked correctly is to look in obj_dat.h and
2352 check the array nid_objs and make sure the objects haven't moved
2353 around (this is important!). Additions are OK, as well as
2354 consistent name changes.
2357 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2360 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2361 The given file adds to whatever has already been seeded into the
2362 random pool through the RANDFILE configuration file option or
2363 environment variable, or the default random state file.
2366 *) mkstack.pl now sorts each macro group into lexical order.
2367 Previously the output order depended on the order the files
2368 appeared in the directory, resulting in needless rewriting
2372 *) Patches to make OpenSSL compile under Win32 again. Mostly
2373 work arounds for the VC++ problem that it treats func() as
2374 func(void). Also stripped out the parts of mkdef.pl that
2375 added extra typesafe functions: these no longer exist.
2378 *) Reorganisation of the stack code. The macros are now all
2379 collected in safestack.h . Each macro is defined in terms of
2380 a "stack macro" of the form SKM_<name>(type, a, b). The
2381 DEBUG_SAFESTACK is now handled in terms of function casts,
2382 this has the advantage of retaining type safety without the
2383 use of additional functions. If DEBUG_SAFESTACK is not defined
2384 then the non typesafe macros are used instead. Also modified the
2385 mkstack.pl script to handle the new form. Needs testing to see
2386 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2387 the default if no major problems. Similar behaviour for ASN1_SET_OF
2388 and PKCS12_STACK_OF.
2391 *) When some versions of IIS use the 'NET' form of private key the
2392 key derivation algorithm is different. Normally MD5(password) is
2393 used as a 128 bit RC4 key. In the modified case
2394 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2395 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2396 as the old Netscape_RSA functions except they have an additional
2397 'sgckey' parameter which uses the modified algorithm. Also added
2398 an -sgckey command line option to the rsa utility. Thanks to
2399 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2400 algorithm to openssl-dev.
2403 *) The evp_local.h macros were using 'c.##kname' which resulted in
2404 invalid expansion on some systems (SCO 5.0.5 for example).
2405 Corrected to 'c.kname'.
2406 [Phillip Porch <root@theporch.com>]
2408 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2409 a STACK of email addresses from a certificate or request, these look
2410 in the subject name and the subject alternative name extensions and
2411 omit any duplicate addresses.
2414 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2415 This makes DSA verification about 2 % faster.
2418 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2419 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2420 plus overhead for 1024 bit moduli).
2421 This makes exponentiations about 0.5 % faster for 1024 bit
2422 exponents (as measured by "openssl speed rsa2048").
2425 *) Rename memory handling macros to avoid conflicts with other
2427 Malloc => OPENSSL_malloc
2428 Malloc_locked => OPENSSL_malloc_locked
2429 Realloc => OPENSSL_realloc
2430 Free => OPENSSL_free
2433 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2434 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2437 *) CygWin32 support.
2438 [John Jarvie <jjarvie@newsguy.com>]
2440 *) The type-safe stack code has been rejigged. It is now only compiled
2441 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2442 by default all type-specific stack functions are "#define"d back to
2443 standard stack functions. This results in more streamlined output
2444 but retains the type-safety checking possibilities of the original
2448 *) The STACK code has been cleaned up, and certain type declarations
2449 that didn't make a lot of sense have been brought in line. This has
2450 also involved a cleanup of sorts in safestack.h to more correctly
2451 map type-safe stack functions onto their plain stack counterparts.
2452 This work has also resulted in a variety of "const"ifications of
2453 lots of the code, especially "_cmp" operations which should normally
2454 be prototyped with "const" parameters anyway.
2457 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2458 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2459 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2460 where all of 'md' is used each time the PRNG is used, but 'state'
2461 is used only indexed by a cyclic counter. As entropy may not be
2462 well distributed from the beginning, 'md' is important as a
2463 chaining variable. However, the output function chains only half
2464 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2465 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2466 in all of 'state' being rewritten, with the new values depending
2467 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2470 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2471 the handshake is continued after ssl_verify_cert_chain();
2472 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2473 can lead to 'unexplainable' connection aborts later.
2474 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2476 *) Major EVP API cipher revision.
2477 Add hooks for extra EVP features. This allows various cipher
2478 parameters to be set in the EVP interface. Support added for variable
2479 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2480 setting of RC2 and RC5 parameters.
2482 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2485 Remove lots of duplicated code from the EVP library. For example *every*
2486 cipher init() function handles the 'iv' in the same way according to the
2487 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2488 for CFB and OFB modes they zero ctx->num.
2490 New functionality allows removal of S/MIME code RC2 hack.
2492 Most of the routines have the same form and so can be declared in terms
2495 By shifting this to the top level EVP_CipherInit() it can be removed from
2496 all individual ciphers. If the cipher wants to handle IVs or keys
2497 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2500 Change lots of functions like EVP_EncryptUpdate() to now return a
2501 value: although software versions of the algorithms cannot fail
2502 any installed hardware versions can.
2505 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2506 this option is set, tolerate broken clients that send the negotiated
2507 protocol version number instead of the requested protocol version
2511 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2512 i.e. non-zero for export ciphersuites, zero otherwise.
2513 Previous versions had this flag inverted, inconsistent with
2514 rsa_tmp_cb (..._TMP_RSA_CB).
2515 [Bodo Moeller; problem reported by Amit Chopra]
2517 *) Add missing DSA library text string. Work around for some IIS
2518 key files with invalid SEQUENCE encoding.
2521 *) Add a document (doc/standards.txt) that list all kinds of standards
2522 and so on that are implemented in OpenSSL.
2525 *) Enhance c_rehash script. Old version would mishandle certificates
2526 with the same subject name hash and wouldn't handle CRLs at all.
2527 Added -fingerprint option to crl utility, to support new c_rehash
2531 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2534 *) Fix for SSL server purpose checking. Server checking was
2535 rejecting certificates which had extended key usage present
2536 but no ssl client purpose.
2537 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2539 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2540 is a little unclear about how a blank password is handled.
2541 Since the password in encoded as a BMPString with terminating
2542 double NULL a zero length password would end up as just the
2543 double NULL. However no password at all is different and is
2544 handled differently in the PKCS#12 key generation code. NS
2545 treats a blank password as zero length. MSIE treats it as no
2546 password on export: but it will try both on import. We now do
2547 the same: PKCS12_parse() tries zero length and no password if
2548 the password is set to "" or NULL (NULL is now a valid password:
2549 it wasn't before) as does the pkcs12 application.
2552 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2553 perror when PEM_read_bio_X509_REQ fails, the error message must
2554 be obtained from the error queue.
2557 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2558 it in ERR_remove_state if appropriate, and change ERR_get_state
2559 accordingly to avoid race conditions (this is necessary because
2560 thread_hash is no longer constant once set).
2563 *) Bugfix for linux-elf makefile.one.
2566 *) RSA_get_default_method() will now cause a default
2567 RSA_METHOD to be chosen if one doesn't exist already.
2568 Previously this was only set during a call to RSA_new()
2569 or RSA_new_method(NULL) meaning it was possible for
2570 RSA_get_default_method() to return NULL.
2573 *) Added native name translation to the existing DSO code
2574 that will convert (if the flag to do so is set) filenames
2575 that are sufficiently small and have no path information
2576 into a canonical native form. Eg. "blah" converted to
2577 "libblah.so" or "blah.dll" etc.
2580 *) New function ERR_error_string_n(e, buf, len) which is like
2581 ERR_error_string(e, buf), but writes at most 'len' bytes
2582 including the 0 terminator. For ERR_error_string_n, 'buf'
2584 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2586 *) CONF library reworked to become more general. A new CONF
2587 configuration file reader "class" is implemented as well as a
2588 new functions (NCONF_*, for "New CONF") to handle it. The now
2589 old CONF_* functions are still there, but are reimplemented to
2590 work in terms of the new functions. Also, a set of functions
2591 to handle the internal storage of the configuration data is
2592 provided to make it easier to write new configuration file
2593 reader "classes" (I can definitely see something reading a
2594 configuration file in XML format, for example), called _CONF_*,
2595 or "the configuration storage API"...
2597 The new configuration file reading functions are:
2599 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2600 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2602 NCONF_default, NCONF_WIN32
2604 NCONF_dump_fp, NCONF_dump_bio
2606 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2607 NCONF_new creates a new CONF object. This works in the same way
2608 as other interfaces in OpenSSL, like the BIO interface.
2609 NCONF_dump_* dump the internal storage of the configuration file,
2610 which is useful for debugging. All other functions take the same
2611 arguments as the old CONF_* functions wth the exception of the
2612 first that must be a `CONF *' instead of a `LHASH *'.
2614 To make it easer to use the new classes with the old CONF_* functions,
2615 the function CONF_set_default_method is provided.
2618 *) Add '-tls1' option to 'openssl ciphers', which was already
2619 mentioned in the documentation but had not been implemented.
2620 (This option is not yet really useful because even the additional
2621 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2624 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2625 OpenSSL-based applications) load shared libraries and bind to
2626 them in a portable way.
2627 [Geoff Thorpe, with contributions from Richard Levitte]
2629 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2631 *) Make sure _lrotl and _lrotr are only used with MSVC.
2633 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2634 (the default implementation of RAND_status).
2636 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2637 to '-clrext' (= clear extensions), as intended and documented.
2638 [Bodo Moeller; inconsistency pointed out by Michael Attili
2641 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2642 was larger than the MD block size.
2643 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2645 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2646 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2647 using the passed key: if the passed key was a private key the result
2648 of X509_print(), for example, would be to print out all the private key
2652 *) des_quad_cksum() byte order bug fix.
2653 [Ulf Möller, using the problem description in krb4-0.9.7, where
2654 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2656 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2658 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2660 *) For easily testing in shell scripts whether some command
2661 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2662 returns with exit code 0 iff no command of the given name is available.
2663 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2664 the output goes to stdout and nothing is printed to stderr.
2665 Additional arguments are always ignored.
2667 Since for each cipher there is a command of the same name,
2668 the 'no-cipher' compilation switches can be tested this way.
2670 ('openssl no-XXX' is not able to detect pseudo-commands such
2671 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2674 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2677 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2678 is set; it will be thrown away anyway because each handshake creates
2680 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2681 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2682 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2683 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2686 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2687 'Q' and 'R' lose their special meanings (quit/renegotiate).
2688 This is part of what -quiet does; unlike -quiet, -ign_eof
2689 does not suppress any output.
2692 *) Add compatibility options to the purpose and trust code. The
2693 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2694 accepts a certificate or CA, this was the previous behaviour,
2695 with all the associated security issues.
2697 X509_TRUST_COMPAT is the old trust behaviour: only and
2698 automatically trust self signed roots in certificate store. A
2699 new trust setting X509_TRUST_DEFAULT is used to specify that
2700 a purpose has no associated trust setting and it should instead
2701 use the value in the default purpose.
2704 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2705 and fix a memory leak.
2708 *) In util/mkerr.pl (which implements 'make errors'), preserve
2709 reason strings from the previous version of the .c file, as
2710 the default to have only downcase letters (and digits) in
2711 automatically generated reasons codes is not always appropriate.
2714 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2715 using strerror. Previously, ERR_reason_error_string() returned
2716 library names as reason strings for SYSerr; but SYSerr is a special
2717 case where small numbers are errno values, not library numbers.
2720 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2721 converts DSA parameters into DH parameters. (When creating parameters,
2722 DSA_generate_parameters is used.)
2725 *) Include 'length' (recommended exponent length) in C code generated
2726 by 'openssl dhparam -C'.
2729 *) The second argument to set_label in perlasm was already being used
2730 so couldn't be used as a "file scope" flag. Moved to third argument
2734 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2735 instead of RAND_bytes for encryption IVs and salts.
2738 *) Include RAND_status() into RAND_METHOD instead of implementing
2739 it only for md_rand.c Otherwise replacing the PRNG by calling
2740 RAND_set_rand_method would be impossible.
2743 *) Don't let DSA_generate_key() enter an infinite loop if the random
2744 number generation fails.
2747 *) New 'rand' application for creating pseudo-random output.
2750 *) Added configuration support for Linux/IA64
2751 [Rolf Haberrecker <rolf@suse.de>]
2753 *) Assembler module support for Mingw32.
2756 *) Shared library support for HPUX (in shlib/).
2757 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2759 *) Shared library support for Solaris gcc.
2760 [Lutz Behnke <behnke@trustcenter.de>]
2762 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2764 *) PKCS7_encrypt() was adding text MIME headers twice because they
2765 were added manually and by SMIME_crlf_copy().
2768 *) In bntest.c don't call BN_rand with zero bits argument.
2769 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2771 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2772 case was implemented. This caused BN_div_recp() to fail occasionally.
2775 *) Add an optional second argument to the set_label() in the perl
2776 assembly language builder. If this argument exists and is set
2777 to 1 it signals that the assembler should use a symbol whose
2778 scope is the entire file, not just the current function. This
2779 is needed with MASM which uses the format label:: for this scope.
2780 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2782 *) Change the ASN1 types so they are typedefs by default. Before
2783 almost all types were #define'd to ASN1_STRING which was causing
2784 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2788 *) Change names of new functions to the new get1/get0 naming
2789 convention: After 'get1', the caller owns a reference count
2790 and has to call ..._free; 'get0' returns a pointer to some
2791 data structure without incrementing reference counters.
2792 (Some of the existing 'get' functions increment a reference
2793 counter, some don't.)
2794 Similarly, 'set1' and 'add1' functions increase reference
2795 counters or duplicate objects.
2798 *) Allow for the possibility of temp RSA key generation failure:
2799 the code used to assume it always worked and crashed on failure.
2802 *) Fix potential buffer overrun problem in BIO_printf().
2803 [Ulf Möller, using public domain code by Patrick Powell; problem
2804 pointed out by David Sacerdote <das33@cornell.edu>]
2806 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2807 RAND_egd() and RAND_status(). In the command line application,
2808 the EGD socket can be specified like a seed file using RANDFILE
2812 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2813 Some CAs (e.g. Verisign) distribute certificates in this form.
2816 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2817 list to exclude them. This means that no special compilation option
2818 is needed to use anonymous DH: it just needs to be included in the
2822 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2823 EVP_MD_type. The old functionality is available in a new macro called
2824 EVP_MD_md(). Change code that uses it and update docs.
2827 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2828 where the 'void *' argument is replaced by a function pointer argument.
2829 Previously 'void *' was abused to point to functions, which works on
2830 many platforms, but is not correct. As these functions are usually
2831 called by macros defined in OpenSSL header files, most source code
2832 should work without changes.
2835 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2836 sections with information on -D... compiler switches used for
2837 compiling the library so that applications can see them. To enable
2838 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2839 must be defined. E.g.,
2840 #define OPENSSL_ALGORITHM_DEFINES
2841 #include <openssl/opensslconf.h>
2842 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2843 [Richard Levitte, Ulf and Bodo Möller]
2845 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2849 *) Change the 'other' type in certificate aux info to a STACK_OF
2850 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2851 the required ASN1 format: arbitrary types determined by an OID.
2854 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2855 argument to 'req'. This is not because the function is newer or
2856 better than others it just uses the work 'NEW' in the certificate
2857 request header lines. Some software needs this.
2860 *) Reorganise password command line arguments: now passwords can be
2861 obtained from various sources. Delete the PEM_cb function and make
2862 it the default behaviour: i.e. if the callback is NULL and the
2863 usrdata argument is not NULL interpret it as a null terminated pass
2864 phrase. If usrdata and the callback are NULL then the pass phrase
2865 is prompted for as usual.
2868 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2869 the support is automatically enabled. The resulting binaries will
2870 autodetect the card and use it if present.
2871 [Ben Laurie and Compaq Inc.]
2873 *) Work around for Netscape hang bug. This sends certificate request
2874 and server done in one record. Since this is perfectly legal in the
2875 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2876 the bugs/SSLv3 entry for more info.
2879 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2882 *) Add -rand argument to smime and pkcs12 applications and read/write
2886 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2889 *) Add command line password options to the remaining applications.
2892 *) Bug fix for BN_div_recp() for numerators with an even number of
2896 *) More tests in bntest.c, and changed test_bn output.
2899 *) ./config recognizes MacOS X now.
2902 *) Bug fix for BN_div() when the first words of num and divsor are
2903 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2906 *) Add support for various broken PKCS#8 formats, and command line
2907 options to produce them.
2910 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2911 get temporary BIGNUMs from a BN_CTX.
2914 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2918 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2919 include a #define from the old name to the new. The original intent
2920 was that statically linked binaries could for example just call
2921 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2922 link with digests. This never worked becayse SSLeay_add_all_digests()
2923 and SSLeay_add_all_ciphers() were in the same source file so calling
2924 one would link with the other. They are now in separate source files.
2927 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2930 *) Use a less unusual form of the Miller-Rabin primality test (it used
2931 a binary algorithm for exponentiation integrated into the Miller-Rabin
2932 loop, our standard modexp algorithms are faster).
2935 *) Support for the EBCDIC character set completed.
2936 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2938 *) Source code cleanups: use const where appropriate, eliminate casts,
2939 use void * instead of char * in lhash.
2942 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2943 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2944 this the server could overwrite ephemeral keys that the client
2948 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2949 using 50 iterations of the Rabin-Miller test.
2951 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2952 iterations of the Rabin-Miller test as required by the appendix
2953 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2954 As BN_is_prime_fasttest includes trial division, DSA parameter
2955 generation becomes much faster.
2957 This implies a change for the callback functions in DSA_is_prime
2958 and DSA_generate_parameters: The callback function is called once
2959 for each positive witness in the Rabin-Miller test, not just
2960 occasionally in the inner loop; and the parameters to the
2961 callback function now provide an iteration count for the outer
2962 loop rather than for the current invocation of the inner loop.
2963 DSA_generate_parameters additionally can call the callback
2964 function with an 'iteration count' of -1, meaning that a
2965 candidate has passed the trial division test (when q is generated
2966 from an application-provided seed, trial division is skipped).
2969 *) New function BN_is_prime_fasttest that optionally does trial
2970 division before starting the Rabin-Miller test and has
2971 an additional BN_CTX * argument (whereas BN_is_prime always
2972 has to allocate at least one BN_CTX).
2973 'callback(1, -1, cb_arg)' is called when a number has passed the
2974 trial division stage.
2977 *) Fix for bug in CRL encoding. The validity dates weren't being handled
2981 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
2984 *) New function BN_pseudo_rand().
2987 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
2988 bignum version of BN_from_montgomery() with the working code from
2989 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
2993 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
2994 made it impossible to use the same SSL_SESSION data structure in
2995 SSL2 clients in multiple threads.
2998 *) The return value of RAND_load_file() no longer counts bytes obtained
2999 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3000 to seed the PRNG (previously an explicit byte count was required).
3001 [Ulf Möller, Bodo Möller]
3003 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3004 used (char *) instead of (void *) and had casts all over the place.
3007 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3010 *) Retain source code compatibility for BN_prime_checks macro:
3011 BN_is_prime(..., BN_prime_checks, ...) now uses
3012 BN_prime_checks_for_size to determine the appropriate number of
3013 Rabin-Miller iterations.
3016 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3017 DH_CHECK_P_NOT_SAFE_PRIME.
3018 (Check if this is true? OpenPGP calls them "strong".)
3021 *) Merge the functionality of "dh" and "gendh" programs into a new program
3022 "dhparam". The old programs are retained for now but will handle DH keys
3023 (instead of parameters) in future.
3026 *) Make the ciphers, s_server and s_client programs check the return values
3027 when a new cipher list is set.
3030 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3031 ciphers. Before when the 56bit ciphers were enabled the sorting was
3034 The syntax for the cipher sorting has been extended to support sorting by
3035 cipher-strength (using the strength_bits hard coded in the tables).
3036 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3038 Fix a bug in the cipher-command parser: when supplying a cipher command
3039 string with an "undefined" symbol (neither command nor alphanumeric
3040 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3041 an error is flagged.
3043 Due to the strength-sorting extension, the code of the
3044 ssl_create_cipher_list() function was completely rearranged. I hope that
3045 the readability was also increased :-)
3046 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3048 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3049 for the first serial number and places 2 in the serial number file. This
3050 avoids problems when the root CA is created with serial number zero and
3051 the first user certificate has the same issuer name and serial number
3055 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3056 the new code. Add documentation for this stuff.
3059 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3060 X509_*() to X509at_*() on the grounds that they don't handle X509
3061 structures and behave in an analagous way to the X509v3 functions:
3062 they shouldn't be called directly but wrapper functions should be used
3065 So we also now have some wrapper functions that call the X509at functions
3066 when passed certificate requests. (TO DO: similar things can be done with
3067 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3068 things. Some of these need some d2i or i2d and print functionality
3069 because they handle more complex structures.)
3072 *) Add missing #ifndefs that caused missing symbols when building libssl
3073 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3074 NO_RSA in ssl/s2*.c.
3075 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3077 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3078 has a return value which indicates the quality of the random data
3079 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3080 error queue. New function RAND_pseudo_bytes() generates output that is
3081 guaranteed to be unique but not unpredictable. RAND_add is like
3082 RAND_seed, but takes an extra argument for an entropy estimate
3083 (RAND_seed always assumes full entropy).
3086 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3087 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3088 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3089 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3090 false-positive rate of at most 2^-80 for random input.
3093 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3096 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3097 in the 0.9.5 release), this returns the chain
3098 from an X509_CTX structure with a dup of the stack and all
3099 the X509 reference counts upped: so the stack will exist
3100 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3103 Also make SSL_SESSION_print() print out the verify return
3107 *) Add manpage for the pkcs12 command. Also change the default
3108 behaviour so MAC iteration counts are used unless the new
3109 -nomaciter option is used. This improves file security and
3110 only older versions of MSIE (4.0 for example) need it.
3113 *) Honor the no-xxx Configure options when creating .DEF files.
3116 *) Add PKCS#10 attributes to field table: challengePassword,
3117 unstructuredName and unstructuredAddress. These are taken from
3118 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3119 international characters are used.
3121 More changes to X509_ATTRIBUTE code: allow the setting of types
3122 based on strings. Remove the 'loc' parameter when adding
3123 attributes because these will be a SET OF encoding which is sorted
3127 *) Initial changes to the 'req' utility to allow request generation
3128 automation. This will allow an application to just generate a template
3129 file containing all the field values and have req construct the
3132 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3133 used all over the place including certificate requests and PKCS#7
3134 structures. They are currently handled manually where necessary with
3135 some primitive wrappers for PKCS#7. The new functions behave in a
3136 manner analogous to the X509 extension functions: they allow
3137 attributes to be looked up by NID and added.
3139 Later something similar to the X509V3 code would be desirable to
3140 automatically handle the encoding, decoding and printing of the
3141 more complex types. The string types like challengePassword can
3142 be handled by the string table functions.
3144 Also modified the multi byte string table handling. Now there is
3145 a 'global mask' which masks out certain types. The table itself
3146 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3147 is useful when for example there is only one permissible type
3148 (as in countryName) and using the mask might result in no valid
3152 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3153 SSL_get_peer_finished to allow applications to obtain the latest
3154 Finished messages sent to the peer or expected from the peer,
3155 respectively. (SSL_get_peer_finished is usually the Finished message
3156 actually received from the peer, otherwise the protocol will be aborted.)
3158 As the Finished message are message digests of the complete handshake
3159 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3160 be used for external authentication procedures when the authentication
3161 provided by SSL/TLS is not desired or is not enough.
3164 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3165 the host supports BWX extension and if Compaq C is present on the
3166 $PATH. Just exploiting of the BWX extension results in 20-30%
3167 performance kick for some algorithms, e.g. DES and RC4 to mention
3168 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3172 *) Add support for MS "fast SGC". This is arguably a violation of the
3173 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3174 weak crypto and after checking the certificate is SGC a second one
3175 with strong crypto. MS SGC stops the first handshake after receiving
3176 the server certificate message and sends a second client hello. Since
3177 a server will typically do all the time consuming operations before
3178 expecting any further messages from the client (server key exchange
3179 is the most expensive) there is little difference between the two.
3181 To get OpenSSL to support MS SGC we have to permit a second client
3182 hello message after we have sent server done. In addition we have to
3183 reset the MAC if we do get this second client hello.
3186 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3187 if a DER encoded private key is RSA or DSA traditional format. Changed
3188 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3189 format DER encoded private key. Newer code should use PKCS#8 format which
3190 has the key type encoded in the ASN1 structure. Added DER private key
3191 support to pkcs8 application.
3194 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3195 ciphersuites has been selected (as required by the SSL 3/TLS 1
3196 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3197 is set, we interpret this as a request to violate the specification
3198 (the worst that can happen is a handshake failure, and 'correct'
3199 behaviour would result in a handshake failure anyway).
3202 *) In SSL_CTX_add_session, take into account that there might be multiple
3203 SSL_SESSION structures with the same session ID (e.g. when two threads
3204 concurrently obtain them from an external cache).
3205 The internal cache can handle only one SSL_SESSION with a given ID,
3206 so if there's a conflict, we now throw out the old one to achieve
3210 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3211 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3212 some routines that use cipher OIDs: some ciphers do not have OIDs
3213 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3217 *) Simplify the trust setting structure and code. Now we just have
3218 two sequences of OIDs for trusted and rejected settings. These will
3219 typically have values the same as the extended key usage extension
3220 and any application specific purposes.
3222 The trust checking code now has a default behaviour: it will just
3223 check for an object with the same NID as the passed id. Functions can
3224 be provided to override either the default behaviour or the behaviour
3225 for a given id. SSL client, server and email already have functions
3226 in place for compatibility: they check the NID and also return "trusted"
3227 if the certificate is self signed.
3230 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3231 traditional format into an EVP_PKEY structure.
3234 *) Add a password callback function PEM_cb() which either prompts for
3235 a password if usr_data is NULL or otherwise assumes it is a null
3236 terminated password. Allow passwords to be passed on command line
3237 environment or config files in a few more utilities.
3240 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3241 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3242 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3243 Update documentation.
3246 *) Support for ASN1 "NULL" type. This could be handled before by using
3247 ASN1_TYPE but there wasn't any function that would try to read a NULL
3248 and produce an error if it couldn't. For compatibility we also have
3249 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3250 don't allocate anything because they don't need to.
3253 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3255 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3257 *) Rebuild of the memory allocation routines used by OpenSSL code and
3258 possibly others as well. The purpose is to make an interface that
3259 provide hooks so anyone can build a separate set of allocation and
3260 deallocation routines to be used by OpenSSL, for example memory
3261 pool implementations, or something else, which was previously hard
3262 since Malloc(), Realloc() and Free() were defined as macros having
3263 the values malloc, realloc and free, respectively (except for Win32
3264 compilations). The same is provided for memory debugging code.
3265 OpenSSL already comes with functionality to find memory leaks, but
3266 this gives people a chance to debug other memory problems.
3268 With these changes, a new set of functions and macros have appeared:
3270 CRYPTO_set_mem_debug_functions() [F]
3271 CRYPTO_get_mem_debug_functions() [F]
3272 CRYPTO_dbg_set_options() [F]
3273 CRYPTO_dbg_get_options() [F]
3274 CRYPTO_malloc_debug_init() [M]
3276 The memory debug functions are NULL by default, unless the library
3277 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3278 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3279 gives the standard debugging functions that come with OpenSSL) or
3280 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3281 provided by the library user) must be used. When the standard
3282 debugging functions are used, CRYPTO_dbg_set_options can be used to
3283 request additional information:
3284 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3285 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3287 Also, things like CRYPTO_set_mem_functions will always give the
3288 expected result (the new set of functions is used for allocation
3289 and deallocation) at all times, regardless of platform and compiler
3292 To finish it up, some functions that were never use in any other
3293 way than through macros have a new API and new semantic:
3296 CRYPTO_dbg_realloc()
3299 All macros of value have retained their old syntax.
3300 [Richard Levitte and Bodo Moeller]
3302 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3303 ordering of SMIMECapabilities wasn't in "strength order" and there
3304 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3308 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3309 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3310 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3312 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3313 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3314 functionality to handle multipart/signed properly) and a utility
3315 called 'smime' to call all this stuff. This is based on code I
3316 originally wrote for Celo who have kindly allowed it to be
3317 included in OpenSSL.
3320 *) Add variants des_set_key_checked and des_set_key_unchecked of
3321 des_set_key (aka des_key_sched). Global variable des_check_key
3322 decides which of these is called by des_set_key; this way
3323 des_check_key behaves as it always did, but applications and
3324 the library itself, which was buggy for des_check_key == 1,
3325 have a cleaner way to pick the version they need.
3328 *) New function PKCS12_newpass() which changes the password of a
3332 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3333 dynamic mix. In both cases the ids can be used as an index into the
3334 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3335 functions so they accept a list of the field values and the
3336 application doesn't need to directly manipulate the X509_TRUST
3340 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3344 *) Modify the way the V3 extension code looks up extensions. This now
3345 works in a similar way to the object code: we have some "standard"
3346 extensions in a static table which is searched with OBJ_bsearch()
3347 and the application can add dynamic ones if needed. The file
3348 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3349 updated whenever a new extension is added to the core code and kept
3350 in ext_nid order. There is a simple program 'tabtest.c' which checks
3351 this. New extensions are not added too often so this file can readily
3352 be maintained manually.
3354 There are two big advantages in doing things this way. The extensions
3355 can be looked up immediately and no longer need to be "added" using
3356 X509V3_add_standard_extensions(): this function now does nothing.
3357 [Side note: I get *lots* of email saying the extension code doesn't
3358 work because people forget to call this function]
3359 Also no dynamic allocation is done unless new extensions are added:
3360 so if we don't add custom extensions there is no need to call
3361 X509V3_EXT_cleanup().
3364 *) Modify enc utility's salting as follows: make salting the default. Add a
3365 magic header, so unsalted files fail gracefully instead of just decrypting
3366 to garbage. This is because not salting is a big security hole, so people
3367 should be discouraged from doing it.
3370 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3371 digest to be passed on the command line but it only used this
3372 parameter when signing a certificate. Modified so all relevant
3373 operations are affected by the digest parameter including the
3374 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3375 DSA key was used because it didn't fix the digest.
3378 *) Initial certificate chain verify code. Currently tests the untrusted
3379 certificates for consistency with the verify purpose (which is set
3380 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3382 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3383 this is because it will reject chains with invalid extensions whereas
3384 every previous version of OpenSSL and SSLeay made no checks at all.
3386 Trust code: checks the root CA for the relevant trust settings. Trust
3387 settings have an initial value consistent with the verify purpose: e.g.
3388 if the verify purpose is for SSL client use it expects the CA to be
3389 trusted for SSL client use. However the default value can be changed to
3390 permit custom trust settings: one example of this would be to only trust
3391 certificates from a specific "secure" set of CAs.
3393 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3394 which should be used for version portability: especially since the
3395 verify structure is likely to change more often now.
3397 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3398 to set them. If not set then assume SSL clients will verify SSL servers
3401 Two new options to the verify program: -untrusted allows a set of
3402 untrusted certificates to be passed in and -purpose which sets the
3403 intended purpose of the certificate. If a purpose is set then the
3404 new chain verify code is used to check extension consistency.
3407 *) Support for the authority information access extension.
3410 *) Modify RSA and DSA PEM read routines to transparently handle
3411 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3412 public keys in a format compatible with certificate
3413 SubjectPublicKeyInfo structures. Unfortunately there were already
3414 functions called *_PublicKey_* which used various odd formats so
3415 these are retained for compatibility: however the DSA variants were
3416 never in a public release so they have been deleted. Changed dsa/rsa
3417 utilities to handle the new format: note no releases ever handled public
3418 keys so we should be OK.
3420 The primary motivation for this change is to avoid the same fiasco
3421 that dogs private keys: there are several incompatible private key
3422 formats some of which are standard and some OpenSSL specific and
3423 require various evil hacks to allow partial transparent handling and
3424 even then it doesn't work with DER formats. Given the option anything
3425 other than PKCS#8 should be dumped: but the other formats have to
3426 stay in the name of compatibility.
3428 With public keys and the benefit of hindsight one standard format
3429 is used which works with EVP_PKEY, RSA or DSA structures: though
3430 it clearly returns an error if you try to read the wrong kind of key.
3432 Added a -pubkey option to the 'x509' utility to output the public key.
3433 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3434 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3435 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3436 that do the same as the EVP_PKEY_assign_*() except they up the
3437 reference count of the added key (they don't "swallow" the
3441 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3442 CRLs would fail if the file contained no certificates or no CRLs:
3443 added a new function to read in both types and return the number
3444 read: this means that if none are read it will be an error. The
3445 DER versions of the certificate and CRL reader would always fail
3446 because it isn't possible to mix certificates and CRLs in DER format
3447 without choking one or the other routine. Changed this to just read
3448 a certificate: this is the best we can do. Also modified the code
3449 in apps/verify.c to take notice of return codes: it was previously
3450 attempting to read in certificates from NULL pointers and ignoring
3451 any errors: this is one reason why the cert and CRL reader seemed
3452 to work. It doesn't check return codes from the default certificate
3453 routines: these may well fail if the certificates aren't installed.
3456 *) Code to support otherName option in GeneralName.
3459 *) First update to verify code. Change the verify utility
3460 so it warns if it is passed a self signed certificate:
3461 for consistency with the normal behaviour. X509_verify
3462 has been modified to it will now verify a self signed
3463 certificate if *exactly* the same certificate appears
3464 in the store: it was previously impossible to trust a
3465 single self signed certificate. This means that:
3466 openssl verify ss.pem
3467 now gives a warning about a self signed certificate but
3468 openssl verify -CAfile ss.pem ss.pem
3472 *) For servers, store verify_result in SSL_SESSION data structure
3473 (and add it to external session representation).
3474 This is needed when client certificate verifications fails,
3475 but an application-provided verification callback (set by
3476 SSL_CTX_set_cert_verify_callback) allows accepting the session
3477 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3478 but returns 1): When the session is reused, we have to set
3479 ssl->verify_result to the appropriate error code to avoid
3481 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3483 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3484 case in PKCS7_dataInit() where the signed PKCS7 structure
3485 didn't contain any existing data because it was being created.
3486 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3488 *) Add a salt to the key derivation routines in enc.c. This
3489 forms the first 8 bytes of the encrypted file. Also add a
3490 -S option to allow a salt to be input on the command line.
3493 *) New function X509_cmp(). Oddly enough there wasn't a function
3494 to compare two certificates. We do this by working out the SHA1
3495 hash and comparing that. X509_cmp() will be needed by the trust
3499 *) SSL_get1_session() is like SSL_get_session(), but increments
3500 the reference count in the SSL_SESSION returned.
3501 [Geoff Thorpe <geoff@eu.c2.net>]
3503 *) Fix for 'req': it was adding a null to request attributes.
3504 Also change the X509_LOOKUP and X509_INFO code to handle
3505 certificate auxiliary information.
3508 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3512 *) Add the possibility to add extra information to the memory leak
3513 detecting output, to form tracebacks, showing from where each
3514 allocation was originated: CRYPTO_push_info("constant string") adds
3515 the string plus current file name and line number to a per-thread
3516 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3517 is like calling CYRPTO_pop_info() until the stack is empty.
3518 Also updated memory leak detection code to be multi-thread-safe.
3521 *) Add options -text and -noout to pkcs7 utility and delete the
3522 encryption options which never did anything. Update docs.
3525 *) Add options to some of the utilities to allow the pass phrase
3526 to be included on either the command line (not recommended on
3527 OSes like Unix) or read from the environment. Update the
3528 manpages and fix a few bugs.
3531 *) Add a few manpages for some of the openssl commands.
3534 *) Fix the -revoke option in ca. It was freeing up memory twice,
3535 leaking and not finding already revoked certificates.
3538 *) Extensive changes to support certificate auxiliary information.
3539 This involves the use of X509_CERT_AUX structure and X509_AUX
3540 functions. An X509_AUX function such as PEM_read_X509_AUX()
3541 can still read in a certificate file in the usual way but it
3542 will also read in any additional "auxiliary information". By
3543 doing things this way a fair degree of compatibility can be
3544 retained: existing certificates can have this information added
3545 using the new 'x509' options.
3547 Current auxiliary information includes an "alias" and some trust
3548 settings. The trust settings will ultimately be used in enhanced
3549 certificate chain verification routines: currently a certificate
3550 can only be trusted if it is self signed and then it is trusted
3554 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3555 The problem was that one of the replacement routines had not been working
3556 since SSLeay releases. For now the offending routine has been replaced
3557 with non-optimised assembler. Even so, this now gives around 95%
3558 performance improvement for 1024 bit RSA signs.
3561 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3562 handling. Most clients have the effective key size in bits equal to
3563 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3564 A few however don't do this and instead use the size of the decrypted key
3565 to determine the RC2 key length and the AlgorithmIdentifier to determine
3566 the effective key length. In this case the effective key length can still
3567 be 40 bits but the key length can be 168 bits for example. This is fixed
3568 by manually forcing an RC2 key into the EVP_PKEY structure because the
3569 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3570 the key length and effective key length are equal.
3573 *) Add a bunch of functions that should simplify the creation of
3574 X509_NAME structures. Now you should be able to do:
3575 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3576 and have it automatically work out the correct field type and fill in
3577 the structures. The more adventurous can try:
3578 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3579 and it will (hopefully) work out the correct multibyte encoding.
3582 *) Change the 'req' utility to use the new field handling and multibyte
3583 copy routines. Before the DN field creation was handled in an ad hoc
3584 way in req, ca, and x509 which was rather broken and didn't support
3585 BMPStrings or UTF8Strings. Since some software doesn't implement
3586 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3587 using the dirstring_type option. See the new comment in the default
3588 openssl.cnf for more info.
3591 *) Make crypto/rand/md_rand.c more robust:
3592 - Assure unique random numbers after fork().
3593 - Make sure that concurrent threads access the global counter and
3594 md serializably so that we never lose entropy in them
3595 or use exactly the same state in multiple threads.
3596 Access to the large state is not always serializable because
3597 the additional locking could be a performance killer, and
3598 md should be large enough anyway.
3601 *) New file apps/app_rand.c with commonly needed functionality
3602 for handling the random seed file.
3604 Use the random seed file in some applications that previously did not:
3606 dsaparam -genkey (which also ignored its '-rand' option),
3609 x509 (when signing).
3610 Except on systems with /dev/urandom, it is crucial to have a random
3611 seed file at least for key creation, DSA signing, and for DH exchanges;
3612 for RSA signatures we could do without one.
3614 gendh and gendsa (unlike genrsa) used to read only the first byte
3615 of each file listed in the '-rand' option. The function as previously
3616 found in genrsa is now in app_rand.c and is used by all programs
3617 that support '-rand'.
3620 *) In RAND_write_file, use mode 0600 for creating files;
3621 don't just chmod when it may be too late.
3624 *) Report an error from X509_STORE_load_locations
3625 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3628 *) New function ASN1_mbstring_copy() this copies a string in either
3629 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3630 into an ASN1_STRING type. A mask of permissible types is passed
3631 and it chooses the "minimal" type to use or an error if not type
3635 *) Add function equivalents to the various macros in asn1.h. The old
3636 macros are retained with an M_ prefix. Code inside the library can
3637 use the M_ macros. External code (including the openssl utility)
3638 should *NOT* in order to be "shared library friendly".
3641 *) Add various functions that can check a certificate's extensions
3642 to see if it usable for various purposes such as SSL client,
3643 server or S/MIME and CAs of these types. This is currently
3644 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3645 verification. Also added a -purpose flag to x509 utility to
3646 print out all the purposes.
3649 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3653 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3654 for, obtain and decode and extension and obtain its critical flag.
3655 This allows all the necessary extension code to be handled in a
3656 single function call.
3659 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3660 platforms. See crypto/rc4/rc4_enc.c for further details.
3663 *) New -noout option to asn1parse. This causes no output to be produced
3664 its main use is when combined with -strparse and -out to extract data
3665 from a file (which may not be in ASN.1 format).
3668 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3669 when producing the local key id.
3670 [Richard Levitte <levitte@stacken.kth.se>]
3672 *) New option -dhparam in s_server. This allows a DH parameter file to be
3673 stated explicitly. If it is not stated then it tries the first server
3674 certificate file. The previous behaviour hard coded the filename
3678 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3679 a public key to be input or output. For example:
3680 openssl rsa -in key.pem -pubout -out pubkey.pem
3681 Also added necessary DSA public key functions to handle this.
3684 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3685 in the message. This was handled by allowing
3686 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3687 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3689 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3690 to the end of the strings whereas this didn't. This would cause problems
3691 if strings read with d2i_ASN1_bytes() were later modified.
3692 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3694 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3695 data and it contains EOF it will end up returning an error. This is
3696 caused by input 46 bytes long. The cause is due to the way base64
3697 BIOs find the start of base64 encoded data. They do this by trying a
3698 trial decode on each line until they find one that works. When they
3699 do a flag is set and it starts again knowing it can pass all the
3700 data directly through the decoder. Unfortunately it doesn't reset
3701 the context it uses. This means that if EOF is reached an attempt
3702 is made to pass two EOFs through the context and this causes the
3703 resulting error. This can also cause other problems as well. As is
3704 usual with these problems it takes *ages* to find and the fix is
3705 trivial: move one line.
3706 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3708 *) Ugly workaround to get s_client and s_server working under Windows. The
3709 old code wouldn't work because it needed to select() on sockets and the
3710 tty (for keypresses and to see if data could be written). Win32 only
3711 supports select() on sockets so we select() with a 1s timeout on the
3712 sockets and then see if any characters are waiting to be read, if none
3713 are present then we retry, we also assume we can always write data to
3714 the tty. This isn't nice because the code then blocks until we've
3715 received a complete line of data and it is effectively polling the
3716 keyboard at 1s intervals: however it's quite a bit better than not
3717 working at all :-) A dedicated Windows application might handle this
3718 with an event loop for example.
3721 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3722 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3723 will be called when RSA_sign() and RSA_verify() are used. This is useful
3724 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3725 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3726 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3727 This necessitated the support of an extra signature type NID_md5_sha1
3728 for SSL signatures and modifications to the SSL library to use it instead
3729 of calling RSA_public_decrypt() and RSA_private_encrypt().
3732 *) Add new -verify -CAfile and -CApath options to the crl program, these
3733 will lookup a CRL issuers certificate and verify the signature in a
3734 similar way to the verify program. Tidy up the crl program so it
3735 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3736 less strict. It will now permit CRL extensions even if it is not
3737 a V2 CRL: this will allow it to tolerate some broken CRLs.
3740 *) Initialize all non-automatic variables each time one of the openssl
3741 sub-programs is started (this is necessary as they may be started
3742 multiple times from the "OpenSSL>" prompt).
3743 [Lennart Bang, Bodo Moeller]
3745 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3746 removing all other RSA functionality (this is what NO_RSA does). This
3747 is so (for example) those in the US can disable those operations covered
3748 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3752 *) Non-copying interface to BIO pairs.
3753 (still largely untested)
3756 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3757 ASCII string. This was handled independently in various places before.
3760 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3761 UTF8 strings a character at a time.
3764 *) Use client_version from client hello to select the protocol
3765 (s23_srvr.c) and for RSA client key exchange verification
3766 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3769 *) Add various utility functions to handle SPKACs, these were previously
3770 handled by poking round in the structure internals. Added new function
3771 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3772 print, verify and generate SPKACs. Based on an original idea from
3773 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3776 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3779 *) Allow the config file extension section to be overwritten on the
3780 command line. Based on an original idea from Massimiliano Pala
3781 <madwolf@comune.modena.it>. The new option is called -extensions
3782 and can be applied to ca, req and x509. Also -reqexts to override
3783 the request extensions in req and -crlexts to override the crl extensions
3787 *) Add new feature to the SPKAC handling in ca. Now you can include
3788 the same field multiple times by preceding it by "XXXX." for example:
3791 this is the same syntax as used in the req config file.
3794 *) Allow certificate extensions to be added to certificate requests. These
3795 are specified in a 'req_extensions' option of the req section of the
3796 config file. They can be printed out with the -text option to req but
3797 are otherwise ignored at present.
3800 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3801 data read consists of only the final block it would not decrypted because
3802 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3803 A misplaced 'break' also meant the decrypted final block might not be
3804 copied until the next read.
3807 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3808 a few extra parameters to the DH structure: these will be useful if
3809 for example we want the value of 'q' or implement X9.42 DH.
3812 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3813 provides hooks that allow the default DSA functions or functions on a
3814 "per key" basis to be replaced. This allows hardware acceleration and
3815 hardware key storage to be handled without major modification to the
3816 library. Also added low level modexp hooks and CRYPTO_EX structure and
3817 associated functions.
3820 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3821 as "read only": it can't be written to and the buffer it points to will
3822 not be freed. Reading from a read only BIO is much more efficient than
3823 a normal memory BIO. This was added because there are several times when
3824 an area of memory needs to be read from a BIO. The previous method was
3825 to create a memory BIO and write the data to it, this results in two
3826 copies of the data and an O(n^2) reading algorithm. There is a new
3827 function BIO_new_mem_buf() which creates a read only memory BIO from
3828 an area of memory. Also modified the PKCS#7 routines to use read only
3832 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3833 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3834 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3835 but a retry condition occured while trying to read the rest.
3838 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3839 NID_pkcs7_encrypted by default: this was wrong since this should almost
3840 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3841 the encrypted data type: this is a more sensible place to put it and it
3842 allows the PKCS#12 code to be tidied up that duplicated this
3846 *) Changed obj_dat.pl script so it takes its input and output files on
3847 the command line. This should avoid shell escape redirection problems
3851 *) Initial support for certificate extension requests, these are included
3852 in things like Xenroll certificate requests. Included functions to allow
3853 extensions to be obtained and added.
3856 *) -crlf option to s_client and s_server for sending newlines as
3857 CRLF (as required by many protocols).
3860 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3862 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3863 [Ralf S. Engelschall]
3865 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3866 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3868 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3872 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3873 DH parameters/keys (q is lost during that conversion, but the resulting
3874 DH parameters contain its length).
3876 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3877 much faster than DH_generate_parameters (which creates parameters
3878 where p = 2*q + 1), and also the smaller q makes DH computations
3879 much more efficient (160-bit exponentiation instead of 1024-bit
3880 exponentiation); so this provides a convenient way to support DHE
3881 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3882 utter importance to use
3883 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3885 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3886 when such DH parameters are used, because otherwise small subgroup
3887 attacks may become possible!
3890 *) Avoid memory leak in i2d_DHparams.
3893 *) Allow the -k option to be used more than once in the enc program:
3894 this allows the same encrypted message to be read by multiple recipients.
3897 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3898 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3899 it will always use the numerical form of the OID, even if it has a short
3903 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3904 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3905 otherwise bn_mod_exp was called. In the case of hardware keys for example
3906 no private key components need be present and it might store extra data
3907 in the RSA structure, which cannot be accessed from bn_mod_exp.
3908 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3909 private key operations.
3912 *) Added support for SPARC Linux.
3915 *) pem_password_cb function type incompatibly changed from
3916 typedef int pem_password_cb(char *buf, int size, int rwflag);
3918 ....(char *buf, int size, int rwflag, void *userdata);
3919 so that applications can pass data to their callbacks:
3920 The PEM[_ASN1]_{read,write}... functions and macros now take an
3921 additional void * argument, which is just handed through whenever
3922 the password callback is called.
3923 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3925 New function SSL_CTX_set_default_passwd_cb_userdata.
3927 Compatibility note: As many C implementations push function arguments
3928 onto the stack in reverse order, the new library version is likely to
3929 interoperate with programs that have been compiled with the old
3930 pem_password_cb definition (PEM_whatever takes some data that
3931 happens to be on the stack as its last argument, and the callback
3932 just ignores this garbage); but there is no guarantee whatsoever that
3935 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3936 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3937 problems not only on Windows, but also on some Unix platforms.
3938 To avoid problematic command lines, these definitions are now in an
3939 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3940 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3943 *) MIPS III/IV assembler module is reimplemented.
3946 *) More DES library cleanups: remove references to srand/rand and
3947 delete an unused file.
3950 *) Add support for the the free Netwide assembler (NASM) under Win32,
3951 since not many people have MASM (ml) and it can be hard to obtain.
3952 This is currently experimental but it seems to work OK and pass all
3953 the tests. Check out INSTALL.W32 for info.
3956 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3957 without temporary keys kept an extra copy of the server key,
3958 and connections with temporary keys did not free everything in case
3962 *) New function RSA_check_key and new openssl rsa option -check
3963 for verifying the consistency of RSA keys.
3964 [Ulf Moeller, Bodo Moeller]
3966 *) Various changes to make Win32 compile work:
3967 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
3968 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
3969 comparison" warnings.
3970 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
3973 *) Add a debugging option to PKCS#5 v2 key generation function: when
3974 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
3975 derived keys are printed to stderr.
3978 *) Copy the flags in ASN1_STRING_dup().
3979 [Roman E. Pavlov <pre@mo.msk.ru>]
3981 *) The x509 application mishandled signing requests containing DSA
3982 keys when the signing key was also DSA and the parameters didn't match.
3984 It was supposed to omit the parameters when they matched the signing key:
3985 the verifying software was then supposed to automatically use the CA's
3986 parameters if they were absent from the end user certificate.
3988 Omitting parameters is no longer recommended. The test was also
3989 the wrong way round! This was probably due to unusual behaviour in
3990 EVP_cmp_parameters() which returns 1 if the parameters match.
3991 This meant that parameters were omitted when they *didn't* match and
3992 the certificate was useless. Certificates signed with 'ca' didn't have
3994 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
3996 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
3997 The interface is as follows:
3998 Applications can use
3999 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4000 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4001 "off" is now the default.
4002 The library internally uses
4003 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4004 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4005 to disable memory-checking temporarily.
4007 Some inconsistent states that previously were possible (and were
4008 even the default) are now avoided.
4010 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4011 with each memory chunk allocated; this is occasionally more helpful
4012 than just having a counter.
4014 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4016 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4020 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4021 which largely parallels "options", but is for changing API behaviour,
4022 whereas "options" are about protocol behaviour.
4023 Initial "mode" flags are:
4025 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4026 a single record has been written.
4027 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4028 retries use the same buffer location.
4029 (But all of the contents must be
4033 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4036 *) Fix problems with no-hmac etc.
4037 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4039 *) New functions RSA_get_default_method(), RSA_set_method() and
4040 RSA_get_method(). These allows replacement of RSA_METHODs without having
4041 to mess around with the internals of an RSA structure.
4044 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4045 Also really enable memory leak checks in openssl.c and in some
4047 [Chad C. Mulligan, Bodo Moeller]
4049 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4050 up the length of negative integers. This has now been simplified to just
4051 store the length when it is first determined and use it later, rather
4052 than trying to keep track of where data is copied and updating it to
4054 [Steve Henson, reported by Brien Wheeler
4055 <bwheeler@authentica-security.com>]
4057 *) Add a new function PKCS7_signatureVerify. This allows the verification
4058 of a PKCS#7 signature but with the signing certificate passed to the
4059 function itself. This contrasts with PKCS7_dataVerify which assumes the
4060 certificate is present in the PKCS#7 structure. This isn't always the
4061 case: certificates can be omitted from a PKCS#7 structure and be
4062 distributed by "out of band" means (such as a certificate database).
4065 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4066 function prototypes in pem.h, also change util/mkdef.pl to add the
4067 necessary function names.
4070 *) mk1mf.pl (used by Windows builds) did not properly read the
4071 options set by Configure in the top level Makefile, and Configure
4072 was not even able to write more than one option correctly.
4073 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4076 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4077 file to be loaded from a BIO or FILE pointer. The BIO version will
4078 for example allow memory BIOs to contain config info.
4081 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4082 Whoever hopes to achieve shared-library compatibility across versions
4083 must use this, not the compile-time macro.
4084 (Exercise 0.9.4: Which is the minimum library version required by
4086 Note: All this applies only to multi-threaded programs, others don't
4090 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4091 through a BIO pair triggered the default case, i.e.
4092 SSLerr(...,SSL_R_UNKNOWN_STATE).
4095 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4096 can use the SSL library even if none of the specific BIOs is
4100 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4101 for the encoded length.
4102 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4104 *) Add initial documentation of the X509V3 functions.
4107 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4108 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4109 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4110 secure PKCS#8 private key format with a high iteration count.
4113 *) Fix determination of Perl interpreter: A perl or perl5
4114 _directory_ in $PATH was also accepted as the interpreter.
4115 [Ralf S. Engelschall]
4117 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4118 wrong with it but it was very old and did things like calling
4119 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4123 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4124 to use the new extension code.
4127 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4128 with macros. This should make it easier to change their form, add extra
4129 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4133 *) Add to configuration table a new entry that can specify an alternative
4134 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4135 according to Mark Crispin <MRC@Panda.COM>.
4139 *) DES CBC did not update the IV. Weird.
4142 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4143 Changing the behaviour of the former might break existing programs --
4144 where IV updating is needed, des_ncbc_encrypt can be used.
4147 *) When bntest is run from "make test" it drives bc to check its
4148 calculations, as well as internally checking them. If an internal check
4149 fails, it needs to cause bc to give a non-zero result or make test carries
4150 on without noticing the failure. Fixed.
4153 *) DES library cleanups.
4156 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4157 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4158 ciphers. NOTE: although the key derivation function has been verified
4159 against some published test vectors it has not been extensively tested
4160 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4164 *) Instead of "mkdir -p", which is not fully portable, use new
4165 Perl script "util/mkdir-p.pl".
4168 *) Rewrite the way password based encryption (PBE) is handled. It used to
4169 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4170 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4171 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4172 the 'parameter' field of the AlgorithmIdentifier is passed to the
4173 underlying key generation function so it must do its own ASN1 parsing.
4174 This has also changed the EVP_PBE_CipherInit() function which now has a
4175 'parameter' argument instead of literal salt and iteration count values
4176 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4179 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4180 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4181 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4182 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4183 value was just used as a "magic string" and not used directly its
4184 value doesn't matter.
4187 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4191 *) "linux-sparc64" configuration (ultrapenguin).
4192 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4193 "linux-sparc" configuration.
4194 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4196 *) config now generates no-xxx options for missing ciphers.
4199 *) Support the EBCDIC character set (work in progress).
4200 File ebcdic.c not yet included because it has a different license.
4201 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4203 *) Support BS2000/OSD-POSIX.
4204 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4206 *) Make callbacks for key generation use void * instead of char *.
4209 *) Make S/MIME samples compile (not yet tested).
4212 *) Additional typesafe stacks.
4215 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4219 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4221 *) New configuration variant "sco5-gcc".
4223 *) Updated some demos.
4224 [Sean O Riordain, Wade Scholine]
4226 *) Add missing BIO_free at exit of pkcs12 application.
4229 *) Fix memory leak in conf.c.
4232 *) Updates for Win32 to assembler version of MD5.
4235 *) Set #! path to perl in apps/der_chop to where we found it
4236 instead of using a fixed path.
4239 *) SHA library changes for irix64-mips4-cc.
4242 *) Improvements for VMS support.
4246 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4248 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4249 This also avoids the problems with SC4.2 and unpatched SC5.
4250 [Andy Polyakov <appro@fy.chalmers.se>]
4252 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4253 These are required because of the typesafe stack would otherwise break
4254 existing code. If old code used a structure member which used to be STACK
4255 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4256 sk_num or sk_value it would produce an error because the num, data members
4257 are not present in STACK_OF. Now it just produces a warning. sk_set
4258 replaces the old method of assigning a value to sk_value
4259 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4260 that does this will no longer work (and should use sk_set instead) but
4261 this could be regarded as a "questionable" behaviour anyway.
4264 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4265 correctly handle encrypted S/MIME data.
4268 *) Change type of various DES function arguments from des_cblock
4269 (which means, in function argument declarations, pointer to char)
4270 to des_cblock * (meaning pointer to array with 8 char elements),
4271 which allows the compiler to do more typechecking; it was like
4272 that back in SSLeay, but with lots of ugly casts.
4274 Introduce new type const_des_cblock.
4277 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4278 problems: find RecipientInfo structure that matches recipient certificate
4279 and initialise the ASN1 structures properly based on passed cipher.
4282 *) Belatedly make the BN tests actually check the results.
4285 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4286 to and from BNs: it was completely broken. New compilation option
4287 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4288 key elements as negative integers.
4291 *) Reorganize and speed up MD5.
4292 [Andy Polyakov <appro@fy.chalmers.se>]
4295 [Richard Levitte <richard@levitte.org>]
4297 *) New option -out to asn1parse to allow the parsed structure to be
4298 output to a file. This is most useful when combined with the -strparse
4299 option to examine the output of things like OCTET STRINGS.
4302 *) Make SSL library a little more fool-proof by not requiring any longer
4303 that SSL_set_{accept,connect}_state be called before
4304 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4305 in many applications because usually everything *appeared* to work as
4306 intended anyway -- now it really works as intended).
4309 *) Move openssl.cnf out of lib/.
4312 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4313 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4314 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4315 [Ralf S. Engelschall]
4317 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4318 handle PKCS#7 enveloped data properly.
4319 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4321 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4322 copying pointers. The cert_st handling is changed by this in
4323 various ways (and thus what used to be known as ctx->default_cert
4324 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4325 any longer when s->cert does not give us what we need).
4326 ssl_cert_instantiate becomes obsolete by this change.
4327 As soon as we've got the new code right (possibly it already is?),
4328 we have solved a couple of bugs of the earlier code where s->cert
4329 was used as if it could not have been shared with other SSL structures.
4331 Note that using the SSL API in certain dirty ways now will result
4332 in different behaviour than observed with earlier library versions:
4333 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4334 does not influence s as it used to.
4336 In order to clean up things more thoroughly, inside SSL_SESSION
4337 we don't use CERT any longer, but a new structure SESS_CERT
4338 that holds per-session data (if available); currently, this is
4339 the peer's certificate chain and, for clients, the server's certificate
4340 and temporary key. CERT holds only those values that can have
4341 meaningful defaults in an SSL_CTX.
4344 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4345 from the internal representation. Various PKCS#7 fixes: remove some
4346 evil casts and set the enc_dig_alg field properly based on the signing
4350 *) Allow PKCS#12 password to be set from the command line or the
4351 environment. Let 'ca' get its config file name from the environment
4352 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4356 *) Allow certificate policies extension to use an IA5STRING for the
4357 organization field. This is contrary to the PKIX definition but
4358 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4362 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4363 without disallowing inline assembler and the like for non-pedantic builds.
4366 *) Support Borland C++ builder.
4367 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4372 *) SHA-1 cleanups and performance enhancements.
4373 [Andy Polyakov <appro@fy.chalmers.se>]
4375 *) Sparc v8plus assembler for the bignum library.
4376 [Andy Polyakov <appro@fy.chalmers.se>]
4378 *) Accept any -xxx and +xxx compiler options in Configure.
4381 *) Update HPUX configuration.
4384 *) Add missing sk_<type>_unshift() function to safestack.h
4385 [Ralf S. Engelschall]
4387 *) New function SSL_CTX_use_certificate_chain_file that sets the
4388 "extra_cert"s in addition to the certificate. (This makes sense
4389 only for "PEM" format files, as chains as a whole are not
4393 *) Support verify_depth from the SSL API.
4394 x509_vfy.c had what can be considered an off-by-one-error:
4395 Its depth (which was not part of the external interface)
4396 was actually counting the number of certificates in a chain;
4397 now it really counts the depth.
4400 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4401 instead of X509err, which often resulted in confusing error
4402 messages since the error codes are not globally unique
4403 (e.g. an alleged error in ssl3_accept when a certificate
4404 didn't match the private key).
4406 *) New function SSL_CTX_set_session_id_context that allows to set a default
4407 value (so that you don't need SSL_set_session_id_context for each
4408 connection using the SSL_CTX).
4411 *) OAEP decoding bug fix.
4414 *) Support INSTALL_PREFIX for package builders, as proposed by
4418 *) New Configure options "threads" and "no-threads". For systems
4419 where the proper compiler options are known (currently Solaris
4420 and Linux), "threads" is the default.
4423 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4426 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4427 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4428 such as /usr/local/bin.
4431 *) "make linux-shared" to build shared libraries.
4432 [Niels Poppe <niels@netbox.org>]
4434 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4437 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4438 extension adding in x509 utility.
4441 *) Remove NOPROTO sections and error code comments.
4444 *) Partial rewrite of the DEF file generator to now parse the ANSI
4448 *) New Configure options --prefix=DIR and --openssldir=DIR.
4451 *) Complete rewrite of the error code script(s). It is all now handled
4452 by one script at the top level which handles error code gathering,
4453 header rewriting and C source file generation. It should be much better
4454 than the old method: it now uses a modified version of Ulf's parser to
4455 read the ANSI prototypes in all header files (thus the old K&R definitions
4456 aren't needed for error creation any more) and do a better job of
4457 translating function codes into names. The old 'ASN1 error code imbedded
4458 in a comment' is no longer necessary and it doesn't use .err files which
4459 have now been deleted. Also the error code call doesn't have to appear all
4460 on one line (which resulted in some large lines...).
4463 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4466 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4467 0 (which usually indicates a closed connection), but continue reading.
4470 *) Fix some race conditions.
4473 *) Add support for CRL distribution points extension. Add Certificate
4474 Policies and CRL distribution points documentation.
4477 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4480 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4481 8 of keying material. Merlin has also confirmed interop with this fix
4482 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4483 [Merlin Hughes <merlin@baltimore.ie>]
4485 *) Fix lots of warnings.
4486 [Richard Levitte <levitte@stacken.kth.se>]
4488 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4489 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4490 [Richard Levitte <levitte@stacken.kth.se>]
4492 *) Fix problems with sizeof(long) == 8.
4493 [Andy Polyakov <appro@fy.chalmers.se>]
4495 *) Change functions to ANSI C.
4498 *) Fix typos in error codes.
4499 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4501 *) Remove defunct assembler files from Configure.
4504 *) SPARC v8 assembler BIGNUM implementation.
4505 [Andy Polyakov <appro@fy.chalmers.se>]
4507 *) Support for Certificate Policies extension: both print and set.
4508 Various additions to support the r2i method this uses.
4511 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4512 return a const string when you are expecting an allocated buffer.
4515 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4516 types DirectoryString and DisplayText.
4519 *) Add code to allow r2i extensions to access the configuration database,
4520 add an LHASH database driver and add several ctx helper functions.
4523 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4524 fail when they extended the size of a BIGNUM.
4527 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4528 support typesafe stack.
4531 *) Fix typo in SSL_[gs]et_options().
4532 [Nils Frostberg <nils@medcom.se>]
4534 *) Delete various functions and files that belonged to the (now obsolete)
4535 old X509V3 handling code.
4538 *) New Configure option "rsaref".
4541 *) Don't auto-generate pem.h.
4544 *) Introduce type-safe ASN.1 SETs.
4547 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4548 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4550 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4551 that links with OpenSSL (well at least cause lots of warnings), but fear
4552 not: the conversion is trivial, and it eliminates loads of evil casts. A
4553 few STACKed things have been converted already. Feel free to convert more.
4554 In the fullness of time, I'll do away with the STACK type altogether.
4557 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4558 specified in <certfile> by updating the entry in the index.txt file.
4559 This way one no longer has to edit the index.txt file manually for
4560 revoking a certificate. The -revoke option does the gory details now.
4561 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4563 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4564 `-text' option at all and this way the `-noout -text' combination was
4565 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4566 [Ralf S. Engelschall]
4568 *) Make sure a corresponding plain text error message exists for the
4569 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4570 verify callback function determined that a certificate was revoked.
4571 [Ralf S. Engelschall]
4573 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4574 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4575 all available cipers including rc5, which was forgotten until now.
4576 In order to let the testing shell script know which algorithms
4577 are available, a new (up to now undocumented) command
4578 "openssl list-cipher-commands" is used.
4581 *) Bugfix: s_client occasionally would sleep in select() when
4582 it should have checked SSL_pending() first.
4585 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4586 the raw DSA values prior to ASN.1 encoding.
4589 *) Tweaks to Configure
4590 [Niels Poppe <niels@netbox.org>]
4592 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4596 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4599 *) New config option to avoid instructions that are illegal on the 80386.
4600 The default code is faster, but requires at least a 486.
4603 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4604 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4605 same as SSL2_VERSION anyway.
4608 *) New "-showcerts" option for s_client.
4611 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4612 application. Various cleanups and fixes.
4615 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4616 modify error routines to work internally. Add error codes and PBE init
4617 to library startup routines.
4620 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4621 packing functions to asn1 and evp. Changed function names and error
4622 codes along the way.
4625 *) PKCS12 integration: and so it begins... First of several patches to
4626 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4627 objects to objects.h
4630 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4631 and display support for Thawte strong extranet extension.
4634 *) Add LinuxPPC support.
4635 [Jeff Dubrule <igor@pobox.org>]
4637 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4638 bn_div_words in alpha.s.
4639 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4641 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4642 OAEP isn't supported when OpenSSL is built with RSAref.
4643 [Ulf Moeller <ulf@fitug.de>]
4645 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4646 so they no longer are missing under -DNOPROTO.
4647 [Soren S. Jorvang <soren@t.dk>]
4650 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4652 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4653 doesn't work when the session is reused. Coming soon!
4656 *) Fix a security hole, that allows sessions to be reused in the wrong
4657 context thus bypassing client cert protection! All software that uses
4658 client certs and session caches in multiple contexts NEEDS PATCHING to
4659 allow session reuse! A fuller solution is in the works.
4660 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4662 *) Some more source tree cleanups (removed obsolete files
4663 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4664 permission on "config" script to be executable) and a fix for the INSTALL
4666 [Ulf Moeller <ulf@fitug.de>]
4668 *) Remove some legacy and erroneous uses of malloc, free instead of
4670 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4672 *) Make rsa_oaep_test return non-zero on error.
4673 [Ulf Moeller <ulf@fitug.de>]
4675 *) Add support for native Solaris shared libraries. Configure
4676 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4677 if someone would make that last step automatic.
4678 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4680 *) ctx_size was not built with the right compiler during "make links". Fixed.
4683 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4684 except NULL ciphers". This means the default cipher list will no longer
4685 enable NULL ciphers. They need to be specifically enabled e.g. with
4686 the string "DEFAULT:eNULL".
4689 *) Fix to RSA private encryption routines: if p < q then it would
4690 occasionally produce an invalid result. This will only happen with
4691 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4694 *) Be less restrictive and allow also `perl util/perlpath.pl
4695 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4696 because this way one can also use an interpreter named `perl5' (which is
4697 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4698 installed as `perl').
4699 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4701 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4702 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4704 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4705 advapi32.lib to Win32 build and change the pem test comparision
4706 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4707 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4708 and crypto/des/ede_cbcm_enc.c.
4711 *) DES quad checksum was broken on big-endian architectures. Fixed.
4714 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4715 Win32 test batch file so it (might) work again. The Win32 test batch file
4716 is horrible: I feel ill....
4719 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4720 in e_os.h. Audit of header files to check ANSI and non ANSI
4721 sections: 10 functions were absent from non ANSI section and not exported
4722 from Windows DLLs. Fixed up libeay.num for new functions.
4725 *) Make `openssl version' output lines consistent.
4726 [Ralf S. Engelschall]
4728 *) Fix Win32 symbol export lists for BIO functions: Added
4729 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4730 to ms/libeay{16,32}.def.
4731 [Ralf S. Engelschall]
4733 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4734 fine under Unix and passes some trivial tests I've now added. But the
4735 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4736 added to make sure no one expects that this stuff really works in the
4737 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4738 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4740 [Ralf S. Engelschall]
4742 *) Fix the generation of two part addresses in perl.
4743 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4745 *) Add config entry for Linux on MIPS.
4746 [John Tobey <jtobey@channel1.com>]
4748 *) Make links whenever Configure is run, unless we are on Windoze.
4751 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4752 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4756 *) Add a useful kludge to allow package maintainers to specify compiler and
4757 other platforms details on the command line without having to patch the
4758 Configure script everytime: One now can use ``perl Configure
4759 <id>:<details>'', i.e. platform ids are allowed to have details appended
4760 to them (seperated by colons). This is treated as there would be a static
4761 pre-configured entry in Configure's %table under key <id> with value
4762 <details> and ``perl Configure <id>'' is called. So, when you want to
4763 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4764 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4765 now, which overrides the FreeBSD-elf entry on-the-fly.
4766 [Ralf S. Engelschall]
4768 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4771 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4772 on the `perl Configure ...' command line. This way one can compile
4773 OpenSSL libraries with Position Independent Code (PIC) which is needed
4774 for linking it into DSOs.
4775 [Ralf S. Engelschall]
4777 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4781 *) Cleaned up the LICENSE document: The official contact for any license
4782 questions now is the OpenSSL core team under openssl-core@openssl.org.
4783 And add a paragraph about the dual-license situation to make sure people
4784 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4785 to the OpenSSL toolkit.
4786 [Ralf S. Engelschall]
4788 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4789 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4790 Additonally cleaned up the `make links' target: Remove unnecessary
4791 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4792 to speed processing and no longer clutter the display with confusing
4793 stuff. Instead only the actually done links are displayed.
4794 [Ralf S. Engelschall]
4796 *) Permit null encryption ciphersuites, used for authentication only. It used
4797 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4798 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4802 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4803 signed attributes when verifying signatures (this would break them),
4804 the detached data encoding was wrong and public keys obtained using
4805 X509_get_pubkey() weren't freed.
4808 *) Add text documentation for the BUFFER functions. Also added a work around
4809 to a Win95 console bug. This was triggered by the password read stuff: the
4810 last character typed gets carried over to the next fread(). If you were
4811 generating a new cert request using 'req' for example then the last
4812 character of the passphrase would be CR which would then enter the first
4816 *) Added the new `Includes OpenSSL Cryptography Software' button as
4817 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4818 button and can be used by applications based on OpenSSL to show the
4819 relationship to the OpenSSL project.
4820 [Ralf S. Engelschall]
4822 *) Remove confusing variables in function signatures in files
4823 ssl/ssl_lib.c and ssl/ssl.h.
4824 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4826 *) Don't install bss_file.c under PREFIX/include/
4827 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4829 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4830 functions that return function pointers and has support for NT specific
4831 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4832 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4833 unsigned to signed types: this was killing the Win32 compile.
4836 *) Add new certificate file to stack functions,
4837 SSL_add_dir_cert_subjects_to_stack() and
4838 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4839 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4840 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4841 This means that Apache-SSL and similar packages don't have to mess around
4842 to add as many CAs as they want to the preferred list.
4845 *) Experiment with doxygen documentation. Currently only partially applied to
4847 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4848 openssl.doxy as the configuration file.
4851 *) Get rid of remaining C++-style comments which strict C compilers hate.
4852 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4854 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4855 compiled in by default: it has problems with large keys.
4858 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4859 DH private keys and/or callback functions which directly correspond to
4860 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4861 is needed for applications which have to configure certificates on a
4862 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4864 For the RSA certificate situation is makes no difference, but
4865 for the DSA certificate situation this fixes the "no shared cipher"
4866 problem where the OpenSSL cipher selection procedure failed because the
4867 temporary keys were not overtaken from the context and the API provided
4868 no way to reconfigure them.
4869 The new functions now let applications reconfigure the stuff and they
4870 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4871 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4872 non-public-API function ssl_cert_instantiate() is used as a helper
4873 function and also to reduce code redundancy inside ssl_rsa.c.
4874 [Ralf S. Engelschall]
4876 *) Move s_server -dcert and -dkey options out of the undocumented feature
4877 area because they are useful for the DSA situation and should be
4878 recognized by the users.
4879 [Ralf S. Engelschall]
4881 *) Fix the cipher decision scheme for export ciphers: the export bits are
4882 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4883 SSL_EXP_MASK. So, the original variable has to be used instead of the
4884 already masked variable.
4885 [Richard Levitte <levitte@stacken.kth.se>]
4887 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4888 [Richard Levitte <levitte@stacken.kth.se>]
4890 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4891 from `int' to `unsigned int' because it's a length and initialized by
4892 EVP_DigestFinal() which expects an `unsigned int *'.
4893 [Richard Levitte <levitte@stacken.kth.se>]
4895 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4896 script. Instead use the usual Shell->Perl transition trick.
4897 [Ralf S. Engelschall]
4899 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4900 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4901 -noout -modulus' as it's already the case for `openssl rsa -noout
4902 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4903 currently the public key is printed (a decision which was already done by
4904 `openssl dsa -modulus' in the past) which serves a similar purpose.
4905 Additionally the NO_RSA no longer completely removes the whole -modulus
4906 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4908 [Ralf S. Engelschall]
4910 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4911 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4912 [Arne Ansper <arne@ats.cyber.ee>]
4914 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4915 to be added. Now both 'req' and 'ca' can use new objects defined in the
4919 *) Add cool BIO that does syslog (or event log on NT).
4920 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4922 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4923 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4924 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4925 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4928 *) Add preliminary config info for new extension code.
4931 *) Make RSA_NO_PADDING really use no padding.
4932 [Ulf Moeller <ulf@fitug.de>]
4934 *) Generate errors when private/public key check is done.
4937 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4938 for some CRL extensions and new objects added.
4941 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4942 key usage extension and fuller support for authority key id.
4945 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4946 padding method for RSA, which is recommended for new applications in PKCS
4947 #1 v2.0 (RFC 2437, October 1998).
4948 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4949 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4950 against Bleichbacher's attack on RSA.
4951 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4954 *) Updates to the new SSL compression code
4955 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4957 *) Fix so that the version number in the master secret, when passed
4958 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
4959 (because the server will not accept higher), that the version number
4960 is 0x03,0x01, not 0x03,0x00
4961 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4963 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
4964 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
4965 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
4968 *) Support for RAW extensions where an arbitrary extension can be
4969 created by including its DER encoding. See apps/openssl.cnf for
4973 *) Make sure latest Perl versions don't interpret some generated C array
4974 code as Perl array code in the crypto/err/err_genc.pl script.
4975 [Lars Weber <3weber@informatik.uni-hamburg.de>]
4977 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
4978 not many people have the assembler. Various Win32 compilation fixes and
4979 update to the INSTALL.W32 file with (hopefully) more accurate Win32
4983 *) Modify configure script 'Configure' to automatically create crypto/date.h
4984 file under Win32 and also build pem.h from pem.org. New script
4985 util/mkfiles.pl to create the MINFO file on environments that can't do a
4986 'make files': perl util/mkfiles.pl >MINFO should work.
4989 *) Major rework of DES function declarations, in the pursuit of correctness
4990 and purity. As a result, many evil casts evaporated, and some weirdness,
4991 too. You may find this causes warnings in your code. Zapping your evil
4992 casts will probably fix them. Mostly.
4995 *) Fix for a typo in asn1.h. Bug fix to object creation script
4996 obj_dat.pl. It considered a zero in an object definition to mean
4997 "end of object": none of the objects in objects.h have any zeros
4998 so it wasn't spotted.
4999 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5001 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5002 Masking (CBCM). In the absence of test vectors, the best I have been able
5003 to do is check that the decrypt undoes the encrypt, so far. Send me test
5004 vectors if you have them.
5007 *) Correct calculation of key length for export ciphers (too much space was
5008 allocated for null ciphers). This has not been tested!
5011 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5012 message is now correct (it understands "crypto" and "ssl" on its
5013 command line). There is also now an "update" option. This will update
5014 the util/ssleay.num and util/libeay.num files with any new functions.
5016 perl util/mkdef.pl crypto ssl update
5017 it will update them.
5020 *) Overhauled the Perl interface (perl/*):
5021 - ported BN stuff to OpenSSL's different BN library
5022 - made the perl/ source tree CVS-aware
5023 - renamed the package from SSLeay to OpenSSL (the files still contain
5024 their history because I've copied them in the repository)
5025 - removed obsolete files (the test scripts will be replaced
5026 by better Test::Harness variants in the future)
5027 [Ralf S. Engelschall]
5029 *) First cut for a very conservative source tree cleanup:
5030 1. merge various obsolete readme texts into doc/ssleay.txt
5031 where we collect the old documents and readme texts.
5032 2. remove the first part of files where I'm already sure that we no
5033 longer need them because of three reasons: either they are just temporary
5034 files which were left by Eric or they are preserved original files where
5035 I've verified that the diff is also available in the CVS via "cvs diff
5036 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5037 the crypto/md/ stuff).
5038 [Ralf S. Engelschall]
5040 *) More extension code. Incomplete support for subject and issuer alt
5041 name, issuer and authority key id. Change the i2v function parameters
5042 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5043 what that's for :-) Fix to ASN1 macro which messed up
5044 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5047 *) Preliminary support for ENUMERATED type. This is largely copied from the
5051 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5052 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5054 *) Make sure `make rehash' target really finds the `openssl' program.
5055 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5057 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5058 like to hear about it if this slows down other processors.
5061 *) Add CygWin32 platform information to Configure script.
5062 [Alan Batie <batie@aahz.jf.intel.com>]
5064 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5065 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5067 *) New program nseq to manipulate netscape certificate sequences
5070 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5074 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5075 but the BN code had some problems that would cause failures when
5076 doing certificate verification and some other functions.
5077 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5079 *) Add ASN1 and PEM code to support netscape certificate sequences.
5082 *) Add ASN1 and PEM code to support netscape certificate sequences.
5085 *) Add several PKIX and private extended key usage OIDs.
5088 *) Modify the 'ca' program to handle the new extension code. Modify
5089 openssl.cnf for new extension format, add comments.
5092 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5093 and add a sample to openssl.cnf so req -x509 now adds appropriate
5097 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5098 error code, add initial support to X509_print() and x509 application.
5101 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5102 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5103 stuff is currently isolated and isn't even compiled yet.
5106 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5107 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5108 Removed the versions check from X509 routines when loading extensions:
5109 this allows certain broken certificates that don't set the version
5110 properly to be processed.
5113 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5114 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5115 can still be regenerated with "make depend".
5118 *) Spelling mistake in C version of CAST-128.
5119 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5121 *) Changes to the error generation code. The perl script err-code.pl
5122 now reads in the old error codes and retains the old numbers, only
5123 adding new ones if necessary. It also only changes the .err files if new
5124 codes are added. The makefiles have been modified to only insert errors
5125 when needed (to avoid needlessly modifying header files). This is done
5126 by only inserting errors if the .err file is newer than the auto generated
5127 C file. To rebuild all the error codes from scratch (the old behaviour)
5128 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5129 or delete all the .err files.
5132 *) CAST-128 was incorrectly implemented for short keys. The C version has
5133 been fixed, but is untested. The assembler versions are also fixed, but
5134 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5135 to regenerate it if needed.
5136 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5137 Hagino <itojun@kame.net>]
5139 *) File was opened incorrectly in randfile.c.
5140 [Ulf Möller <ulf@fitug.de>]
5142 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5143 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5144 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5145 al: it's just almost always a UTCTime. Note this patch adds new error
5146 codes so do a "make errors" if there are problems.
5149 *) Correct Linux 1 recognition in config.
5150 [Ulf Möller <ulf@fitug.de>]
5152 *) Remove pointless MD5 hash when using DSA keys in ca.
5153 [Anonymous <nobody@replay.com>]
5155 *) Generate an error if given an empty string as a cert directory. Also
5156 generate an error if handed NULL (previously returned 0 to indicate an
5157 error, but didn't set one).
5158 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5160 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5163 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5164 parameters. This was causing a warning which killed off the Win32 compile.
5167 *) Remove C++ style comments from crypto/bn/bn_local.h.
5168 [Neil Costigan <neil.costigan@celocom.com>]
5170 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5171 based on a text string, looking up short and long names and finally
5172 "dot" format. The "dot" format stuff didn't work. Added new function
5173 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5174 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5175 OID is not part of the table.
5178 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5179 X509_LOOKUP_by_alias().
5182 *) Sort openssl functions by name.
5185 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5186 encryption from sample DSA keys (in case anyone is interested the password
5190 *) Make _all_ *_free functions accept a NULL pointer.
5191 [Frans Heymans <fheymans@isaserver.be>]
5193 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5195 [Anonymous <nobody@replay.com>]
5197 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5198 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5200 *) Don't blow it for numeric -newkey arguments to apps/req.
5201 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5203 *) Temp key "for export" tests were wrong in s3_srvr.c.
5204 [Anonymous <nobody@replay.com>]
5206 *) Add prototype for temp key callback functions
5207 SSL_CTX_set_tmp_{rsa,dh}_callback().
5210 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5211 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5214 *) X509_name_add_entry() freed the wrong thing after an error.
5215 [Arne Ansper <arne@ats.cyber.ee>]
5217 *) rsa_eay.c would attempt to free a NULL context.
5218 [Arne Ansper <arne@ats.cyber.ee>]
5220 *) BIO_s_socket() had a broken should_retry() on Windoze.
5221 [Arne Ansper <arne@ats.cyber.ee>]
5223 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5224 [Arne Ansper <arne@ats.cyber.ee>]
5226 *) Make sure the already existing X509_STORE->depth variable is initialized
5227 in X509_STORE_new(), but document the fact that this variable is still
5228 unused in the certificate verification process.
5229 [Ralf S. Engelschall]
5231 *) Fix the various library and apps files to free up pkeys obtained from
5232 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5235 *) Fix reference counting in X509_PUBKEY_get(). This makes
5236 demos/maurice/example2.c work, amongst others, probably.
5237 [Steve Henson and Ben Laurie]
5239 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5240 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5241 are no longer created. This way we have a single and consistent command
5242 line interface `openssl <command>', similar to `cvs <command>'.
5243 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5245 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5246 BIT STRING wrapper always have zero unused bits.
5249 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5252 *) Make the top-level INSTALL documentation easier to understand.
5255 *) Makefiles updated to exit if an error occurs in a sub-directory
5256 make (including if user presses ^C) [Paul Sutton]
5258 *) Make Montgomery context stuff explicit in RSA data structure.
5261 *) Fix build order of pem and err to allow for generated pem.h.
5264 *) Fix renumbering bug in X509_NAME_delete_entry().
5267 *) Enhanced the err-ins.pl script so it makes the error library number
5268 global and can add a library name. This is needed for external ASN1 and
5269 other error libraries.
5272 *) Fixed sk_insert which never worked properly.
5275 *) Fix ASN1 macros so they can handle indefinite length construted
5276 EXPLICIT tags. Some non standard certificates use these: they can now
5280 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5281 into a single doc/ssleay.txt bundle. This way the information is still
5282 preserved but no longer messes up this directory. Now it's new room for
5283 the new set of documenation files.
5284 [Ralf S. Engelschall]
5286 *) SETs were incorrectly DER encoded. This was a major pain, because they
5287 shared code with SEQUENCEs, which aren't coded the same. This means that
5288 almost everything to do with SETs or SEQUENCEs has either changed name or
5289 number of arguments.
5290 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5292 *) Fix test data to work with the above.
5295 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5296 was already fixed by Eric for 0.9.1 it seems.
5297 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5299 *) Autodetect FreeBSD3.
5302 *) Fix various bugs in Configure. This affects the following platforms:
5306 unixware-2.0-pentium
5310 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5311 before they are needed.
5314 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5318 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5320 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5321 changed SSLeay to OpenSSL in version strings.
5322 [Ralf S. Engelschall]
5324 *) Some fixups to the top-level documents.
5327 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5328 because the symlink to include/ was missing.
5329 [Ralf S. Engelschall]
5331 *) Incorporated the popular no-RSA/DSA-only patches
5332 which allow to compile a RSA-free SSLeay.
5333 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5335 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5336 when "ssleay" is still not found.
5337 [Ralf S. Engelschall]
5339 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5340 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5342 *) Updated the README file.
5343 [Ralf S. Engelschall]
5345 *) Added various .cvsignore files in the CVS repository subdirs
5346 to make a "cvs update" really silent.
5347 [Ralf S. Engelschall]
5349 *) Recompiled the error-definition header files and added
5350 missing symbols to the Win32 linker tables.
5351 [Ralf S. Engelschall]
5353 *) Cleaned up the top-level documents;
5354 o new files: CHANGES and LICENSE
5355 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5356 o merged COPYRIGHT into LICENSE
5357 o removed obsolete TODO file
5358 o renamed MICROSOFT to INSTALL.W32
5359 [Ralf S. Engelschall]
5361 *) Removed dummy files from the 0.9.1b source tree:
5362 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5363 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5364 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5365 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5366 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5367 [Ralf S. Engelschall]
5369 *) Added various platform portability fixes.
5372 *) The Genesis of the OpenSSL rpject:
5373 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5374 Young and Tim J. Hudson created while they were working for C2Net until
5376 [The OpenSSL Project]
5379 Changes between 0.9.0b and 0.9.1b [not released]
5381 *) Updated a few CA certificates under certs/
5384 *) Changed some BIGNUM api stuff.
5387 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5388 DGUX x86, Linux Alpha, etc.
5391 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5392 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5396 *) Add -strparse option to asn1pars program which parses nested
5398 [Dr Stephen Henson <shenson@bigfoot.com>]
5400 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5403 *) DSA fix for "ca" program.
5406 *) Added "-genkey" option to "dsaparam" program.
5409 *) Added RIPE MD160 (rmd160) message digest.
5412 *) Added -a (all) option to "ssleay version" command.
5415 *) Added PLATFORM define which is the id given to Configure.
5418 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5421 *) Extended the ASN.1 parser routines.
5424 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5427 *) Added a BN_CTX to the BN library.
5430 *) Fixed the weak key values in DES library
5433 *) Changed API in EVP library for cipher aliases.
5436 *) Added support for RC2/64bit cipher.
5439 *) Converted the lhash library to the crypto/mem.c functions.
5442 *) Added more recognized ASN.1 object ids.
5445 *) Added more RSA padding checks for SSL/TLS.
5448 *) Added BIO proxy/filter functionality.
5451 *) Added extra_certs to SSL_CTX which can be used
5452 send extra CA certificates to the client in the CA cert chain sending
5453 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5456 *) Now Fortezza is denied in the authentication phase because
5457 this is key exchange mechanism is not supported by SSLeay at all.
5460 *) Additional PKCS1 checks.
5463 *) Support the string "TLSv1" for all TLS v1 ciphers.
5466 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5467 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5470 *) Fixed a few memory leaks.
5473 *) Fixed various code and comment typos.
5476 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5477 bytes sent in the client random.
5478 [Edward Bishop <ebishop@spyglass.com>]