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 +) New functions/macros
17 SSL_CTX_set_msg_callback(ctx, cb)
18 SSL_CTX_set_msg_callback_arg(ctx, arg)
19 SSL_set_msg_callback(ssl, cb)
20 SSL_set_msg_callback_arg(ssl, arg)
22 to request calling a callback function
24 void cb(int write_p, int version, int content_type,
25 const void *buf, size_t len, SSL *ssl, void *arg)
27 whenever a protocol message has been completely received
28 (write_p == 0) or sent (write_p == 1). Here 'version' is the
29 protocol version according to which the SSL library interprets
30 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
31 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
32 the content type as defined in the SSL 3.0/TLS 1.0 protocol
33 specification (change_cipher_spec(20), alert(21), handshake(22)).
34 'buf' and 'len' point to the actual message, 'ssl' to the
35 SSL object, and 'arg' is the application-defined value set by
36 SSL[_CTX]_set_msg_callback_arg().
38 'openssl s_client' and 'openssl s_server' have new '-msg' options
39 to enable a callback that displays all protocol messages.
42 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
43 messages are stored in a single piece (fixed-length part and
44 variable-length part combined) and fix various bugs found on the way.
47 +) Change the shared library support so shared libraries are built as
48 soon as the corresponding static library is finished, and thereby get
49 openssl and the test programs linked against the shared library.
50 This still only happens when the keyword "shard" has been given to
51 the configuration scripts.
53 NOTE: shared library support is still an experimental thing, and
54 backward binary compatibility is still not guaranteed.
55 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
57 +) Add support for Subject Information Access extension.
58 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
60 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
61 additional bytes when new memory had to be allocated, not just
62 when reusing an existing buffer.
65 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
66 instead. BIO_gethostbyname() does not know what timeouts are
67 appropriate, so entries would stay in cache even when they have
69 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
71 +) New command line and configuration option 'utf8' for the req command.
72 This allows field values to be specified as UTF8 strings.
75 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
76 runs for the former and machine-readable output for the latter.
79 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
80 of the e-mail address in the DN (i.e., it will go into a certificate
81 extension only). The new configuration file option 'email_in_dn = no'
83 [Massimiliano Pala madwolf@openca.org]
85 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
86 faced with a pathologically small ClientHello fragment that does
87 not contain client_version: Instead of aborting with an error,
88 simply choose the highest available protocol version (i.e.,
89 TLS 1.0 unless it is disabled). In practice, ClientHello
90 messages are never sent like this, but this change gives us
91 strictly correct behaviour at least for TLS.
94 +) Change all functions with names starting with des_ to be starting
95 with DES_ instead. This because there are increasing clashes with
96 libdes and other des libraries that are currently used by other
97 projects. The old libdes interface is provided, as well as crypt(),
98 if openssl/des_old.h is included. Note that crypt() is no longer
99 declared in openssl/des.h.
101 NOTE: This is a major break of an old API into a new one. Software
102 authors are encouraged to switch to the DES_ style functions. Some
103 time in the future, des_old.h and the libdes compatibility functions
104 will be completely removed.
107 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
108 never resets s->method to s->ctx->method when called from within
109 one of the SSL handshake functions.
110 [Bodo Moeller; problem pointed out by Niko Baric]
112 +) Test for certificates which contain unsupported critical extensions.
113 If such a certificate is found during a verify operation it is
114 rejected by default: this behaviour can be overridden by either
115 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
116 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
117 X509_supported_extension() has also been added which returns 1 if a
118 particular extension is supported.
121 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
122 (sent using the client's version number) if client_version is
123 smaller than the protocol version in use. Also change
124 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
125 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
126 the client will at least see that alert.
129 +) Modify the behaviour of EVP cipher functions in similar way to digests
130 to retain compatibility with existing code.
133 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
134 compatibility with existing code. In particular the 'ctx' parameter does
135 not have to be to be initialized before the call to EVP_DigestInit() and
136 it is tidied up after a call to EVP_DigestFinal(). New function
137 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
138 EVP_MD_CTX_copy() changed to not require the destination to be
139 initialized valid and new function EVP_MD_CTX_copy_ex() added which
140 requires the destination to be valid.
142 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
143 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
146 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
147 so that complete 'Handshake' protocol structures are kept in memory
148 instead of overwriting 'msg_type' and 'length' with 'body' data.
151 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
155 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
156 [Massimo Santin via Richard Levitte]
158 +) Major restructuring to the underlying ENGINE code. This includes
159 reduction of linker bloat, separation of pure "ENGINE" manipulation
160 (initialisation, etc) from functionality dealing with implementations
161 of specific crypto iterfaces. This change also introduces integrated
162 support for symmetric ciphers and digest implementations - so ENGINEs
163 can now accelerate these by providing EVP_CIPHER and EVP_MD
164 implementations of their own. This is detailed in crypto/engine/README
165 as it couldn't be adequately described here. However, there are a few
166 API changes worth noting - some RSA, DSA, DH, and RAND functions that
167 were changed in the original introduction of ENGINE code have now
168 reverted back - the hooking from this code to ENGINE is now a good
169 deal more passive and at run-time, operations deal directly with
170 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
171 dereferencing through an ENGINE pointer any more. Also, the ENGINE
172 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
173 they were not being used by the framework as there is no concept of a
174 BIGNUM_METHOD and they could not be generalised to the new
175 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
176 ENGINE_cpy() has been removed as it cannot be consistently defined in
180 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
183 +) Change mkdef.pl to sort symbols that get the same entry number,
184 and make sure the automatically generated functions ERR_load_*
185 become part of libeay.num as well.
188 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
189 client receives HelloRequest while in a handshake.
190 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
192 +) New function SSL_renegotiate_pending(). This returns true once
193 renegotiation has been requested (either SSL_renegotiate() call
194 or HelloRequest/ClientHello receveived from the peer) and becomes
195 false once a handshake has been completed.
196 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
197 sends a HelloRequest, but does not ensure that a handshake takes
198 place. SSL_renegotiate_pending() is useful for checking if the
199 client has followed the request.)
202 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
203 By default, clients may request session resumption even during
204 renegotiation (if session ID contexts permit); with this option,
205 session resumption is possible only in the first handshake.
208 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
209 should end in 'break', not 'goto end' which circuments various
210 cleanups done in state SSL_ST_OK. But session related stuff
211 must be disabled for SSL_ST_OK in the case that we just sent a
214 Also avoid some overhead by not calling ssl_init_wbio_buffer()
215 before just sending a HelloRequest.
216 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
218 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
219 reveal whether illegal block cipher padding was found or a MAC
220 verification error occured. (Neither SSLerr() codes nor alerts
221 are directly visible to potential attackers, but the information
222 may leak via logfiles.)
224 Similar changes are not required for the SSL 2.0 implementation
225 because the number of padding bytes is sent in clear for SSL 2.0,
226 and the extra bytes are just ignored. However ssl/s2_pkt.c
227 failed to verify that the purported number of padding bytes is in
231 +) Add some demos for certificate and certificate request creation.
234 +) Make maximum certificate chain size accepted from the peer application
235 settable (SSL*_get/set_max_cert_list()), as proposed by
236 "Douglas E. Engert" <deengert@anl.gov>.
239 +) Add support for shared libraries for Unixware-7 and support including
240 shared libraries for OpenUNIX-8 (Boyd Lynn Gerber <gerberb@zenez.com>).
243 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
244 'wristwatch attack' using huge encoding parameters (cf.
245 James H. Manger's CRYPTO 2001 paper). Note that the
246 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
247 encoding parameters and hence was not vulnerable.
250 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
251 be done prior to destruction. Use this to unload error strings from
252 ENGINEs that load their own error strings. NB: This adds two new API
253 functions to "get" and "set" this destroy handler in an ENGINE.
256 +) Alter all existing ENGINE implementations (except "openssl" and
257 "openbsd") to dynamically instantiate their own error strings. This
258 makes them more flexible to be built both as statically-linked ENGINEs
259 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
260 Also, add stub code to each that makes building them as self-contained
261 shared-libraries easier (see README.ENGINE).
264 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
265 implementations into applications that are completely implemented in
266 self-contained shared-libraries. The "dynamic" ENGINE exposes control
267 commands that can be used to configure what shared-library to load and
268 to control aspects of the way it is handled. Also, made an update to
269 the README.ENGINE file that brings its information up-to-date and
270 provides some information and instructions on the "dynamic" ENGINE
271 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
275 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
277 *) Make it possible to unload ranges of ERR strings with a new
278 "ERR_unload_strings" function.
281 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
282 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
283 followed by modular reduction.
284 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
286 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
287 equivalent based on BN_pseudo_rand() instead of BN_rand().
290 +) Add a copy() function to EVP_MD.
293 +) Make EVP_MD routines take a context pointer instead of just the
294 md_data void pointer.
297 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
298 that the digest can only process a single chunk of data
299 (typically because it is provided by a piece of
300 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
301 is only going to provide a single chunk of data, and hence the
302 framework needn't accumulate the data for oneshot drivers.
305 +) As with "ERR", make it possible to replace the underlying "ex_data"
306 functions. This change also alters the storage and management of global
307 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
308 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
309 index counters. The API functions that use this state have been changed
310 to take a "class_index" rather than pointers to the class's local STACK
311 and counter, and there is now an API function to dynamically create new
312 classes. This centralisation allows us to (a) plug a lot of the
313 thread-safety problems that existed, and (b) makes it possible to clean
314 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
315 such data would previously have always leaked in application code and
316 workarounds were in place to make the memory debugging turn a blind eye
317 to it. Application code that doesn't use this new function will still
318 leak as before, but their memory debugging output will announce it now
319 rather than letting it slide.
321 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
322 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
323 has a return value to indicate success or failure.
326 +) Make it possible to replace the underlying "ERR" functions such that the
327 global state (2 LHASH tables and 2 locks) is only used by the "default"
328 implementation. This change also adds two functions to "get" and "set"
329 the implementation prior to it being automatically set the first time
330 any other ERR function takes place. Ie. an application can call "get",
331 pass the return value to a module it has just loaded, and that module
332 can call its own "set" function using that value. This means the
333 module's "ERR" operations will use (and modify) the error state in the
334 application and not in its own statically linked copy of OpenSSL code.
337 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
338 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
339 the operation, and provides a more encapsulated way for external code
340 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
341 to use these functions rather than manually incrementing the counts.
343 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
346 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
347 This function was broken, as the check for a new client hello message
348 to handle SGC did not allow these large messages.
349 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
352 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
355 +) Add EVP test program.
358 +) Add symmetric cipher support to ENGINE. Expect the API to change!
361 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
362 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
363 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
364 These allow a CRL to be built without having to access X509_CRL fields
365 directly. Modify 'ca' application to use new functions.
368 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
369 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
372 *) Rework the configuration and shared library support for Tru64 Unix.
373 The configuration part makes use of modern compiler features and
374 still retains old compiler behavior for those that run older versions
375 of the OS. The shared library support part includes a variant that
376 uses the RPATH feature, and is available through the special
377 configuration target "alpha-cc-rpath", which will never be selected
379 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
381 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
382 with the same message size as in ssl3_get_certificate_request().
383 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
384 messages might inadvertently be reject as too long.
385 [Petr Lampa <lampa@fee.vutbr.cz>]
387 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
388 bug workarounds. Rollback attack detection is a security feature.
389 The problem will only arise on OpenSSL servers when TLSv1 is not
390 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
391 Software authors not wanting to support TLSv1 will have special reasons
392 for their choice and can explicitly enable this option.
393 [Bodo Moeller, Lutz Jaenicke]
395 +) Rationalise EVP so it can be extended: don't include a union of
396 cipher/digest structures, add init/cleanup functions. This also reduces
397 the number of header dependencies.
400 +) Make DES key schedule conform to the usual scheme, as well as
401 correcting its structure. This means that calls to DES functions
402 now have to pass a pointer to a des_key_schedule instead of a
403 plain des_key_schedule (which was actually always a pointer
407 +) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
410 *) Modified SSL library such that the verify_callback that has been set
411 specificly for an SSL object with SSL_set_verify() is actually being
412 used. Before the change, a verify_callback set with this function was
413 ignored and the verify_callback() set in the SSL_CTX at the time of
414 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
415 to allow the necessary settings.
418 +) Initial reduction of linker bloat: the use of some functions, such as
419 PEM causes large amounts of unused functions to be linked in due to
420 poor organisation. For example pem_all.c contains every PEM function
421 which has a knock on effect of linking in large amounts of (unused)
422 ASN1 code. Grouping together similar functions and splitting unrelated
423 functions prevents this.
426 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
427 explicitly to NULL, as at least on Solaris 8 this seems not always to be
428 done automatically (in contradiction to the requirements of the C
429 standard). This made problems when used from OpenSSH.
432 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
433 dh->length and always used
435 BN_rand_range(priv_key, dh->p).
437 BN_rand_range() is not necessary for Diffie-Hellman, and this
438 specific range makes Diffie-Hellman unnecessarily inefficient if
439 dh->length (recommended exponent length) is much smaller than the
440 length of dh->p. We could use BN_rand_range() if the order of
441 the subgroup was stored in the DH structure, but we only have
446 BN_rand(priv_key, l, ...)
448 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
454 RSA_eay_public_encrypt
455 RSA_eay_private_decrypt
456 RSA_eay_private_encrypt (signing)
457 RSA_eay_public_decrypt (signature verification)
459 (default implementations for RSA_public_encrypt,
460 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
461 always reject numbers >= n.
464 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
465 to synchronize access to 'locking_thread'. This is necessary on
466 systems where access to 'locking_thread' (an 'unsigned long'
467 variable) is not atomic.
470 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
471 *before* setting the 'crypto_lock_rand' flag. The previous code had
472 a race condition if 0 is a valid thread ID.
473 [Travis Vitek <vitek@roguewave.com>]
475 +) Cleanup of EVP macros.
478 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
482 +) Add initial OCSP responder support to ocsp application. The
483 revocation information is handled using the text based index
484 use by the ca application. The responder can either handle
485 requests generated internally, supplied in files (for example
486 via a CGI script) or using an internal minimal server.
489 +) Add configuration choices to get zlib compression for TLS.
492 +) Changes to Kerberos SSL for RFC 2712 compliance:
493 1. Implemented real KerberosWrapper, instead of just using
494 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
495 2. Implemented optional authenticator field of KerberosWrapper.
497 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
498 and authenticator structs; see crypto/krb5/.
500 Generalized Kerberos calls to support multiple Kerberos libraries.
501 [Vern Staats <staatsvr@asc.hpc.mil>,
502 Jeffrey Altman <jaltman@columbia.edu>
505 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
506 already does with RSA. testdsa.h now has 'priv_key/pub_key'
507 values for each of the key sizes rather than having just
508 parameters (and 'speed' generating keys each time).
511 -) OpenSSL 0.9.6b released [9 July 2001]
513 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
514 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
515 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
516 PRNG state recovery was possible based on the output of
517 one PRNG request appropriately sized to gain knowledge on
518 'md' followed by enough consecutive 1-byte PRNG requests
519 to traverse all of 'state'.
521 1. When updating 'md_local' (the current thread's copy of 'md')
522 during PRNG output generation, hash all of the previous
523 'md_local' value, not just the half used for PRNG output.
525 2. Make the number of bytes from 'state' included into the hash
526 independent from the number of PRNG bytes requested.
528 The first measure alone would be sufficient to avoid
529 Markku-Juhani's attack. (Actually it had never occurred
530 to me that the half of 'md_local' used for chaining was the
531 half from which PRNG output bytes were taken -- I had always
532 assumed that the secret half would be used.) The second
533 measure makes sure that additional data from 'state' is never
534 mixed into 'md_local' in small portions; this heuristically
535 further strengthens the PRNG.
538 +) Speed up EVP routines.
541 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
542 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
543 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
544 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
546 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
547 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
548 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
551 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
553 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
556 *) Fix crypto/bn/asm/mips3.s.
559 *) When only the key is given to "enc", the IV is undefined. Print out
560 an error message in this case.
563 +) Added the OS2-EMX target.
564 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
566 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
567 to support NCONF routines in extension code. New function CONF_set_nconf()
568 to allow functions which take an NCONF to also handle the old LHASH
569 structure: this means that the old CONF compatible routines can be
570 retained (in particular wrt extensions) without having to duplicate the
571 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
574 *) Handle special case when X509_NAME is empty in X509 printing routines.
577 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
578 positive and less than q.
581 +) Enhance the general user interface with mechanisms for inner control
582 and with possibilities to have yes/no kind of prompts.
585 +) Change all calls to low level digest routines in the library and
586 applications to use EVP. Add missing calls to HMAC_cleanup() and
587 don't assume HMAC_CTX can be copied using memcpy().
588 [Verdon Walker <VWalker@novell.com>, Steve Henson]
590 +) Add the possibility to control engines through control names but with
591 arbitrary arguments instead of just a string.
592 Change the key loaders to take a UI_METHOD instead of a callback
593 function pointer. NOTE: this breaks binary compatibility with earlier
594 versions of OpenSSL [engine].
595 Adapt the nCipher code for these new conditions and add a card insertion
599 +) Enhance the general user interface with mechanisms to better support
600 dialog box interfaces, application-defined prompts, the possibility
601 to use defaults (for example default passwords from somewhere else)
602 and interrupts/cancellations.
605 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
606 used: it isn't thread safe and the add_lock_callback should handle
608 [Paul Rose <Paul.Rose@bridge.com>]
610 *) Verify that incoming data obeys the block size in
611 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
614 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
615 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
619 [Ulf Möller, Bodo Möller]
621 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
622 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
623 when fixing the server behaviour for backwards-compatible 'client
624 hello' messages. (Note that the attack is impractical against
625 SSL 3.0 and TLS 1.0 anyway because length and version checking
626 means that the probability of guessing a valid ciphertext is
627 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
630 Before 0.9.5, the countermeasure (hide the error by generating a
631 random 'decryption result') did not work properly because
632 ERR_clear_error() was missing, meaning that SSL_get_error() would
633 detect the supposedly ignored error.
635 Both problems are now fixed.
638 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
639 (previously it was 1024).
642 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
643 tidy up some unnecessarily weird code in 'sk_new()').
644 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
646 +) Change the key loading routines for ENGINEs to use the same kind
647 callback (pem_password_cb) as all other routines that need this
651 *) Fix for compatibility mode trust settings: ignore trust settings
652 unless some valid trust or reject settings are present.
655 *) Fix for blowfish EVP: its a variable length cipher.
658 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
659 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
660 than this minimum value is recommended.
663 +) New random seeder for OpenVMS, using the system process statistics
664 that are easily reachable.
667 +) Windows apparently can't transparently handle global
668 variables defined in DLLs. Initialisations such as:
670 const ASN1_ITEM *it = &ASN1_INTEGER_it;
672 wont compile. This is used by the any applications that need to
673 declare their own ASN1 modules. This was fixed by adding the option
674 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
675 needed for static libraries under Win32.
678 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
679 setting of purpose and trust fields. New X509_STORE trust and
680 purpose functions and tidy up setting in other SSL functions.
683 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
684 structure. These are inherited by X509_STORE_CTX when it is
685 initialised. This allows various defaults to be set in the
686 X509_STORE structure (such as flags for CRL checking and custom
687 purpose or trust settings) for functions which only use X509_STORE_CTX
688 internally such as S/MIME.
690 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
691 trust settings if they are not set in X509_STORE. This allows X509_STORE
692 purposes and trust (in S/MIME for example) to override any set by default.
694 Add command line options for CRL checking to smime, s_client and s_server
698 +) Initial CRL based revocation checking. If the CRL checking flag(s)
699 are set then the CRL is looked up in the X509_STORE structure and
700 its validity and signature checked, then if the certificate is found
701 in the CRL the verify fails with a revoked error.
703 Various new CRL related callbacks added to X509_STORE_CTX structure.
705 Command line options added to 'verify' application to support this.
707 This needs some additional work, such as being able to handle multiple
708 CRLs with different times, extension based lookup (rather than just
709 by subject name) and ultimately more complete V2 CRL extension
713 +) Add a general user interface API (crypto/ui/). This is designed
714 to replace things like des_read_password and friends (backward
715 compatibility functions using this new API are provided).
716 The purpose is to remove prompting functions from the DES code
717 section as well as provide for prompting through dialog boxes in
718 a window system and the like.
721 *) Fix various bugs related to DSA S/MIME verification. Handle missing
722 parameters in DSA public key structures and return an error in the
723 DSA routines if parameters are absent.
726 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
727 in the current directory if neither $RANDFILE nor $HOME was set.
728 RAND_file_name() in 0.9.6a returned NULL in this case. This has
729 caused some confusion to Windows users who haven't defined $HOME.
730 Thus RAND_file_name() is changed again: e_os.h can define a
731 DEFAULT_HOME, which will be used if $HOME is not set.
732 For Windows, we use "C:"; on other platforms, we still require
733 environment variables.
735 +) Add "ex_data" support to ENGINE so implementations can add state at a
736 per-structure level rather than having to store it globally.
739 +) Make it possible for ENGINE structures to be copied when retrieved by
740 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
741 This causes the "original" ENGINE structure to act like a template,
742 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
743 operational state can be localised to each ENGINE structure, despite the
744 fact they all share the same "methods". New ENGINE structures returned in
745 this case have no functional references and the return value is the single
746 structural reference. This matches the single structural reference returned
747 by ENGINE_by_id() normally, when it is incremented on the pre-existing
751 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
752 needs to match any other type at all we need to manually clear the
756 +) Changes to the "openssl engine" utility to include;
757 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
758 about an ENGINE's available control commands.
759 - executing control commands from command line arguments using the
760 '-pre' and '-post' switches. '-post' is only used if '-t' is
761 specified and the ENGINE is successfully initialised. The syntax for
762 the individual commands are colon-separated, for example;
763 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
766 +) New dynamic control command support for ENGINEs. ENGINEs can now
767 declare their own commands (numbers), names (strings), descriptions,
768 and input types for run-time discovery by calling applications. A
769 subset of these commands are implicitly classed as "executable"
770 depending on their input type, and only these can be invoked through
771 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
772 can be based on user input, config files, etc). The distinction is
773 that "executable" commands cannot return anything other than a boolean
774 result and can only support numeric or string input, whereas some
775 discoverable commands may only be for direct use through
776 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
777 pointers, or other custom uses. The "executable" commands are to
778 support parameterisations of ENGINE behaviour that can be
779 unambiguously defined by ENGINEs and used consistently across any
780 OpenSSL-based application. Commands have been added to all the
781 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
782 control over shared-library paths without source code alterations.
785 +) Changed all ENGINE implementations to dynamically allocate their
786 ENGINEs rather than declaring them statically. Apart from this being
787 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
788 this also allows the implementations to compile without using the
789 internal engine_int.h header.
792 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
793 'const' value. Any code that should be able to modify a RAND_METHOD
794 should already have non-const pointers to it (ie. they should only
795 modify their own ones).
798 +) Made a variety of little tweaks to the ENGINE code.
799 - "atalla" and "ubsec" string definitions were moved from header files
800 to C code. "nuron" string definitions were placed in variables
801 rather than hard-coded - allowing parameterisation of these values
802 later on via ctrl() commands.
803 - Removed unused "#if 0"'d code.
804 - Fixed engine list iteration code so it uses ENGINE_free() to release
805 structural references.
806 - Constified the RAND_METHOD element of ENGINE structures.
807 - Constified various get/set functions as appropriate and added
808 missing functions (including a catch-all ENGINE_cpy that duplicates
809 all ENGINE values onto a new ENGINE except reference counts/state).
810 - Removed NULL parameter checks in get/set functions. Setting a method
811 or function to NULL is a way of cancelling out a previously set
812 value. Passing a NULL ENGINE parameter is just plain stupid anyway
813 and doesn't justify the extra error symbols and code.
814 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
815 flags from engine_int.h to engine.h.
816 - Changed prototypes for ENGINE handler functions (init(), finish(),
817 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
820 *) Move 'if (!initialized) RAND_poll()' into regions protected by
821 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
822 having multiple threads call RAND_poll() concurrently.
825 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
826 combination of a flag and a thread ID variable.
827 Otherwise while one thread is in ssleay_rand_bytes (which sets the
828 flag), *other* threads can enter ssleay_add_bytes without obeying
829 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
830 that they do not hold after the first thread unsets add_do_not_lock).
833 +) Implement binary inversion algorithm for BN_mod_inverse in addition
834 to the algorithm using long division. The binary algorithm can be
835 used only if the modulus is odd. On 32-bit systems, it is faster
836 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
837 roughly 5-15% for 256-bit moduli), so we use it only for moduli
838 up to 450 bits. In 64-bit environments, the binary algorithm
839 appears to be advantageous for much longer moduli; here we use it
840 for moduli up to 2048 bits.
843 *) Change bctest again: '-x' expressions are not available in all
847 -) OpenSSL 0.9.6a released [5 Apr 2001]
849 *) Fix a couple of memory leaks in PKCS7_dataDecode()
850 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
852 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
853 the default extension for executables, if any. Also, make the perl
854 scripts that use symlink() to test if it really exists and use "cp"
855 if it doesn't. All this made OpenSSL compilable and installable in
859 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
860 could not support the combine flag in choice fields.
863 -) Fix for asn1_GetSequence() for indefinite length constructed data.
864 If SEQUENCE is length is indefinite just set c->slen to the total
865 amount of data available.
866 [Steve Henson, reported by shige@FreeBSD.org]
867 [This change does not apply to 0.9.7.]
869 *) Change bctest to avoid here-documents inside command substitution
870 (workaround for FreeBSD /bin/sh bug).
871 For compatibility with Ultrix, avoid shell functions (introduced
872 in the bctest version that searches along $PATH).
875 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
876 with des_encrypt() defined on some operating systems, like Solaris
880 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
881 On the Importance of Eliminating Errors in Cryptographic
882 Computations, J. Cryptology 14 (2001) 2, 101-119,
883 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
886 *) MIPS assembler BIGNUM division bug fix.
889 *) Disabled incorrect Alpha assembler code.
892 -) Fix PKCS#7 decode routines so they correctly update the length
893 after reading an EOC for the EXPLICIT tag.
895 [This change does not apply to 0.9.7.]
897 *) Fix bug in PKCS#12 key generation routines. This was triggered
898 if a 3DES key was generated with a 0 initial byte. Include
899 PKCS12_BROKEN_KEYGEN compilation option to retain the old
900 (but broken) behaviour.
903 *) Enhance bctest to search for a working bc along $PATH and print
905 [Tim Rice <tim@multitalents.net> via Richard Levitte]
907 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
908 extensions from a certificate request to the certificate.
911 +) Allow multiple 'certopt' and 'nameopt' options to be separated
912 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
913 file: this allows the display of the certificate about to be
914 signed to be customised, to allow certain fields to be included
915 or excluded and extension details. The old system didn't display
916 multicharacter strings properly, omitted fields not in the policy
917 and couldn't display additional details such as extensions.
920 *) Fix memory leaks in err.c: free err_data string if necessary;
921 don't write to the wrong index in ERR_set_error_data.
924 +) Function EC_POINTs_mul for simultaneous scalar multiplication
925 of an arbitrary number of elliptic curve points, optionally
926 including the generator defined for the EC_GROUP.
927 EC_POINT_mul is a simple wrapper function for the typical case
928 that the point list has just one item (besides the optional
932 +) First EC_METHODs for curves over GF(p):
934 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
935 operations and provides various method functions that can also
936 operate with faster implementations of modular arithmetic.
938 EC_GFp_mont_method() reuses most functions that are part of
939 EC_GFp_simple_method, but uses Montgomery arithmetic.
941 [Bodo Moeller; point addition and point doubling
942 implementation directly derived from source code provided by
943 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
945 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
948 Curves are EC_GROUP objects (with an optional group generator)
949 based on EC_METHODs that are built into the library.
951 Points are EC_POINT objects based on EC_GROUP objects.
953 Most of the framework would be able to handle curves over arbitrary
954 finite fields, but as there are no obvious types for fields other
955 than GF(p), some functions are limited to that for now.
958 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
959 that the file contains a complete HTTP response.
962 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
963 change the def and num file printf format specifier from "%-40sXXX"
964 to "%-39s XXX". The latter will always guarantee a space after the
965 field while the former will cause them to run together if the field
966 is 40 of more characters long.
969 +) Constify the cipher and digest 'method' functions and structures
970 and modify related functions to take constant EVP_MD and EVP_CIPHER
974 *) Implement ssl23_peek (analogous to ssl23_read), which previously
978 *) Replace rdtsc with _emit statements for VC++ version 5.
979 [Jeremy Cooper <jeremy@baymoo.org>]
981 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
982 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
985 +) Modify EVP_Digest*() routines so they now return values. Although the
986 internal software routines can never fail additional hardware versions
990 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
992 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
993 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
996 ERR_R_NESTED_ASN1_ERROR
998 ERR_R_MISSING_ASN1_EOS
999 were 4 .. 9, conflicting with
1000 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1002 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1003 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1005 Add new error code 'ERR_R_INTERNAL_ERROR'.
1008 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1012 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1013 sets the subject name for a new request or supersedes the
1014 subject name in a given request. Formats that can be parsed are
1015 'CN=Some Name, OU=myOU, C=IT'
1017 'CN=Some Name/OU=myOU/C=IT'.
1019 Add options '-batch' and '-verbose' to 'openssl req'.
1020 [Massimiliano Pala <madwolf@hackmasters.net>]
1022 *) Make it possible to reuse SSLv2 sessions.
1025 +) Introduce the possibility to access global variables through
1026 functions on platform were that's the best way to handle exporting
1027 global variables in shared libraries. To enable this functionality,
1028 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1029 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1030 is normally done by Configure or something similar).
1032 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1033 in the source file (foo.c) like this:
1035 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1036 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1038 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1039 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1041 OPENSSL_DECLARE_GLOBAL(int,foo);
1042 #define foo OPENSSL_GLOBAL_REF(foo)
1043 OPENSSL_DECLARE_GLOBAL(double,bar);
1044 #define bar OPENSSL_GLOBAL_REF(bar)
1046 The #defines are very important, and therefore so is including the
1047 header file everywhere where the defined globals are used.
1049 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1050 of ASN.1 items, but that structure is a bit different.
1052 The largest change is in util/mkdef.pl which has been enhanced with
1053 better and easier to understand logic to choose which symbols should
1054 go into the Windows .def files as well as a number of fixes and code
1055 cleanup (among others, algorithm keywords are now sorted
1056 lexicographically to avoid constant rewrites).
1059 *) In copy_email() check for >= 0 as a return value for
1060 X509_NAME_get_index_by_NID() since 0 is a valid index.
1061 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1063 +) In BN_div() keep a copy of the sign of 'num' before writing the
1064 result to 'rm' because if rm==num the value will be overwritten
1065 and produce the wrong result if 'num' is negative: this caused
1066 problems with BN_mod() and BN_nnmod().
1069 +) Function OCSP_request_verify(). This checks the signature on an
1070 OCSP request and verifies the signer certificate. The signer
1071 certificate is just checked for a generic purpose and OCSP request
1075 +) Add OCSP_check_validity() function to check the validity of OCSP
1076 responses. OCSP responses are prepared in real time and may only
1077 be a few seconds old. Simply checking that the current time lies
1078 between thisUpdate and nextUpdate max reject otherwise valid responses
1079 caused by either OCSP responder or client clock inaccuracy. Instead
1080 we allow thisUpdate and nextUpdate to fall within a certain period of
1081 the current time. The age of the response can also optionally be
1082 checked. Two new options -validity_period and -status_age added to
1086 +) If signature or public key algorithm is unrecognized print out its
1087 OID rather that just UNKNOWN.
1090 *) Avoid coredump with unsupported or invalid public keys by checking if
1091 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1092 PKCS7_verify() fails with non detached data.
1095 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1096 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1097 ID to be generated from the issuer certificate alone which can then be
1098 passed to OCSP_id_issuer_cmp().
1101 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1102 ASN1 modules to export functions returning ASN1_ITEM pointers
1103 instead of the ASN1_ITEM structures themselves. This adds several
1104 new macros which allow the underlying ASN1 function/structure to
1105 be accessed transparently. As a result code should not use ASN1_ITEM
1106 references directly (such as &X509_it) but instead use the relevant
1107 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1108 use of the new ASN1 code on platforms where exporting structures
1109 is problematical (for example in shared libraries) but exporting
1110 functions returning pointers to structures is not.
1113 +) Add support for overriding the generation of SSL/TLS session IDs.
1114 These callbacks can be registered either in an SSL_CTX or per SSL.
1115 The purpose of this is to allow applications to control, if they wish,
1116 the arbitrary values chosen for use as session IDs, particularly as it
1117 can be useful for session caching in multiple-server environments. A
1118 command-line switch for testing this (and any client code that wishes
1119 to use such a feature) has been added to "s_server".
1120 [Geoff Thorpe, Lutz Jaenicke]
1122 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1123 of the form '#if defined(...) || defined(...) || ...' and
1124 '#if !defined(...) && !defined(...) && ...'. This also avoids
1125 the growing number of special cases it was previously handling.
1128 *) Don't use getenv in library functions when run as setuid/setgid.
1129 New function OPENSSL_issetugid().
1132 +) Make all configuration macros available for application by making
1133 sure they are available in opensslconf.h, by giving them names starting
1134 with "OPENSSL_" to avoid conflicts with other packages and by making
1135 sure e_os2.h will cover all platform-specific cases together with
1137 Additionally, it is now possible to define configuration/platform-
1138 specific names (called "system identities"). In the C code, these
1139 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1140 macro with the name beginning with "OPENSSL_SYS_", which is determined
1141 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1145 +) New option -set_serial to 'req' and 'x509' this allows the serial
1146 number to use to be specified on the command line. Previously self
1147 signed certificates were hard coded with serial number 0 and the
1148 CA options of 'x509' had to use a serial number in a file which was
1152 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1153 due to incorrect handling of multi-threading:
1155 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1157 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1159 3. Count how many times MemCheck_off() has been called so that
1160 nested use can be treated correctly. This also avoids
1161 inband-signalling in the previous code (which relied on the
1162 assumption that thread ID 0 is impossible).
1165 +) New options to 'ca' utility to support V2 CRL entry extensions.
1166 Currently CRL reason, invalidity date and hold instruction are
1167 supported. Add new CRL extensions to V3 code and some new objects.
1170 *) Add "-rand" option also to s_client and s_server.
1173 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1174 disable standard block padding (aka PKCS#5 padding) in the EVP
1175 API, which was previously mandatory. This means that the data is
1176 not padded in any way and so the total length much be a multiple
1177 of the block size, otherwise an error occurs.
1180 +) Initial (incomplete) OCSP SSL support.
1183 *) Fix CPU detection on Irix 6.x.
1184 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1185 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1187 +) New function OCSP_parse_url(). This splits up a URL into its host,
1188 port and path components: primarily to parse OCSP URLs. New -url
1189 option to ocsp utility.
1192 +) New nonce behavior. The return value of OCSP_check_nonce() now
1193 reflects the various checks performed. Applications can decide
1194 whether to tolerate certain situations such as an absent nonce
1195 in a response when one was present in a request: the ocsp application
1196 just prints out a warning. New function OCSP_add1_basic_nonce()
1197 this is to allow responders to include a nonce in a response even if
1198 the request is nonce-less.
1201 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1204 [This change does not apply to 0.9.7.]
1206 *) Use the cached encoding of an X509_NAME structure rather than
1207 copying it. This is apparently the reason for the libsafe "errors"
1208 but the code is actually correct.
1211 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1212 skipped when using openssl x509 multiple times on a single input file,
1213 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1216 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1217 set string type: to handle setting ASN1_TIME structures. Fix ca
1218 utility to correctly initialize revocation date of CRLs.
1221 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1222 the clients preferred ciphersuites and rather use its own preferences.
1223 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1224 Internet Explorer by ensuring unchanged hash method during stepup.
1225 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1228 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1229 to aes and add a new 'exist' option to print out symbols that don't
1233 +) Additional options to ocsp utility to allow flags to be set and
1234 additional certificates supplied.
1237 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1238 OCSP client a number of certificate to only verify the response
1242 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1243 Bleichenbacher's DSA attack.
1244 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1245 to be set and top=0 forces the highest bit to be set; top=-1 is new
1246 and leaves the highest bit random.
1247 [Ulf Moeller, Bodo Moeller]
1249 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1250 handle the new API. Currently only ECB, CBC modes supported. Add new
1251 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1252 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1253 [Ben Laurie, Steve Henson]
1255 *) In the NCONF_...-based implementations for CONF_... queries
1256 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1257 a temporary CONF structure with the data component set to NULL
1258 (which gives segmentation faults in lh_retrieve).
1259 Instead, use NULL for the CONF pointer in CONF_get_string and
1260 CONF_get_number (which may use environment variables) and directly
1261 return NULL from CONF_get_section.
1264 *) Fix potential buffer overrun for EBCDIC.
1267 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1268 request to response.
1271 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1272 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1273 extract information from a certificate request. OCSP_response_create()
1274 creates a response and optionally adds a basic response structure.
1275 OCSP_basic_add1_status() adds a complete single response to a basic
1276 response and returns the OCSP_SINGLERESP structure just added (to allow
1277 extensions to be included for example). OCSP_basic_add1_cert() adds a
1278 certificate to a basic response and OCSP_basic_sign() signs a basic
1279 response with various flags. New helper functions ASN1_TIME_check()
1280 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1281 (converts ASN1_TIME to GeneralizedTime).
1284 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1285 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1286 structure from a certificate. X509_pubkey_digest() digests the public_key
1287 contents: this is used in various key identifiers.
1290 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1291 keyUsage if basicConstraints absent for a CA.
1294 *) Make SMIME_write_PKCS7() write mail header values with a format that
1295 is more generally accepted (no spaces before the semicolon), since
1296 some programs can't parse those values properly otherwise. Also make
1297 sure BIO's that break lines after each write do not create invalid
1301 +) Make sk_sort() tolerate a NULL argument.
1302 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1304 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1305 passed by the function are trusted implicitly. If any of them signed the
1306 response then it is assumed to be valid and is not verified.
1309 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1310 macros previously used would not encode an empty SEQUENCE OF
1311 and break the signature.
1313 [This change does not apply to 0.9.7.]
1315 *) Zero the premaster secret after deriving the master secret in
1319 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1320 to data. This was previously part of the PKCS7 ASN1 code. This
1321 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1322 [Steve Henson, reported by Kenneth R. Robinette
1323 <support@securenetterm.com>]
1325 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1326 routines: without these tracing memory leaks is very painful.
1327 Fix leaks in PKCS12 and PKCS7 routines.
1330 *) Add some EVP_add_digest_alias registrations (as found in
1331 OpenSSL_add_all_digests()) to SSL_library_init()
1332 aka OpenSSL_add_ssl_algorithms(). This provides improved
1333 compatibility with peers using X.509 certificates
1334 with unconventional AlgorithmIdentifier OIDs.
1337 *) Fix for Irix with NO_ASM.
1338 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1340 *) ./config script fixes.
1341 [Ulf Moeller, Richard Levitte]
1343 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1344 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1345 effectively meant GeneralizedTime would never be used. Now it
1346 is initialised to -1 but X509_time_adj() now has to check the value
1347 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1348 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1349 [Steve Henson, reported by Kenneth R. Robinette
1350 <support@securenetterm.com>]
1352 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1353 result in a zero length in the ASN1_INTEGER structure which was
1354 not consistent with the structure when d2i_ASN1_INTEGER() was used
1355 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1356 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1357 where it did not print out a minus for negative ASN1_INTEGER.
1360 *) Fix 'openssl passwd -1'.
1363 +) Add summary printout to ocsp utility. The various functions which
1364 convert status values to strings have been renamed to:
1365 OCSP_response_status_str(), OCSP_cert_status_str() and
1366 OCSP_crl_reason_str() and are no longer static. New options
1367 to verify nonce values and to disable verification. OCSP response
1368 printout format cleaned up.
1371 +) Add additional OCSP certificate checks. These are those specified
1372 in RFC2560. This consists of two separate checks: the CA of the
1373 certificate being checked must either be the OCSP signer certificate
1374 or the issuer of the OCSP signer certificate. In the latter case the
1375 OCSP signer certificate must contain the OCSP signing extended key
1376 usage. This check is performed by attempting to match the OCSP
1377 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1378 in the OCSP_CERTID structures of the response.
1381 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1382 and related routines. This uses the standard OpenSSL certificate
1383 verify routines to perform initial checks (just CA validity) and
1384 to obtain the certificate chain. Then additional checks will be
1385 performed on the chain. Currently the root CA is checked to see
1386 if it is explicitly trusted for OCSP signing. This is used to set
1387 a root CA as a global signing root: that is any certificate that
1388 chains to that CA is an acceptable OCSP signing certificate.
1391 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1392 extensions from a separate configuration file.
1393 As when reading extensions from the main configuration file,
1394 the '-extensions ...' option may be used for specifying the
1396 [Massimiliano Pala <madwolf@comune.modena.it>]
1398 *) Change PKCS12_key_gen_asc() so it can cope with non null
1399 terminated strings whose length is passed in the passlen
1400 parameter, for example from PEM callbacks. This was done
1401 by adding an extra length parameter to asc2uni().
1402 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1404 +) New OCSP utility. Allows OCSP requests to be generated or
1405 read. The request can be sent to a responder and the output
1406 parsed, outputed or printed in text form. Not complete yet:
1407 still needs to check the OCSP response validity.
1410 +) New subcommands for 'openssl ca':
1411 'openssl ca -status <serial>' prints the status of the cert with
1412 the given serial number (according to the index file).
1413 'openssl ca -updatedb' updates the expiry status of certificates
1415 [Massimiliano Pala <madwolf@comune.modena.it>]
1417 +) New '-newreq-nodes' command option to CA.pl. This is like
1418 '-newreq', but calls 'openssl req' with the '-nodes' option
1419 so that the resulting key is not encrypted.
1420 [Damien Miller <djm@mindrot.org>]
1422 +) New configuration for the GNU Hurd.
1423 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1425 +) Initial code to implement OCSP basic response verify. This
1426 is currently incomplete. Currently just finds the signer's
1427 certificate and verifies the signature on the response.
1430 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1431 value of OPENSSLDIR. This is available via the new '-d' option
1432 to 'openssl version', and is also included in 'openssl version -a'.
1435 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1436 call failed, free the DSA structure.
1439 +) Allowing defining memory allocation callbacks that will be given
1440 file name and line number information in additional arguments
1441 (a const char* and an int). The basic functionality remains, as
1442 well as the original possibility to just replace malloc(),
1443 realloc() and free() by functions that do not know about these
1444 additional arguments. To register and find out the current
1445 settings for extended allocation functions, the following
1446 functions are provided:
1448 CRYPTO_set_mem_ex_functions
1449 CRYPTO_set_locked_mem_ex_functions
1450 CRYPTO_get_mem_ex_functions
1451 CRYPTO_get_locked_mem_ex_functions
1453 These work the same way as CRYPTO_set_mem_functions and friends.
1454 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1455 extended allocation function is enabled.
1456 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1457 a conventional allocation function is enabled.
1458 [Richard Levitte, Bodo Moeller]
1460 *) Fix to uni2asc() to cope with zero length Unicode strings.
1461 These are present in some PKCS#12 files.
1464 +) Finish off removing the remaining LHASH function pointer casts.
1465 There should no longer be any prototype-casting required when using
1466 the LHASH abstraction, and any casts that remain are "bugs". See
1467 the callback types and macros at the head of lhash.h for details
1468 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1471 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1472 If /dev/[u]random devices are not available or do not return enough
1473 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1475 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1476 /etc/entropy will be queried once each in this sequence, quering stops
1477 when enough entropy was collected without querying more sockets.
1480 +) Change the Unix RAND_poll() variant to be able to poll several
1481 random devices, as specified by DEVRANDOM, until a sufficient amount
1482 of data has been collected. We spend at most 10 ms on each file
1483 (select timeout) and read in non-blocking mode. DEVRANDOM now
1484 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1485 (previously it was just the string "/dev/urandom"), so on typical
1486 platforms the 10 ms delay will never occur.
1487 Also separate out the Unix variant to its own file, rand_unix.c.
1488 For VMS, there's a currently-empty rand_vms.c.
1491 +) Move OCSP client related routines to ocsp_cl.c. These
1492 provide utility functions which an application needing
1493 to issue a request to an OCSP responder and analyse the
1494 response will typically need: as opposed to those which an
1495 OCSP responder itself would need which will be added later.
1497 OCSP_request_sign() signs an OCSP request with an API similar
1498 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1499 response. OCSP_response_get1_basic() extracts basic response
1500 from response. OCSP_resp_find_status(): finds and extracts status
1501 information from an OCSP_CERTID structure (which will be created
1502 when the request structure is built). These are built from lower
1503 level functions which work on OCSP_SINGLERESP structures but
1504 wont normally be used unless the application wishes to examine
1505 extensions in the OCSP response for example.
1507 Replace nonce routines with a pair of functions.
1508 OCSP_request_add1_nonce() adds a nonce value and optionally
1509 generates a random value. OCSP_check_nonce() checks the
1510 validity of the nonce in an OCSP response.
1513 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1514 This doesn't copy the supplied OCSP_CERTID and avoids the
1515 need to free up the newly created id. Change return type
1516 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1517 This can then be used to add extensions to the request.
1518 Deleted OCSP_request_new(), since most of its functionality
1519 is now in OCSP_REQUEST_new() (and the case insensitive name
1520 clash) apart from the ability to set the request name which
1521 will be added elsewhere.
1524 +) Update OCSP API. Remove obsolete extensions argument from
1525 various functions. Extensions are now handled using the new
1526 OCSP extension code. New simple OCSP HTTP function which
1527 can be used to send requests and parse the response.
1530 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1531 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1532 uses the special reorder version of SET OF to sort the attributes
1533 and reorder them to match the encoded order. This resolves a long
1534 standing problem: a verify on a PKCS7 structure just after signing
1535 it used to fail because the attribute order did not match the
1536 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1537 it uses the received order. This is necessary to tolerate some broken
1538 software that does not order SET OF. This is handled by encoding
1539 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1540 to produce the required SET OF.
1543 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1544 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1545 files to get correct declarations of the ASN.1 item variables.
1548 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1549 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1550 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1551 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1552 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1553 ASN1_ITEM and no wrapper functions.
1556 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1557 replace the old function pointer based I/O routines. Change most of
1558 the *_d2i_bio() and *_d2i_fp() functions to use these.
1561 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1562 lines, recognice more "algorithms" that can be deselected, and make
1563 it complain about algorithm deselection that isn't recognised.
1566 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1567 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1568 to use new functions. Add NO_ASN1_OLD which can be set to remove
1569 some old style ASN1 functions: this can be used to determine if old
1570 code will still work when these eventually go away.
1573 +) New extension functions for OCSP structures, these follow the
1574 same conventions as certificates and CRLs.
1577 +) New function X509V3_add1_i2d(). This automatically encodes and
1578 adds an extension. Its behaviour can be customised with various
1579 flags to append, replace or delete. Various wrappers added for
1580 certifcates and CRLs.
1583 +) Fix to avoid calling the underlying ASN1 print routine when
1584 an extension cannot be parsed. Correct a typo in the
1585 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1588 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1589 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1590 when writing a 32767 byte record.
1591 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1593 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1594 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1596 (RSA objects have a reference count access to which is protected
1597 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1598 so they are meant to be shared between threads.)
1599 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1600 "Reddie, Steven" <Steven.Reddie@ca.com>]
1602 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1603 entries for variables.
1606 *) Fix a deadlock in CRYPTO_mem_leaks().
1609 +) Add functionality to apps/openssl.c for detecting locking
1610 problems: As the program is single-threaded, all we have
1611 to do is register a locking callback using an array for
1612 storing which locks are currently held by the program.
1615 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1616 SSL_get_ex_data_X509_STORE_idx(), which is used in
1617 ssl_verify_cert_chain() and thus can be called at any time
1618 during TLS/SSL handshakes so that thread-safety is essential.
1619 Unfortunately, the ex_data design is not at all suited
1620 for multi-threaded use, so it probably should be abolished.
1623 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1624 [Broadcom, tweaked and integrated by Geoff Thorpe]
1626 +) Move common extension printing code to new function
1627 X509V3_print_extensions(). Reorganise OCSP print routines and
1628 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1631 +) New function X509_signature_print() to remove duplication in some
1635 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1636 set (this was treated exactly the same as SET OF previously). This
1637 is used to reorder the STACK representing the structure to match the
1638 encoding. This will be used to get round a problem where a PKCS7
1639 structure which was signed could not be verified because the STACK
1640 order did not reflect the encoded order.
1643 +) Reimplement the OCSP ASN1 module using the new code.
1646 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1647 for its ASN1 operations. The old style function pointers still exist
1648 for now but they will eventually go away.
1651 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1652 completely replaces the old ASN1 functionality with a table driven
1653 encoder and decoder which interprets an ASN1_ITEM structure describing
1654 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1655 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1656 has also been converted to the new form.
1659 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1660 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1661 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1662 for negative moduli.
1665 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1666 of not touching the result's sign bit.
1669 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1673 +) Changed the LHASH code to use prototypes for callbacks, and created
1674 macros to declare and implement thin (optionally static) functions
1675 that provide type-safety and avoid function pointer casting for the
1676 type-specific callbacks.
1679 *) Use better test patterns in bntest.
1682 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1684 [Veers Staats <staatsvr@asc.hpc.mil>,
1685 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1687 *) rand_win.c fix for Borland C.
1690 *) BN_rshift bugfix for n == 0.
1693 +) Reformat the FAQ so the different questions and answers can be divided
1694 in sections depending on the subject.
1697 +) Have the zlib compression code load ZLIB.DLL dynamically under
1701 +) New function BN_mod_sqrt for computing square roots modulo a prime
1702 (using the probabilistic Tonelli-Shanks algorithm unless
1703 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1704 be handled deterministically).
1705 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1707 *) Add a 'bctest' script that checks for some known 'bc' bugs
1708 so that 'make test' does not abort just because 'bc' is broken.
1711 *) Store verify_result within SSL_SESSION also for client side to
1712 avoid potential security hole. (Re-used sessions on the client side
1713 always resulted in verify_result==X509_V_OK, not using the original
1714 result of the server certificate verification.)
1717 +) Make BN_mod_inverse faster by explicitly handling small quotients
1718 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1719 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1722 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1723 SSL3_RT_APPLICATION_DATA, return 0.
1724 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1728 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1729 releases, have been re-implemented by renaming the previous
1730 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1731 and ssl3_read_internal, respectively, and adding 'peek' parameters
1732 to them. The new ssl[23]_{read,peek} functions are calls to
1733 ssl[23]_read_internal with the 'peek' flag set appropriately.
1734 A 'peek' parameter has also been added to ssl3_read_bytes, which
1735 does the actual work for ssl3_read_internal.
1738 +) New function BN_kronecker.
1741 +) Fix BN_gcd so that it works on negative inputs; the result is
1742 positive unless both parameters are zero.
1743 Previously something reasonably close to an infinite loop was
1744 possible because numbers could be growing instead of shrinking
1745 in the implementation of Euclid's algorithm.
1748 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1749 sign of the number in question.
1751 Fix BN_is_word(a,w) to work correctly for w == 0.
1753 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1754 because its test if the absolute value of 'a' equals 'w'.
1755 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1756 it exists mostly for use in the implementations of BN_is_zero(),
1757 BN_is_one(), and BN_is_word().
1760 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1761 the method-specific "init()" handler. Also clean up ex_data after
1762 calling the method-specific "finish()" handler. Previously, this was
1763 happening the other way round.
1766 +) New function BN_swap.
1769 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1770 the exponentiation functions are more likely to produce reasonable
1771 results on negative inputs.
1774 +) Change BN_mod_mul so that the result is always non-negative.
1775 Previously, it could be negative if one of the factors was negative;
1776 I don't think anyone really wanted that behaviour.
1779 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1780 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1781 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1782 and add new functions:
1791 BN_mod_lshift1_quick
1795 These functions always generate non-negative results.
1797 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1798 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1800 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1801 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1802 be reduced modulo m.
1803 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1805 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1806 was actually never needed) and in BN_mul(). The removal in BN_mul()
1807 required a small change in bn_mul_part_recursive() and the addition
1808 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1809 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1810 bn_sub_words() and bn_add_words() except they take arrays with
1814 +) In 'openssl passwd', verify passwords read from the terminal
1815 unless the '-salt' option is used (which usually means that
1816 verification would just waste user's time since the resulting
1817 hash is going to be compared with some given password hash)
1818 or the new '-noverify' option is used.
1820 This is an incompatible change, but it does not affect
1821 non-interactive use of 'openssl passwd' (passwords on the command
1822 line, '-stdin' option, '-in ...' option) and thus should not
1826 +) Remove all references to RSAref, since there's no more need for it.
1829 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
1830 The previous value, 12, was not always sufficient for BN_mod_exp().
1833 +) Make DSO load along a path given through an environment variable
1834 (SHLIB_PATH) with shl_load().
1837 +) Constify the ENGINE code as a result of BIGNUM constification.
1838 Also constify the RSA code and most things related to it. In a
1839 few places, most notable in the depth of the ASN.1 code, ugly
1840 casts back to non-const were required (to be solved at a later
1844 +) Make it so the openssl application has all engines loaded by default.
1847 +) Constify the BIGNUM routines a little more.
1850 *) Make sure that shared libraries get the internal name engine with
1851 the full version number and not just 0. This should mark the
1852 shared libraries as not backward compatible. Of course, this should
1853 be changed again when we can guarantee backward binary compatibility.
1856 +) Add the following functions:
1858 ENGINE_load_cswift()
1860 ENGINE_load_atalla()
1862 ENGINE_load_builtin_engines()
1864 That way, an application can itself choose if external engines that
1865 are built-in in OpenSSL shall ever be used or not. The benefit is
1866 that applications won't have to be linked with libdl or other dso
1867 libraries unless it's really needed.
1869 Changed 'openssl engine' to load all engines on demand.
1870 Changed the engine header files to avoid the duplication of some
1871 declarations (they differed!).
1874 +) 'openssl engine' can now list capabilities.
1877 +) Better error reporting in 'openssl engine'.
1880 +) Never call load_dh_param(NULL) in s_server.
1883 +) Add engine application. It can currently list engines by name and
1884 identity, and test if they are actually available.
1887 +) Add support for shared libraries under Irix.
1888 [Albert Chin-A-Young <china@thewrittenword.com>]
1890 +) Improve RPM specification file by forcing symbolic linking and making
1891 sure the installed documentation is also owned by root.root.
1892 [Damien Miller <djm@mindrot.org>]
1894 +) Add configuration option to build on Linux on both big-endian and
1896 [Ralf Baechle <ralf@uni-koblenz.de>]
1898 +) Give the OpenSSL applications more possibilities to make use of
1899 keys (public as well as private) handled by engines.
1902 +) Add OCSP code that comes from CertCo.
1905 +) Add VMS support for the Rijndael code.
1908 +) Added untested support for Nuron crypto accelerator.
1911 +) Add support for external cryptographic devices. This code was
1912 previously distributed separately as the "engine" branch.
1913 [Geoff Thorpe, Richard Levitte]
1915 +) Rework the filename-translation in the DSO code. It is now possible to
1916 have far greater control over how a "name" is turned into a filename
1917 depending on the operating environment and any oddities about the
1918 different shared library filenames on each system.
1921 +) Support threads on FreeBSD-elf in Configure.
1924 +) Add the possibility to create shared libraries on HP-UX
1927 +) Fix for SHA1 assembly problem with MASM: it produces
1928 warnings about corrupt line number information when assembling
1929 with debugging information. This is caused by the overlapping
1931 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1934 NCONF_get_number() has no error checking at all. As a replacement,
1935 NCONF_get_number_e() is defined (_e for "error checking") and is
1936 promoted strongly. The old NCONF_get_number is kept around for
1937 binary backward compatibility.
1938 Make it possible for methods to load from something other than a BIO,
1939 by providing a function pointer that is given a name instead of a BIO.
1940 For example, this could be used to load configuration data from an
1944 *) Fix typo in get_cert_by_subject() in by_dir.c
1945 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
1947 *) Rework the system to generate shared libraries:
1949 - Make note of the expected extension for the shared libraries and
1950 if there is a need for symbolic links from for example libcrypto.so.0
1951 to libcrypto.so.0.9.7. There is extended info in Configure for
1954 - Make as few rebuilds of the shared libraries as possible.
1956 - Still avoid linking the OpenSSL programs with the shared libraries.
1958 - When installing, install the shared libraries separately from the
1962 +) Fix for non blocking accept BIOs. Added new I/O special reason
1963 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1964 with non blocking I/O was not possible because no retry code was
1965 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1969 +) Added the beginnings of Rijndael support.
1972 +) Fix for bug in DirectoryString mask setting. Add support for
1973 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1974 to allow certificate printing to more controllable, additional
1975 'certopt' option to 'x509' to allow new printing options to be
1979 +) Clean old EAY MD5 hack from e_os.h.
1982 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
1984 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
1985 and not in SSL_clear because the latter is also used by the
1986 accept/connect functions; previously, the settings made by
1987 SSL_set_read_ahead would be lost during the handshake.
1988 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
1990 *) Correct util/mkdef.pl to be selective about disabled algorithms.
1991 Previously, it would create entries for disableed algorithms no
1995 *) Added several new manual pages for SSL_* function.
1998 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2000 *) In ssl23_get_client_hello, generate an error message when faced
2001 with an initial SSL 3.0/TLS record that is too small to contain the
2002 first two bytes of the ClientHello message, i.e. client_version.
2003 (Note that this is a pathologic case that probably has never happened
2004 in real life.) The previous approach was to use the version number
2005 from the record header as a substitute; but our protocol choice
2006 should not depend on that one because it is not authenticated
2007 by the Finished messages.
2010 *) More robust randomness gathering functions for Windows.
2011 [Jeffrey Altman <jaltman@columbia.edu>]
2013 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2014 not set then we don't setup the error code for issuer check errors
2015 to avoid possibly overwriting other errors which the callback does
2016 handle. If an application does set the flag then we assume it knows
2017 what it is doing and can handle the new informational codes
2021 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2022 a general "ANY" type, as such it should be able to decode anything
2023 including tagged types. However it didn't check the class so it would
2024 wrongly interpret tagged types in the same way as their universal
2025 counterpart and unknown types were just rejected. Changed so that the
2026 tagged and unknown types are handled in the same way as a SEQUENCE:
2027 that is the encoding is stored intact. There is also a new type
2028 "V_ASN1_OTHER" which is used when the class is not universal, in this
2029 case we have no idea what the actual type is so we just lump them all
2033 *) On VMS, stdout may very well lead to a file that is written to
2034 in a record-oriented fashion. That means that every write() will
2035 write a separate record, which will be read separately by the
2036 programs trying to read from it. This can be very confusing.
2038 The solution is to put a BIO filter in the way that will buffer
2039 text until a linefeed is reached, and then write everything a
2040 line at a time, so every record written will be an actual line,
2041 not chunks of lines and not (usually doesn't happen, but I've
2042 seen it once) several lines in one record. BIO_f_linebuffer() is
2045 Currently, it's a VMS-only method, because that's where it has
2046 been tested well enough.
2049 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2050 it can return incorrect results.
2051 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2052 but it was in 0.9.6-beta[12].)
2055 *) Disable the check for content being present when verifying detached
2056 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2057 include zero length content when signing messages.
2060 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2061 BIO_ctrl (for BIO pairs).
2064 *) Add DSO method for VMS.
2067 *) Bug fix: Montgomery multiplication could produce results with the
2071 *) Add RPM specification openssl.spec and modify it to build three
2072 packages. The default package contains applications, application
2073 documentation and run-time libraries. The devel package contains
2074 include files, static libraries and function documentation. The
2075 doc package contains the contents of the doc directory. The original
2076 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2079 *) Add a large number of documentation files for many SSL routines.
2080 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2082 *) Add a configuration entry for Sony News 4.
2083 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2085 *) Don't set the two most significant bits to one when generating a
2086 random number < q in the DSA library.
2089 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2090 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2091 the underlying transport is blocking) if a handshake took place.
2092 (The default behaviour is needed by applications such as s_client
2093 and s_server that use select() to determine when to use SSL_read;
2094 but for applications that know in advance when to expect data, it
2095 just makes things more complicated.)
2098 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2102 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2103 work better on such systems.
2104 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2106 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2107 Update PKCS12_parse() so it copies the friendlyName and the
2108 keyid to the certificates aux info.
2111 *) Fix bug in PKCS7_verify() which caused an infinite loop
2112 if there was more than one signature.
2113 [Sven Uszpelkat <su@celocom.de>]
2115 *) Major change in util/mkdef.pl to include extra information
2116 about each symbol, as well as presentig variables as well
2117 as functions. This change means that there's n more need
2118 to rebuild the .num files when some algorithms are excluded.
2121 *) Allow the verify time to be set by an application,
2122 rather than always using the current time.
2125 *) Phase 2 verify code reorganisation. The certificate
2126 verify code now looks up an issuer certificate by a
2127 number of criteria: subject name, authority key id
2128 and key usage. It also verifies self signed certificates
2129 by the same criteria. The main comparison function is
2130 X509_check_issued() which performs these checks.
2132 Lot of changes were necessary in order to support this
2133 without completely rewriting the lookup code.
2135 Authority and subject key identifier are now cached.
2137 The LHASH 'certs' is X509_STORE has now been replaced
2138 by a STACK_OF(X509_OBJECT). This is mainly because an
2139 LHASH can't store or retrieve multiple objects with
2140 the same hash value.
2142 As a result various functions (which were all internal
2143 use only) have changed to handle the new X509_STORE
2144 structure. This will break anything that messed round
2145 with X509_STORE internally.
2147 The functions X509_STORE_add_cert() now checks for an
2148 exact match, rather than just subject name.
2150 The X509_STORE API doesn't directly support the retrieval
2151 of multiple certificates matching a given criteria, however
2152 this can be worked round by performing a lookup first
2153 (which will fill the cache with candidate certificates)
2154 and then examining the cache for matches. This is probably
2155 the best we can do without throwing out X509_LOOKUP
2156 entirely (maybe later...).
2158 The X509_VERIFY_CTX structure has been enhanced considerably.
2160 All certificate lookup operations now go via a get_issuer()
2161 callback. Although this currently uses an X509_STORE it
2162 can be replaced by custom lookups. This is a simple way
2163 to bypass the X509_STORE hackery necessary to make this
2164 work and makes it possible to use more efficient techniques
2165 in future. A very simple version which uses a simple
2166 STACK for its trusted certificate store is also provided
2167 using X509_STORE_CTX_trusted_stack().
2169 The verify_cb() and verify() callbacks now have equivalents
2170 in the X509_STORE_CTX structure.
2172 X509_STORE_CTX also has a 'flags' field which can be used
2173 to customise the verify behaviour.
2176 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2177 excludes S/MIME capabilities.
2180 *) When a certificate request is read in keep a copy of the
2181 original encoding of the signed data and use it when outputing
2182 again. Signatures then use the original encoding rather than
2183 a decoded, encoded version which may cause problems if the
2184 request is improperly encoded.
2187 *) For consistency with other BIO_puts implementations, call
2188 buffer_write(b, ...) directly in buffer_puts instead of calling
2191 In BIO_puts, increment b->num_write as in BIO_write.
2192 [Peter.Sylvester@EdelWeb.fr]
2194 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2195 BN_zero, we may not return a BIGNUM with an array consisting of
2199 *) Avoid calling abort() from within the library when problems are
2200 detected, except if preprocessor symbols have been defined
2201 (such as REF_CHECK, BN_DEBUG etc.).
2204 *) New openssl application 'rsautl'. This utility can be
2205 used for low level RSA operations. DER public key
2206 BIO/fp routines also added.
2209 *) New Configure entry and patches for compiling on QNX 4.
2210 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2212 *) A demo state-machine implementation was sponsored by
2213 Nuron (http://www.nuron.com/) and is now available in
2214 demos/state_machine.
2217 *) New options added to the 'dgst' utility for signature
2218 generation and verification.
2221 *) Unrecognized PKCS#7 content types are now handled via a
2222 catch all ASN1_TYPE structure. This allows unsupported
2223 types to be stored as a "blob" and an application can
2224 encode and decode it manually.
2227 *) Fix various signed/unsigned issues to make a_strex.c
2229 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2231 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2232 length if passed a buffer. ASN1_INTEGER_to_BN failed
2233 if passed a NULL BN and its argument was negative.
2234 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2236 *) Modification to PKCS#7 encoding routines to output definite
2237 length encoding. Since currently the whole structures are in
2238 memory there's not real point in using indefinite length
2239 constructed encoding. However if OpenSSL is compiled with
2240 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2243 *) Added BIO_vprintf() and BIO_vsnprintf().
2246 *) Added more prefixes to parse for in the the strings written
2247 through a logging bio, to cover all the levels that are available
2248 through syslog. The prefixes are now:
2250 PANIC, EMERG, EMR => LOG_EMERG
2251 ALERT, ALR => LOG_ALERT
2252 CRIT, CRI => LOG_CRIT
2253 ERROR, ERR => LOG_ERR
2254 WARNING, WARN, WAR => LOG_WARNING
2255 NOTICE, NOTE, NOT => LOG_NOTICE
2256 INFO, INF => LOG_INFO
2257 DEBUG, DBG => LOG_DEBUG
2259 and as before, if none of those prefixes are present at the
2260 beginning of the string, LOG_ERR is chosen.
2262 On Win32, the LOG_* levels are mapped according to this:
2264 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2265 LOG_WARNING => EVENTLOG_WARNING_TYPE
2266 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2270 *) Made it possible to reconfigure with just the configuration
2271 argument "reconf" or "reconfigure". The command line arguments
2272 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2273 and are retrieved from there when reconfiguring.
2277 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2279 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2282 *) The obj_dat.pl script was messing up the sorting of object
2283 names. The reason was that it compared the quoted version
2284 of strings as a result "OCSP" > "OCSP Signing" because
2285 " > SPACE. Changed script to store unquoted versions of
2286 names and add quotes on output. It was also omitting some
2287 names from the lookup table if they were given a default
2288 value (that is if SN is missing it is given the same
2289 value as LN and vice versa), these are now added on the
2290 grounds that if an object has a name we should be able to
2291 look it up. Finally added warning output when duplicate
2292 short or long names are found.
2295 *) Changes needed for Tandem NSK.
2296 [Scott Uroff <scott@xypro.com>]
2298 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2299 RSA_padding_check_SSLv23(), special padding was never detected
2300 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2301 version rollback attacks was not effective.
2303 In s23_clnt.c, don't use special rollback-attack detection padding
2304 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2305 client; similarly, in s23_srvr.c, don't do the rollback check if
2306 SSL 2.0 is the only protocol enabled in the server.
2309 *) Make it possible to get hexdumps of unprintable data with 'openssl
2310 asn1parse'. By implication, the functions ASN1_parse_dump() and
2311 BIO_dump_indent() are added.
2314 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2315 these print out strings and name structures based on various
2316 flags including RFC2253 support and proper handling of
2317 multibyte characters. Added options to the 'x509' utility
2318 to allow the various flags to be set.
2321 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2322 Also change the functions X509_cmp_current_time() and
2323 X509_gmtime_adj() work with an ASN1_TIME structure,
2324 this will enable certificates using GeneralizedTime in validity
2325 dates to be checked.
2328 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2329 negative public key encodings) on by default,
2330 NO_NEG_PUBKEY_BUG can be set to disable it.
2333 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2334 content octets. An i2c_ASN1_OBJECT is unnecessary because
2335 the encoding can be trivially obtained from the structure.
2338 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2339 not read locks (CRYPTO_r_[un]lock).
2342 *) A first attempt at creating official support for shared
2343 libraries through configuration. I've kept it so the
2344 default is static libraries only, and the OpenSSL programs
2345 are always statically linked for now, but there are
2346 preparations for dynamic linking in place.
2347 This has been tested on Linux and Tru64.
2350 *) Randomness polling function for Win9x, as described in:
2351 Peter Gutmann, Software Generation of Practically Strong
2355 *) Fix so PRNG is seeded in req if using an already existing
2359 *) New options to smime application. -inform and -outform
2360 allow alternative formats for the S/MIME message including
2361 PEM and DER. The -content option allows the content to be
2362 specified separately. This should allow things like Netscape
2363 form signing output easier to verify.
2366 *) Fix the ASN1 encoding of tags using the 'long form'.
2369 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2370 STRING types. These convert content octets to and from the
2371 underlying type. The actual tag and length octets are
2372 already assumed to have been read in and checked. These
2373 are needed because all other string types have virtually
2374 identical handling apart from the tag. By having versions
2375 of the ASN1 functions that just operate on content octets
2376 IMPLICIT tagging can be handled properly. It also allows
2377 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2378 and ASN1_INTEGER are identical apart from the tag.
2381 *) Change the handling of OID objects as follows:
2383 - New object identifiers are inserted in objects.txt, following
2384 the syntax given in objects.README.
2385 - objects.pl is used to process obj_mac.num and create a new
2387 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2390 This is currently kind of a hack, and the perl code in objects.pl
2391 isn't very elegant, but it works as I intended. The simplest way
2392 to check that it worked correctly is to look in obj_dat.h and
2393 check the array nid_objs and make sure the objects haven't moved
2394 around (this is important!). Additions are OK, as well as
2395 consistent name changes.
2398 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2401 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2402 The given file adds to whatever has already been seeded into the
2403 random pool through the RANDFILE configuration file option or
2404 environment variable, or the default random state file.
2407 *) mkstack.pl now sorts each macro group into lexical order.
2408 Previously the output order depended on the order the files
2409 appeared in the directory, resulting in needless rewriting
2413 *) Patches to make OpenSSL compile under Win32 again. Mostly
2414 work arounds for the VC++ problem that it treats func() as
2415 func(void). Also stripped out the parts of mkdef.pl that
2416 added extra typesafe functions: these no longer exist.
2419 *) Reorganisation of the stack code. The macros are now all
2420 collected in safestack.h . Each macro is defined in terms of
2421 a "stack macro" of the form SKM_<name>(type, a, b). The
2422 DEBUG_SAFESTACK is now handled in terms of function casts,
2423 this has the advantage of retaining type safety without the
2424 use of additional functions. If DEBUG_SAFESTACK is not defined
2425 then the non typesafe macros are used instead. Also modified the
2426 mkstack.pl script to handle the new form. Needs testing to see
2427 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2428 the default if no major problems. Similar behaviour for ASN1_SET_OF
2429 and PKCS12_STACK_OF.
2432 *) When some versions of IIS use the 'NET' form of private key the
2433 key derivation algorithm is different. Normally MD5(password) is
2434 used as a 128 bit RC4 key. In the modified case
2435 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2436 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2437 as the old Netscape_RSA functions except they have an additional
2438 'sgckey' parameter which uses the modified algorithm. Also added
2439 an -sgckey command line option to the rsa utility. Thanks to
2440 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2441 algorithm to openssl-dev.
2444 *) The evp_local.h macros were using 'c.##kname' which resulted in
2445 invalid expansion on some systems (SCO 5.0.5 for example).
2446 Corrected to 'c.kname'.
2447 [Phillip Porch <root@theporch.com>]
2449 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2450 a STACK of email addresses from a certificate or request, these look
2451 in the subject name and the subject alternative name extensions and
2452 omit any duplicate addresses.
2455 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2456 This makes DSA verification about 2 % faster.
2459 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2460 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2461 plus overhead for 1024 bit moduli).
2462 This makes exponentiations about 0.5 % faster for 1024 bit
2463 exponents (as measured by "openssl speed rsa2048").
2466 *) Rename memory handling macros to avoid conflicts with other
2468 Malloc => OPENSSL_malloc
2469 Malloc_locked => OPENSSL_malloc_locked
2470 Realloc => OPENSSL_realloc
2471 Free => OPENSSL_free
2474 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2475 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2478 *) CygWin32 support.
2479 [John Jarvie <jjarvie@newsguy.com>]
2481 *) The type-safe stack code has been rejigged. It is now only compiled
2482 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2483 by default all type-specific stack functions are "#define"d back to
2484 standard stack functions. This results in more streamlined output
2485 but retains the type-safety checking possibilities of the original
2489 *) The STACK code has been cleaned up, and certain type declarations
2490 that didn't make a lot of sense have been brought in line. This has
2491 also involved a cleanup of sorts in safestack.h to more correctly
2492 map type-safe stack functions onto their plain stack counterparts.
2493 This work has also resulted in a variety of "const"ifications of
2494 lots of the code, especially "_cmp" operations which should normally
2495 be prototyped with "const" parameters anyway.
2498 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2499 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2500 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2501 where all of 'md' is used each time the PRNG is used, but 'state'
2502 is used only indexed by a cyclic counter. As entropy may not be
2503 well distributed from the beginning, 'md' is important as a
2504 chaining variable. However, the output function chains only half
2505 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2506 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2507 in all of 'state' being rewritten, with the new values depending
2508 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2511 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2512 the handshake is continued after ssl_verify_cert_chain();
2513 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2514 can lead to 'unexplainable' connection aborts later.
2515 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2517 *) Major EVP API cipher revision.
2518 Add hooks for extra EVP features. This allows various cipher
2519 parameters to be set in the EVP interface. Support added for variable
2520 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2521 setting of RC2 and RC5 parameters.
2523 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2526 Remove lots of duplicated code from the EVP library. For example *every*
2527 cipher init() function handles the 'iv' in the same way according to the
2528 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2529 for CFB and OFB modes they zero ctx->num.
2531 New functionality allows removal of S/MIME code RC2 hack.
2533 Most of the routines have the same form and so can be declared in terms
2536 By shifting this to the top level EVP_CipherInit() it can be removed from
2537 all individual ciphers. If the cipher wants to handle IVs or keys
2538 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2541 Change lots of functions like EVP_EncryptUpdate() to now return a
2542 value: although software versions of the algorithms cannot fail
2543 any installed hardware versions can.
2546 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2547 this option is set, tolerate broken clients that send the negotiated
2548 protocol version number instead of the requested protocol version
2552 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2553 i.e. non-zero for export ciphersuites, zero otherwise.
2554 Previous versions had this flag inverted, inconsistent with
2555 rsa_tmp_cb (..._TMP_RSA_CB).
2556 [Bodo Moeller; problem reported by Amit Chopra]
2558 *) Add missing DSA library text string. Work around for some IIS
2559 key files with invalid SEQUENCE encoding.
2562 *) Add a document (doc/standards.txt) that list all kinds of standards
2563 and so on that are implemented in OpenSSL.
2566 *) Enhance c_rehash script. Old version would mishandle certificates
2567 with the same subject name hash and wouldn't handle CRLs at all.
2568 Added -fingerprint option to crl utility, to support new c_rehash
2572 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2575 *) Fix for SSL server purpose checking. Server checking was
2576 rejecting certificates which had extended key usage present
2577 but no ssl client purpose.
2578 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2580 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2581 is a little unclear about how a blank password is handled.
2582 Since the password in encoded as a BMPString with terminating
2583 double NULL a zero length password would end up as just the
2584 double NULL. However no password at all is different and is
2585 handled differently in the PKCS#12 key generation code. NS
2586 treats a blank password as zero length. MSIE treats it as no
2587 password on export: but it will try both on import. We now do
2588 the same: PKCS12_parse() tries zero length and no password if
2589 the password is set to "" or NULL (NULL is now a valid password:
2590 it wasn't before) as does the pkcs12 application.
2593 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2594 perror when PEM_read_bio_X509_REQ fails, the error message must
2595 be obtained from the error queue.
2598 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2599 it in ERR_remove_state if appropriate, and change ERR_get_state
2600 accordingly to avoid race conditions (this is necessary because
2601 thread_hash is no longer constant once set).
2604 *) Bugfix for linux-elf makefile.one.
2607 *) RSA_get_default_method() will now cause a default
2608 RSA_METHOD to be chosen if one doesn't exist already.
2609 Previously this was only set during a call to RSA_new()
2610 or RSA_new_method(NULL) meaning it was possible for
2611 RSA_get_default_method() to return NULL.
2614 *) Added native name translation to the existing DSO code
2615 that will convert (if the flag to do so is set) filenames
2616 that are sufficiently small and have no path information
2617 into a canonical native form. Eg. "blah" converted to
2618 "libblah.so" or "blah.dll" etc.
2621 *) New function ERR_error_string_n(e, buf, len) which is like
2622 ERR_error_string(e, buf), but writes at most 'len' bytes
2623 including the 0 terminator. For ERR_error_string_n, 'buf'
2625 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2627 *) CONF library reworked to become more general. A new CONF
2628 configuration file reader "class" is implemented as well as a
2629 new functions (NCONF_*, for "New CONF") to handle it. The now
2630 old CONF_* functions are still there, but are reimplemented to
2631 work in terms of the new functions. Also, a set of functions
2632 to handle the internal storage of the configuration data is
2633 provided to make it easier to write new configuration file
2634 reader "classes" (I can definitely see something reading a
2635 configuration file in XML format, for example), called _CONF_*,
2636 or "the configuration storage API"...
2638 The new configuration file reading functions are:
2640 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2641 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2643 NCONF_default, NCONF_WIN32
2645 NCONF_dump_fp, NCONF_dump_bio
2647 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2648 NCONF_new creates a new CONF object. This works in the same way
2649 as other interfaces in OpenSSL, like the BIO interface.
2650 NCONF_dump_* dump the internal storage of the configuration file,
2651 which is useful for debugging. All other functions take the same
2652 arguments as the old CONF_* functions wth the exception of the
2653 first that must be a `CONF *' instead of a `LHASH *'.
2655 To make it easer to use the new classes with the old CONF_* functions,
2656 the function CONF_set_default_method is provided.
2659 *) Add '-tls1' option to 'openssl ciphers', which was already
2660 mentioned in the documentation but had not been implemented.
2661 (This option is not yet really useful because even the additional
2662 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2665 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2666 OpenSSL-based applications) load shared libraries and bind to
2667 them in a portable way.
2668 [Geoff Thorpe, with contributions from Richard Levitte]
2670 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2672 *) Make sure _lrotl and _lrotr are only used with MSVC.
2674 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2675 (the default implementation of RAND_status).
2677 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2678 to '-clrext' (= clear extensions), as intended and documented.
2679 [Bodo Moeller; inconsistency pointed out by Michael Attili
2682 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2683 was larger than the MD block size.
2684 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2686 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2687 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2688 using the passed key: if the passed key was a private key the result
2689 of X509_print(), for example, would be to print out all the private key
2693 *) des_quad_cksum() byte order bug fix.
2694 [Ulf Möller, using the problem description in krb4-0.9.7, where
2695 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2697 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2699 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2701 *) For easily testing in shell scripts whether some command
2702 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2703 returns with exit code 0 iff no command of the given name is available.
2704 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2705 the output goes to stdout and nothing is printed to stderr.
2706 Additional arguments are always ignored.
2708 Since for each cipher there is a command of the same name,
2709 the 'no-cipher' compilation switches can be tested this way.
2711 ('openssl no-XXX' is not able to detect pseudo-commands such
2712 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2715 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2718 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2719 is set; it will be thrown away anyway because each handshake creates
2721 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2722 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2723 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2724 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2727 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2728 'Q' and 'R' lose their special meanings (quit/renegotiate).
2729 This is part of what -quiet does; unlike -quiet, -ign_eof
2730 does not suppress any output.
2733 *) Add compatibility options to the purpose and trust code. The
2734 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2735 accepts a certificate or CA, this was the previous behaviour,
2736 with all the associated security issues.
2738 X509_TRUST_COMPAT is the old trust behaviour: only and
2739 automatically trust self signed roots in certificate store. A
2740 new trust setting X509_TRUST_DEFAULT is used to specify that
2741 a purpose has no associated trust setting and it should instead
2742 use the value in the default purpose.
2745 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2746 and fix a memory leak.
2749 *) In util/mkerr.pl (which implements 'make errors'), preserve
2750 reason strings from the previous version of the .c file, as
2751 the default to have only downcase letters (and digits) in
2752 automatically generated reasons codes is not always appropriate.
2755 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2756 using strerror. Previously, ERR_reason_error_string() returned
2757 library names as reason strings for SYSerr; but SYSerr is a special
2758 case where small numbers are errno values, not library numbers.
2761 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2762 converts DSA parameters into DH parameters. (When creating parameters,
2763 DSA_generate_parameters is used.)
2766 *) Include 'length' (recommended exponent length) in C code generated
2767 by 'openssl dhparam -C'.
2770 *) The second argument to set_label in perlasm was already being used
2771 so couldn't be used as a "file scope" flag. Moved to third argument
2775 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2776 instead of RAND_bytes for encryption IVs and salts.
2779 *) Include RAND_status() into RAND_METHOD instead of implementing
2780 it only for md_rand.c Otherwise replacing the PRNG by calling
2781 RAND_set_rand_method would be impossible.
2784 *) Don't let DSA_generate_key() enter an infinite loop if the random
2785 number generation fails.
2788 *) New 'rand' application for creating pseudo-random output.
2791 *) Added configuration support for Linux/IA64
2792 [Rolf Haberrecker <rolf@suse.de>]
2794 *) Assembler module support for Mingw32.
2797 *) Shared library support for HPUX (in shlib/).
2798 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2800 *) Shared library support for Solaris gcc.
2801 [Lutz Behnke <behnke@trustcenter.de>]
2803 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2805 *) PKCS7_encrypt() was adding text MIME headers twice because they
2806 were added manually and by SMIME_crlf_copy().
2809 *) In bntest.c don't call BN_rand with zero bits argument.
2810 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
2812 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
2813 case was implemented. This caused BN_div_recp() to fail occasionally.
2816 *) Add an optional second argument to the set_label() in the perl
2817 assembly language builder. If this argument exists and is set
2818 to 1 it signals that the assembler should use a symbol whose
2819 scope is the entire file, not just the current function. This
2820 is needed with MASM which uses the format label:: for this scope.
2821 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
2823 *) Change the ASN1 types so they are typedefs by default. Before
2824 almost all types were #define'd to ASN1_STRING which was causing
2825 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
2829 *) Change names of new functions to the new get1/get0 naming
2830 convention: After 'get1', the caller owns a reference count
2831 and has to call ..._free; 'get0' returns a pointer to some
2832 data structure without incrementing reference counters.
2833 (Some of the existing 'get' functions increment a reference
2834 counter, some don't.)
2835 Similarly, 'set1' and 'add1' functions increase reference
2836 counters or duplicate objects.
2839 *) Allow for the possibility of temp RSA key generation failure:
2840 the code used to assume it always worked and crashed on failure.
2843 *) Fix potential buffer overrun problem in BIO_printf().
2844 [Ulf Möller, using public domain code by Patrick Powell; problem
2845 pointed out by David Sacerdote <das33@cornell.edu>]
2847 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
2848 RAND_egd() and RAND_status(). In the command line application,
2849 the EGD socket can be specified like a seed file using RANDFILE
2853 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
2854 Some CAs (e.g. Verisign) distribute certificates in this form.
2857 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
2858 list to exclude them. This means that no special compilation option
2859 is needed to use anonymous DH: it just needs to be included in the
2863 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
2864 EVP_MD_type. The old functionality is available in a new macro called
2865 EVP_MD_md(). Change code that uses it and update docs.
2868 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
2869 where the 'void *' argument is replaced by a function pointer argument.
2870 Previously 'void *' was abused to point to functions, which works on
2871 many platforms, but is not correct. As these functions are usually
2872 called by macros defined in OpenSSL header files, most source code
2873 should work without changes.
2876 *) <openssl/opensslconf.h> (which is created by Configure) now contains
2877 sections with information on -D... compiler switches used for
2878 compiling the library so that applications can see them. To enable
2879 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
2880 must be defined. E.g.,
2881 #define OPENSSL_ALGORITHM_DEFINES
2882 #include <openssl/opensslconf.h>
2883 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
2884 [Richard Levitte, Ulf and Bodo Möller]
2886 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
2890 *) Change the 'other' type in certificate aux info to a STACK_OF
2891 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
2892 the required ASN1 format: arbitrary types determined by an OID.
2895 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
2896 argument to 'req'. This is not because the function is newer or
2897 better than others it just uses the work 'NEW' in the certificate
2898 request header lines. Some software needs this.
2901 *) Reorganise password command line arguments: now passwords can be
2902 obtained from various sources. Delete the PEM_cb function and make
2903 it the default behaviour: i.e. if the callback is NULL and the
2904 usrdata argument is not NULL interpret it as a null terminated pass
2905 phrase. If usrdata and the callback are NULL then the pass phrase
2906 is prompted for as usual.
2909 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
2910 the support is automatically enabled. The resulting binaries will
2911 autodetect the card and use it if present.
2912 [Ben Laurie and Compaq Inc.]
2914 *) Work around for Netscape hang bug. This sends certificate request
2915 and server done in one record. Since this is perfectly legal in the
2916 SSL/TLS protocol it isn't a "bug" option and is on by default. See
2917 the bugs/SSLv3 entry for more info.
2920 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
2923 *) Add -rand argument to smime and pkcs12 applications and read/write
2927 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
2930 *) Add command line password options to the remaining applications.
2933 *) Bug fix for BN_div_recp() for numerators with an even number of
2937 *) More tests in bntest.c, and changed test_bn output.
2940 *) ./config recognizes MacOS X now.
2943 *) Bug fix for BN_div() when the first words of num and divsor are
2944 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
2947 *) Add support for various broken PKCS#8 formats, and command line
2948 options to produce them.
2951 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
2952 get temporary BIGNUMs from a BN_CTX.
2955 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
2959 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
2960 include a #define from the old name to the new. The original intent
2961 was that statically linked binaries could for example just call
2962 SSLeay_add_all_ciphers() to just add ciphers to the table and not
2963 link with digests. This never worked becayse SSLeay_add_all_digests()
2964 and SSLeay_add_all_ciphers() were in the same source file so calling
2965 one would link with the other. They are now in separate source files.
2968 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
2971 *) Use a less unusual form of the Miller-Rabin primality test (it used
2972 a binary algorithm for exponentiation integrated into the Miller-Rabin
2973 loop, our standard modexp algorithms are faster).
2976 *) Support for the EBCDIC character set completed.
2977 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
2979 *) Source code cleanups: use const where appropriate, eliminate casts,
2980 use void * instead of char * in lhash.
2983 *) Bugfix: ssl3_send_server_key_exchange was not restartable
2984 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
2985 this the server could overwrite ephemeral keys that the client
2989 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
2990 using 50 iterations of the Rabin-Miller test.
2992 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
2993 iterations of the Rabin-Miller test as required by the appendix
2994 to FIPS PUB 186[-1]) instead of DSA_is_prime.
2995 As BN_is_prime_fasttest includes trial division, DSA parameter
2996 generation becomes much faster.
2998 This implies a change for the callback functions in DSA_is_prime
2999 and DSA_generate_parameters: The callback function is called once
3000 for each positive witness in the Rabin-Miller test, not just
3001 occasionally in the inner loop; and the parameters to the
3002 callback function now provide an iteration count for the outer
3003 loop rather than for the current invocation of the inner loop.
3004 DSA_generate_parameters additionally can call the callback
3005 function with an 'iteration count' of -1, meaning that a
3006 candidate has passed the trial division test (when q is generated
3007 from an application-provided seed, trial division is skipped).
3010 *) New function BN_is_prime_fasttest that optionally does trial
3011 division before starting the Rabin-Miller test and has
3012 an additional BN_CTX * argument (whereas BN_is_prime always
3013 has to allocate at least one BN_CTX).
3014 'callback(1, -1, cb_arg)' is called when a number has passed the
3015 trial division stage.
3018 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3022 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3025 *) New function BN_pseudo_rand().
3028 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3029 bignum version of BN_from_montgomery() with the working code from
3030 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3034 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3035 made it impossible to use the same SSL_SESSION data structure in
3036 SSL2 clients in multiple threads.
3039 *) The return value of RAND_load_file() no longer counts bytes obtained
3040 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3041 to seed the PRNG (previously an explicit byte count was required).
3042 [Ulf Möller, Bodo Möller]
3044 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3045 used (char *) instead of (void *) and had casts all over the place.
3048 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3051 *) Retain source code compatibility for BN_prime_checks macro:
3052 BN_is_prime(..., BN_prime_checks, ...) now uses
3053 BN_prime_checks_for_size to determine the appropriate number of
3054 Rabin-Miller iterations.
3057 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3058 DH_CHECK_P_NOT_SAFE_PRIME.
3059 (Check if this is true? OpenPGP calls them "strong".)
3062 *) Merge the functionality of "dh" and "gendh" programs into a new program
3063 "dhparam". The old programs are retained for now but will handle DH keys
3064 (instead of parameters) in future.
3067 *) Make the ciphers, s_server and s_client programs check the return values
3068 when a new cipher list is set.
3071 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3072 ciphers. Before when the 56bit ciphers were enabled the sorting was
3075 The syntax for the cipher sorting has been extended to support sorting by
3076 cipher-strength (using the strength_bits hard coded in the tables).
3077 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3079 Fix a bug in the cipher-command parser: when supplying a cipher command
3080 string with an "undefined" symbol (neither command nor alphanumeric
3081 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3082 an error is flagged.
3084 Due to the strength-sorting extension, the code of the
3085 ssl_create_cipher_list() function was completely rearranged. I hope that
3086 the readability was also increased :-)
3087 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3089 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3090 for the first serial number and places 2 in the serial number file. This
3091 avoids problems when the root CA is created with serial number zero and
3092 the first user certificate has the same issuer name and serial number
3096 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3097 the new code. Add documentation for this stuff.
3100 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3101 X509_*() to X509at_*() on the grounds that they don't handle X509
3102 structures and behave in an analagous way to the X509v3 functions:
3103 they shouldn't be called directly but wrapper functions should be used
3106 So we also now have some wrapper functions that call the X509at functions
3107 when passed certificate requests. (TO DO: similar things can be done with
3108 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3109 things. Some of these need some d2i or i2d and print functionality
3110 because they handle more complex structures.)
3113 *) Add missing #ifndefs that caused missing symbols when building libssl
3114 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3115 NO_RSA in ssl/s2*.c.
3116 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3118 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3119 has a return value which indicates the quality of the random data
3120 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3121 error queue. New function RAND_pseudo_bytes() generates output that is
3122 guaranteed to be unique but not unpredictable. RAND_add is like
3123 RAND_seed, but takes an extra argument for an entropy estimate
3124 (RAND_seed always assumes full entropy).
3127 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3128 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3129 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3130 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3131 false-positive rate of at most 2^-80 for random input.
3134 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3137 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3138 in the 0.9.5 release), this returns the chain
3139 from an X509_CTX structure with a dup of the stack and all
3140 the X509 reference counts upped: so the stack will exist
3141 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3144 Also make SSL_SESSION_print() print out the verify return
3148 *) Add manpage for the pkcs12 command. Also change the default
3149 behaviour so MAC iteration counts are used unless the new
3150 -nomaciter option is used. This improves file security and
3151 only older versions of MSIE (4.0 for example) need it.
3154 *) Honor the no-xxx Configure options when creating .DEF files.
3157 *) Add PKCS#10 attributes to field table: challengePassword,
3158 unstructuredName and unstructuredAddress. These are taken from
3159 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3160 international characters are used.
3162 More changes to X509_ATTRIBUTE code: allow the setting of types
3163 based on strings. Remove the 'loc' parameter when adding
3164 attributes because these will be a SET OF encoding which is sorted
3168 *) Initial changes to the 'req' utility to allow request generation
3169 automation. This will allow an application to just generate a template
3170 file containing all the field values and have req construct the
3173 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3174 used all over the place including certificate requests and PKCS#7
3175 structures. They are currently handled manually where necessary with
3176 some primitive wrappers for PKCS#7. The new functions behave in a
3177 manner analogous to the X509 extension functions: they allow
3178 attributes to be looked up by NID and added.
3180 Later something similar to the X509V3 code would be desirable to
3181 automatically handle the encoding, decoding and printing of the
3182 more complex types. The string types like challengePassword can
3183 be handled by the string table functions.
3185 Also modified the multi byte string table handling. Now there is
3186 a 'global mask' which masks out certain types. The table itself
3187 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3188 is useful when for example there is only one permissible type
3189 (as in countryName) and using the mask might result in no valid
3193 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3194 SSL_get_peer_finished to allow applications to obtain the latest
3195 Finished messages sent to the peer or expected from the peer,
3196 respectively. (SSL_get_peer_finished is usually the Finished message
3197 actually received from the peer, otherwise the protocol will be aborted.)
3199 As the Finished message are message digests of the complete handshake
3200 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3201 be used for external authentication procedures when the authentication
3202 provided by SSL/TLS is not desired or is not enough.
3205 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3206 the host supports BWX extension and if Compaq C is present on the
3207 $PATH. Just exploiting of the BWX extension results in 20-30%
3208 performance kick for some algorithms, e.g. DES and RC4 to mention
3209 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3213 *) Add support for MS "fast SGC". This is arguably a violation of the
3214 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3215 weak crypto and after checking the certificate is SGC a second one
3216 with strong crypto. MS SGC stops the first handshake after receiving
3217 the server certificate message and sends a second client hello. Since
3218 a server will typically do all the time consuming operations before
3219 expecting any further messages from the client (server key exchange
3220 is the most expensive) there is little difference between the two.
3222 To get OpenSSL to support MS SGC we have to permit a second client
3223 hello message after we have sent server done. In addition we have to
3224 reset the MAC if we do get this second client hello.
3227 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3228 if a DER encoded private key is RSA or DSA traditional format. Changed
3229 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3230 format DER encoded private key. Newer code should use PKCS#8 format which
3231 has the key type encoded in the ASN1 structure. Added DER private key
3232 support to pkcs8 application.
3235 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3236 ciphersuites has been selected (as required by the SSL 3/TLS 1
3237 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3238 is set, we interpret this as a request to violate the specification
3239 (the worst that can happen is a handshake failure, and 'correct'
3240 behaviour would result in a handshake failure anyway).
3243 *) In SSL_CTX_add_session, take into account that there might be multiple
3244 SSL_SESSION structures with the same session ID (e.g. when two threads
3245 concurrently obtain them from an external cache).
3246 The internal cache can handle only one SSL_SESSION with a given ID,
3247 so if there's a conflict, we now throw out the old one to achieve
3251 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3252 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3253 some routines that use cipher OIDs: some ciphers do not have OIDs
3254 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3258 *) Simplify the trust setting structure and code. Now we just have
3259 two sequences of OIDs for trusted and rejected settings. These will
3260 typically have values the same as the extended key usage extension
3261 and any application specific purposes.
3263 The trust checking code now has a default behaviour: it will just
3264 check for an object with the same NID as the passed id. Functions can
3265 be provided to override either the default behaviour or the behaviour
3266 for a given id. SSL client, server and email already have functions
3267 in place for compatibility: they check the NID and also return "trusted"
3268 if the certificate is self signed.
3271 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3272 traditional format into an EVP_PKEY structure.
3275 *) Add a password callback function PEM_cb() which either prompts for
3276 a password if usr_data is NULL or otherwise assumes it is a null
3277 terminated password. Allow passwords to be passed on command line
3278 environment or config files in a few more utilities.
3281 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3282 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3283 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3284 Update documentation.
3287 *) Support for ASN1 "NULL" type. This could be handled before by using
3288 ASN1_TYPE but there wasn't any function that would try to read a NULL
3289 and produce an error if it couldn't. For compatibility we also have
3290 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3291 don't allocate anything because they don't need to.
3294 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3296 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3298 *) Rebuild of the memory allocation routines used by OpenSSL code and
3299 possibly others as well. The purpose is to make an interface that
3300 provide hooks so anyone can build a separate set of allocation and
3301 deallocation routines to be used by OpenSSL, for example memory
3302 pool implementations, or something else, which was previously hard
3303 since Malloc(), Realloc() and Free() were defined as macros having
3304 the values malloc, realloc and free, respectively (except for Win32
3305 compilations). The same is provided for memory debugging code.
3306 OpenSSL already comes with functionality to find memory leaks, but
3307 this gives people a chance to debug other memory problems.
3309 With these changes, a new set of functions and macros have appeared:
3311 CRYPTO_set_mem_debug_functions() [F]
3312 CRYPTO_get_mem_debug_functions() [F]
3313 CRYPTO_dbg_set_options() [F]
3314 CRYPTO_dbg_get_options() [F]
3315 CRYPTO_malloc_debug_init() [M]
3317 The memory debug functions are NULL by default, unless the library
3318 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3319 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3320 gives the standard debugging functions that come with OpenSSL) or
3321 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3322 provided by the library user) must be used. When the standard
3323 debugging functions are used, CRYPTO_dbg_set_options can be used to
3324 request additional information:
3325 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3326 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3328 Also, things like CRYPTO_set_mem_functions will always give the
3329 expected result (the new set of functions is used for allocation
3330 and deallocation) at all times, regardless of platform and compiler
3333 To finish it up, some functions that were never use in any other
3334 way than through macros have a new API and new semantic:
3337 CRYPTO_dbg_realloc()
3340 All macros of value have retained their old syntax.
3341 [Richard Levitte and Bodo Moeller]
3343 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3344 ordering of SMIMECapabilities wasn't in "strength order" and there
3345 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3349 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3350 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3351 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3353 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3354 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3355 functionality to handle multipart/signed properly) and a utility
3356 called 'smime' to call all this stuff. This is based on code I
3357 originally wrote for Celo who have kindly allowed it to be
3358 included in OpenSSL.
3361 *) Add variants des_set_key_checked and des_set_key_unchecked of
3362 des_set_key (aka des_key_sched). Global variable des_check_key
3363 decides which of these is called by des_set_key; this way
3364 des_check_key behaves as it always did, but applications and
3365 the library itself, which was buggy for des_check_key == 1,
3366 have a cleaner way to pick the version they need.
3369 *) New function PKCS12_newpass() which changes the password of a
3373 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3374 dynamic mix. In both cases the ids can be used as an index into the
3375 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3376 functions so they accept a list of the field values and the
3377 application doesn't need to directly manipulate the X509_TRUST
3381 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3385 *) Modify the way the V3 extension code looks up extensions. This now
3386 works in a similar way to the object code: we have some "standard"
3387 extensions in a static table which is searched with OBJ_bsearch()
3388 and the application can add dynamic ones if needed. The file
3389 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3390 updated whenever a new extension is added to the core code and kept
3391 in ext_nid order. There is a simple program 'tabtest.c' which checks
3392 this. New extensions are not added too often so this file can readily
3393 be maintained manually.
3395 There are two big advantages in doing things this way. The extensions
3396 can be looked up immediately and no longer need to be "added" using
3397 X509V3_add_standard_extensions(): this function now does nothing.
3398 [Side note: I get *lots* of email saying the extension code doesn't
3399 work because people forget to call this function]
3400 Also no dynamic allocation is done unless new extensions are added:
3401 so if we don't add custom extensions there is no need to call
3402 X509V3_EXT_cleanup().
3405 *) Modify enc utility's salting as follows: make salting the default. Add a
3406 magic header, so unsalted files fail gracefully instead of just decrypting
3407 to garbage. This is because not salting is a big security hole, so people
3408 should be discouraged from doing it.
3411 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3412 digest to be passed on the command line but it only used this
3413 parameter when signing a certificate. Modified so all relevant
3414 operations are affected by the digest parameter including the
3415 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3416 DSA key was used because it didn't fix the digest.
3419 *) Initial certificate chain verify code. Currently tests the untrusted
3420 certificates for consistency with the verify purpose (which is set
3421 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3423 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3424 this is because it will reject chains with invalid extensions whereas
3425 every previous version of OpenSSL and SSLeay made no checks at all.
3427 Trust code: checks the root CA for the relevant trust settings. Trust
3428 settings have an initial value consistent with the verify purpose: e.g.
3429 if the verify purpose is for SSL client use it expects the CA to be
3430 trusted for SSL client use. However the default value can be changed to
3431 permit custom trust settings: one example of this would be to only trust
3432 certificates from a specific "secure" set of CAs.
3434 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3435 which should be used for version portability: especially since the
3436 verify structure is likely to change more often now.
3438 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3439 to set them. If not set then assume SSL clients will verify SSL servers
3442 Two new options to the verify program: -untrusted allows a set of
3443 untrusted certificates to be passed in and -purpose which sets the
3444 intended purpose of the certificate. If a purpose is set then the
3445 new chain verify code is used to check extension consistency.
3448 *) Support for the authority information access extension.
3451 *) Modify RSA and DSA PEM read routines to transparently handle
3452 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3453 public keys in a format compatible with certificate
3454 SubjectPublicKeyInfo structures. Unfortunately there were already
3455 functions called *_PublicKey_* which used various odd formats so
3456 these are retained for compatibility: however the DSA variants were
3457 never in a public release so they have been deleted. Changed dsa/rsa
3458 utilities to handle the new format: note no releases ever handled public
3459 keys so we should be OK.
3461 The primary motivation for this change is to avoid the same fiasco
3462 that dogs private keys: there are several incompatible private key
3463 formats some of which are standard and some OpenSSL specific and
3464 require various evil hacks to allow partial transparent handling and
3465 even then it doesn't work with DER formats. Given the option anything
3466 other than PKCS#8 should be dumped: but the other formats have to
3467 stay in the name of compatibility.
3469 With public keys and the benefit of hindsight one standard format
3470 is used which works with EVP_PKEY, RSA or DSA structures: though
3471 it clearly returns an error if you try to read the wrong kind of key.
3473 Added a -pubkey option to the 'x509' utility to output the public key.
3474 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3475 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3476 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3477 that do the same as the EVP_PKEY_assign_*() except they up the
3478 reference count of the added key (they don't "swallow" the
3482 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3483 CRLs would fail if the file contained no certificates or no CRLs:
3484 added a new function to read in both types and return the number
3485 read: this means that if none are read it will be an error. The
3486 DER versions of the certificate and CRL reader would always fail
3487 because it isn't possible to mix certificates and CRLs in DER format
3488 without choking one or the other routine. Changed this to just read
3489 a certificate: this is the best we can do. Also modified the code
3490 in apps/verify.c to take notice of return codes: it was previously
3491 attempting to read in certificates from NULL pointers and ignoring
3492 any errors: this is one reason why the cert and CRL reader seemed
3493 to work. It doesn't check return codes from the default certificate
3494 routines: these may well fail if the certificates aren't installed.
3497 *) Code to support otherName option in GeneralName.
3500 *) First update to verify code. Change the verify utility
3501 so it warns if it is passed a self signed certificate:
3502 for consistency with the normal behaviour. X509_verify
3503 has been modified to it will now verify a self signed
3504 certificate if *exactly* the same certificate appears
3505 in the store: it was previously impossible to trust a
3506 single self signed certificate. This means that:
3507 openssl verify ss.pem
3508 now gives a warning about a self signed certificate but
3509 openssl verify -CAfile ss.pem ss.pem
3513 *) For servers, store verify_result in SSL_SESSION data structure
3514 (and add it to external session representation).
3515 This is needed when client certificate verifications fails,
3516 but an application-provided verification callback (set by
3517 SSL_CTX_set_cert_verify_callback) allows accepting the session
3518 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3519 but returns 1): When the session is reused, we have to set
3520 ssl->verify_result to the appropriate error code to avoid
3522 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3524 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3525 case in PKCS7_dataInit() where the signed PKCS7 structure
3526 didn't contain any existing data because it was being created.
3527 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3529 *) Add a salt to the key derivation routines in enc.c. This
3530 forms the first 8 bytes of the encrypted file. Also add a
3531 -S option to allow a salt to be input on the command line.
3534 *) New function X509_cmp(). Oddly enough there wasn't a function
3535 to compare two certificates. We do this by working out the SHA1
3536 hash and comparing that. X509_cmp() will be needed by the trust
3540 *) SSL_get1_session() is like SSL_get_session(), but increments
3541 the reference count in the SSL_SESSION returned.
3542 [Geoff Thorpe <geoff@eu.c2.net>]
3544 *) Fix for 'req': it was adding a null to request attributes.
3545 Also change the X509_LOOKUP and X509_INFO code to handle
3546 certificate auxiliary information.
3549 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3553 *) Add the possibility to add extra information to the memory leak
3554 detecting output, to form tracebacks, showing from where each
3555 allocation was originated: CRYPTO_push_info("constant string") adds
3556 the string plus current file name and line number to a per-thread
3557 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3558 is like calling CYRPTO_pop_info() until the stack is empty.
3559 Also updated memory leak detection code to be multi-thread-safe.
3562 *) Add options -text and -noout to pkcs7 utility and delete the
3563 encryption options which never did anything. Update docs.
3566 *) Add options to some of the utilities to allow the pass phrase
3567 to be included on either the command line (not recommended on
3568 OSes like Unix) or read from the environment. Update the
3569 manpages and fix a few bugs.
3572 *) Add a few manpages for some of the openssl commands.
3575 *) Fix the -revoke option in ca. It was freeing up memory twice,
3576 leaking and not finding already revoked certificates.
3579 *) Extensive changes to support certificate auxiliary information.
3580 This involves the use of X509_CERT_AUX structure and X509_AUX
3581 functions. An X509_AUX function such as PEM_read_X509_AUX()
3582 can still read in a certificate file in the usual way but it
3583 will also read in any additional "auxiliary information". By
3584 doing things this way a fair degree of compatibility can be
3585 retained: existing certificates can have this information added
3586 using the new 'x509' options.
3588 Current auxiliary information includes an "alias" and some trust
3589 settings. The trust settings will ultimately be used in enhanced
3590 certificate chain verification routines: currently a certificate
3591 can only be trusted if it is self signed and then it is trusted
3595 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3596 The problem was that one of the replacement routines had not been working
3597 since SSLeay releases. For now the offending routine has been replaced
3598 with non-optimised assembler. Even so, this now gives around 95%
3599 performance improvement for 1024 bit RSA signs.
3602 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3603 handling. Most clients have the effective key size in bits equal to
3604 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3605 A few however don't do this and instead use the size of the decrypted key
3606 to determine the RC2 key length and the AlgorithmIdentifier to determine
3607 the effective key length. In this case the effective key length can still
3608 be 40 bits but the key length can be 168 bits for example. This is fixed
3609 by manually forcing an RC2 key into the EVP_PKEY structure because the
3610 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3611 the key length and effective key length are equal.
3614 *) Add a bunch of functions that should simplify the creation of
3615 X509_NAME structures. Now you should be able to do:
3616 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3617 and have it automatically work out the correct field type and fill in
3618 the structures. The more adventurous can try:
3619 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3620 and it will (hopefully) work out the correct multibyte encoding.
3623 *) Change the 'req' utility to use the new field handling and multibyte
3624 copy routines. Before the DN field creation was handled in an ad hoc
3625 way in req, ca, and x509 which was rather broken and didn't support
3626 BMPStrings or UTF8Strings. Since some software doesn't implement
3627 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3628 using the dirstring_type option. See the new comment in the default
3629 openssl.cnf for more info.
3632 *) Make crypto/rand/md_rand.c more robust:
3633 - Assure unique random numbers after fork().
3634 - Make sure that concurrent threads access the global counter and
3635 md serializably so that we never lose entropy in them
3636 or use exactly the same state in multiple threads.
3637 Access to the large state is not always serializable because
3638 the additional locking could be a performance killer, and
3639 md should be large enough anyway.
3642 *) New file apps/app_rand.c with commonly needed functionality
3643 for handling the random seed file.
3645 Use the random seed file in some applications that previously did not:
3647 dsaparam -genkey (which also ignored its '-rand' option),
3650 x509 (when signing).
3651 Except on systems with /dev/urandom, it is crucial to have a random
3652 seed file at least for key creation, DSA signing, and for DH exchanges;
3653 for RSA signatures we could do without one.
3655 gendh and gendsa (unlike genrsa) used to read only the first byte
3656 of each file listed in the '-rand' option. The function as previously
3657 found in genrsa is now in app_rand.c and is used by all programs
3658 that support '-rand'.
3661 *) In RAND_write_file, use mode 0600 for creating files;
3662 don't just chmod when it may be too late.
3665 *) Report an error from X509_STORE_load_locations
3666 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3669 *) New function ASN1_mbstring_copy() this copies a string in either
3670 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3671 into an ASN1_STRING type. A mask of permissible types is passed
3672 and it chooses the "minimal" type to use or an error if not type
3676 *) Add function equivalents to the various macros in asn1.h. The old
3677 macros are retained with an M_ prefix. Code inside the library can
3678 use the M_ macros. External code (including the openssl utility)
3679 should *NOT* in order to be "shared library friendly".
3682 *) Add various functions that can check a certificate's extensions
3683 to see if it usable for various purposes such as SSL client,
3684 server or S/MIME and CAs of these types. This is currently
3685 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3686 verification. Also added a -purpose flag to x509 utility to
3687 print out all the purposes.
3690 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3694 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3695 for, obtain and decode and extension and obtain its critical flag.
3696 This allows all the necessary extension code to be handled in a
3697 single function call.
3700 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3701 platforms. See crypto/rc4/rc4_enc.c for further details.
3704 *) New -noout option to asn1parse. This causes no output to be produced
3705 its main use is when combined with -strparse and -out to extract data
3706 from a file (which may not be in ASN.1 format).
3709 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3710 when producing the local key id.
3711 [Richard Levitte <levitte@stacken.kth.se>]
3713 *) New option -dhparam in s_server. This allows a DH parameter file to be
3714 stated explicitly. If it is not stated then it tries the first server
3715 certificate file. The previous behaviour hard coded the filename
3719 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3720 a public key to be input or output. For example:
3721 openssl rsa -in key.pem -pubout -out pubkey.pem
3722 Also added necessary DSA public key functions to handle this.
3725 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3726 in the message. This was handled by allowing
3727 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3728 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3730 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3731 to the end of the strings whereas this didn't. This would cause problems
3732 if strings read with d2i_ASN1_bytes() were later modified.
3733 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3735 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3736 data and it contains EOF it will end up returning an error. This is
3737 caused by input 46 bytes long. The cause is due to the way base64
3738 BIOs find the start of base64 encoded data. They do this by trying a
3739 trial decode on each line until they find one that works. When they
3740 do a flag is set and it starts again knowing it can pass all the
3741 data directly through the decoder. Unfortunately it doesn't reset
3742 the context it uses. This means that if EOF is reached an attempt
3743 is made to pass two EOFs through the context and this causes the
3744 resulting error. This can also cause other problems as well. As is
3745 usual with these problems it takes *ages* to find and the fix is
3746 trivial: move one line.
3747 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
3749 *) Ugly workaround to get s_client and s_server working under Windows. The
3750 old code wouldn't work because it needed to select() on sockets and the
3751 tty (for keypresses and to see if data could be written). Win32 only
3752 supports select() on sockets so we select() with a 1s timeout on the
3753 sockets and then see if any characters are waiting to be read, if none
3754 are present then we retry, we also assume we can always write data to
3755 the tty. This isn't nice because the code then blocks until we've
3756 received a complete line of data and it is effectively polling the
3757 keyboard at 1s intervals: however it's quite a bit better than not
3758 working at all :-) A dedicated Windows application might handle this
3759 with an event loop for example.
3762 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
3763 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
3764 will be called when RSA_sign() and RSA_verify() are used. This is useful
3765 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
3766 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
3767 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
3768 This necessitated the support of an extra signature type NID_md5_sha1
3769 for SSL signatures and modifications to the SSL library to use it instead
3770 of calling RSA_public_decrypt() and RSA_private_encrypt().
3773 *) Add new -verify -CAfile and -CApath options to the crl program, these
3774 will lookup a CRL issuers certificate and verify the signature in a
3775 similar way to the verify program. Tidy up the crl program so it
3776 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
3777 less strict. It will now permit CRL extensions even if it is not
3778 a V2 CRL: this will allow it to tolerate some broken CRLs.
3781 *) Initialize all non-automatic variables each time one of the openssl
3782 sub-programs is started (this is necessary as they may be started
3783 multiple times from the "OpenSSL>" prompt).
3784 [Lennart Bang, Bodo Moeller]
3786 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
3787 removing all other RSA functionality (this is what NO_RSA does). This
3788 is so (for example) those in the US can disable those operations covered
3789 by the RSA patent while allowing storage and parsing of RSA keys and RSA
3793 *) Non-copying interface to BIO pairs.
3794 (still largely untested)
3797 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
3798 ASCII string. This was handled independently in various places before.
3801 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
3802 UTF8 strings a character at a time.
3805 *) Use client_version from client hello to select the protocol
3806 (s23_srvr.c) and for RSA client key exchange verification
3807 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
3810 *) Add various utility functions to handle SPKACs, these were previously
3811 handled by poking round in the structure internals. Added new function
3812 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
3813 print, verify and generate SPKACs. Based on an original idea from
3814 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
3817 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
3820 *) Allow the config file extension section to be overwritten on the
3821 command line. Based on an original idea from Massimiliano Pala
3822 <madwolf@comune.modena.it>. The new option is called -extensions
3823 and can be applied to ca, req and x509. Also -reqexts to override
3824 the request extensions in req and -crlexts to override the crl extensions
3828 *) Add new feature to the SPKAC handling in ca. Now you can include
3829 the same field multiple times by preceding it by "XXXX." for example:
3832 this is the same syntax as used in the req config file.
3835 *) Allow certificate extensions to be added to certificate requests. These
3836 are specified in a 'req_extensions' option of the req section of the
3837 config file. They can be printed out with the -text option to req but
3838 are otherwise ignored at present.
3841 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
3842 data read consists of only the final block it would not decrypted because
3843 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
3844 A misplaced 'break' also meant the decrypted final block might not be
3845 copied until the next read.
3848 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
3849 a few extra parameters to the DH structure: these will be useful if
3850 for example we want the value of 'q' or implement X9.42 DH.
3853 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
3854 provides hooks that allow the default DSA functions or functions on a
3855 "per key" basis to be replaced. This allows hardware acceleration and
3856 hardware key storage to be handled without major modification to the
3857 library. Also added low level modexp hooks and CRYPTO_EX structure and
3858 associated functions.
3861 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
3862 as "read only": it can't be written to and the buffer it points to will
3863 not be freed. Reading from a read only BIO is much more efficient than
3864 a normal memory BIO. This was added because there are several times when
3865 an area of memory needs to be read from a BIO. The previous method was
3866 to create a memory BIO and write the data to it, this results in two
3867 copies of the data and an O(n^2) reading algorithm. There is a new
3868 function BIO_new_mem_buf() which creates a read only memory BIO from
3869 an area of memory. Also modified the PKCS#7 routines to use read only
3873 *) Bugfix: ssl23_get_client_hello did not work properly when called in
3874 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
3875 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
3876 but a retry condition occured while trying to read the rest.
3879 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
3880 NID_pkcs7_encrypted by default: this was wrong since this should almost
3881 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
3882 the encrypted data type: this is a more sensible place to put it and it
3883 allows the PKCS#12 code to be tidied up that duplicated this
3887 *) Changed obj_dat.pl script so it takes its input and output files on
3888 the command line. This should avoid shell escape redirection problems
3892 *) Initial support for certificate extension requests, these are included
3893 in things like Xenroll certificate requests. Included functions to allow
3894 extensions to be obtained and added.
3897 *) -crlf option to s_client and s_server for sending newlines as
3898 CRLF (as required by many protocols).
3901 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
3903 *) Install libRSAglue.a when OpenSSL is built with RSAref.
3904 [Ralf S. Engelschall]
3906 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
3907 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
3909 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
3913 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
3914 DH parameters/keys (q is lost during that conversion, but the resulting
3915 DH parameters contain its length).
3917 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
3918 much faster than DH_generate_parameters (which creates parameters
3919 where p = 2*q + 1), and also the smaller q makes DH computations
3920 much more efficient (160-bit exponentiation instead of 1024-bit
3921 exponentiation); so this provides a convenient way to support DHE
3922 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
3923 utter importance to use
3924 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3926 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
3927 when such DH parameters are used, because otherwise small subgroup
3928 attacks may become possible!
3931 *) Avoid memory leak in i2d_DHparams.
3934 *) Allow the -k option to be used more than once in the enc program:
3935 this allows the same encrypted message to be read by multiple recipients.
3938 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
3939 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
3940 it will always use the numerical form of the OID, even if it has a short
3944 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
3945 method only got called if p,q,dmp1,dmq1,iqmp components were present,
3946 otherwise bn_mod_exp was called. In the case of hardware keys for example
3947 no private key components need be present and it might store extra data
3948 in the RSA structure, which cannot be accessed from bn_mod_exp.
3949 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
3950 private key operations.
3953 *) Added support for SPARC Linux.
3956 *) pem_password_cb function type incompatibly changed from
3957 typedef int pem_password_cb(char *buf, int size, int rwflag);
3959 ....(char *buf, int size, int rwflag, void *userdata);
3960 so that applications can pass data to their callbacks:
3961 The PEM[_ASN1]_{read,write}... functions and macros now take an
3962 additional void * argument, which is just handed through whenever
3963 the password callback is called.
3964 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
3966 New function SSL_CTX_set_default_passwd_cb_userdata.
3968 Compatibility note: As many C implementations push function arguments
3969 onto the stack in reverse order, the new library version is likely to
3970 interoperate with programs that have been compiled with the old
3971 pem_password_cb definition (PEM_whatever takes some data that
3972 happens to be on the stack as its last argument, and the callback
3973 just ignores this garbage); but there is no guarantee whatsoever that
3976 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
3977 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
3978 problems not only on Windows, but also on some Unix platforms.
3979 To avoid problematic command lines, these definitions are now in an
3980 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
3981 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
3984 *) MIPS III/IV assembler module is reimplemented.
3987 *) More DES library cleanups: remove references to srand/rand and
3988 delete an unused file.
3991 *) Add support for the the free Netwide assembler (NASM) under Win32,
3992 since not many people have MASM (ml) and it can be hard to obtain.
3993 This is currently experimental but it seems to work OK and pass all
3994 the tests. Check out INSTALL.W32 for info.
3997 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
3998 without temporary keys kept an extra copy of the server key,
3999 and connections with temporary keys did not free everything in case
4003 *) New function RSA_check_key and new openssl rsa option -check
4004 for verifying the consistency of RSA keys.
4005 [Ulf Moeller, Bodo Moeller]
4007 *) Various changes to make Win32 compile work:
4008 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4009 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4010 comparison" warnings.
4011 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4014 *) Add a debugging option to PKCS#5 v2 key generation function: when
4015 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4016 derived keys are printed to stderr.
4019 *) Copy the flags in ASN1_STRING_dup().
4020 [Roman E. Pavlov <pre@mo.msk.ru>]
4022 *) The x509 application mishandled signing requests containing DSA
4023 keys when the signing key was also DSA and the parameters didn't match.
4025 It was supposed to omit the parameters when they matched the signing key:
4026 the verifying software was then supposed to automatically use the CA's
4027 parameters if they were absent from the end user certificate.
4029 Omitting parameters is no longer recommended. The test was also
4030 the wrong way round! This was probably due to unusual behaviour in
4031 EVP_cmp_parameters() which returns 1 if the parameters match.
4032 This meant that parameters were omitted when they *didn't* match and
4033 the certificate was useless. Certificates signed with 'ca' didn't have
4035 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4037 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4038 The interface is as follows:
4039 Applications can use
4040 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4041 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4042 "off" is now the default.
4043 The library internally uses
4044 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4045 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4046 to disable memory-checking temporarily.
4048 Some inconsistent states that previously were possible (and were
4049 even the default) are now avoided.
4051 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4052 with each memory chunk allocated; this is occasionally more helpful
4053 than just having a counter.
4055 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4057 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4061 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4062 which largely parallels "options", but is for changing API behaviour,
4063 whereas "options" are about protocol behaviour.
4064 Initial "mode" flags are:
4066 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4067 a single record has been written.
4068 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4069 retries use the same buffer location.
4070 (But all of the contents must be
4074 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4077 *) Fix problems with no-hmac etc.
4078 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4080 *) New functions RSA_get_default_method(), RSA_set_method() and
4081 RSA_get_method(). These allows replacement of RSA_METHODs without having
4082 to mess around with the internals of an RSA structure.
4085 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4086 Also really enable memory leak checks in openssl.c and in some
4088 [Chad C. Mulligan, Bodo Moeller]
4090 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4091 up the length of negative integers. This has now been simplified to just
4092 store the length when it is first determined and use it later, rather
4093 than trying to keep track of where data is copied and updating it to
4095 [Steve Henson, reported by Brien Wheeler
4096 <bwheeler@authentica-security.com>]
4098 *) Add a new function PKCS7_signatureVerify. This allows the verification
4099 of a PKCS#7 signature but with the signing certificate passed to the
4100 function itself. This contrasts with PKCS7_dataVerify which assumes the
4101 certificate is present in the PKCS#7 structure. This isn't always the
4102 case: certificates can be omitted from a PKCS#7 structure and be
4103 distributed by "out of band" means (such as a certificate database).
4106 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4107 function prototypes in pem.h, also change util/mkdef.pl to add the
4108 necessary function names.
4111 *) mk1mf.pl (used by Windows builds) did not properly read the
4112 options set by Configure in the top level Makefile, and Configure
4113 was not even able to write more than one option correctly.
4114 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4117 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4118 file to be loaded from a BIO or FILE pointer. The BIO version will
4119 for example allow memory BIOs to contain config info.
4122 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4123 Whoever hopes to achieve shared-library compatibility across versions
4124 must use this, not the compile-time macro.
4125 (Exercise 0.9.4: Which is the minimum library version required by
4127 Note: All this applies only to multi-threaded programs, others don't
4131 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4132 through a BIO pair triggered the default case, i.e.
4133 SSLerr(...,SSL_R_UNKNOWN_STATE).
4136 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4137 can use the SSL library even if none of the specific BIOs is
4141 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4142 for the encoded length.
4143 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4145 *) Add initial documentation of the X509V3 functions.
4148 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4149 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4150 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4151 secure PKCS#8 private key format with a high iteration count.
4154 *) Fix determination of Perl interpreter: A perl or perl5
4155 _directory_ in $PATH was also accepted as the interpreter.
4156 [Ralf S. Engelschall]
4158 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4159 wrong with it but it was very old and did things like calling
4160 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4164 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4165 to use the new extension code.
4168 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4169 with macros. This should make it easier to change their form, add extra
4170 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4174 *) Add to configuration table a new entry that can specify an alternative
4175 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4176 according to Mark Crispin <MRC@Panda.COM>.
4180 *) DES CBC did not update the IV. Weird.
4183 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4184 Changing the behaviour of the former might break existing programs --
4185 where IV updating is needed, des_ncbc_encrypt can be used.
4188 *) When bntest is run from "make test" it drives bc to check its
4189 calculations, as well as internally checking them. If an internal check
4190 fails, it needs to cause bc to give a non-zero result or make test carries
4191 on without noticing the failure. Fixed.
4194 *) DES library cleanups.
4197 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4198 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4199 ciphers. NOTE: although the key derivation function has been verified
4200 against some published test vectors it has not been extensively tested
4201 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4205 *) Instead of "mkdir -p", which is not fully portable, use new
4206 Perl script "util/mkdir-p.pl".
4209 *) Rewrite the way password based encryption (PBE) is handled. It used to
4210 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4211 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4212 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4213 the 'parameter' field of the AlgorithmIdentifier is passed to the
4214 underlying key generation function so it must do its own ASN1 parsing.
4215 This has also changed the EVP_PBE_CipherInit() function which now has a
4216 'parameter' argument instead of literal salt and iteration count values
4217 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4220 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4221 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4222 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4223 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4224 value was just used as a "magic string" and not used directly its
4225 value doesn't matter.
4228 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4232 *) "linux-sparc64" configuration (ultrapenguin).
4233 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4234 "linux-sparc" configuration.
4235 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4237 *) config now generates no-xxx options for missing ciphers.
4240 *) Support the EBCDIC character set (work in progress).
4241 File ebcdic.c not yet included because it has a different license.
4242 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4244 *) Support BS2000/OSD-POSIX.
4245 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4247 *) Make callbacks for key generation use void * instead of char *.
4250 *) Make S/MIME samples compile (not yet tested).
4253 *) Additional typesafe stacks.
4256 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4260 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4262 *) New configuration variant "sco5-gcc".
4264 *) Updated some demos.
4265 [Sean O Riordain, Wade Scholine]
4267 *) Add missing BIO_free at exit of pkcs12 application.
4270 *) Fix memory leak in conf.c.
4273 *) Updates for Win32 to assembler version of MD5.
4276 *) Set #! path to perl in apps/der_chop to where we found it
4277 instead of using a fixed path.
4280 *) SHA library changes for irix64-mips4-cc.
4283 *) Improvements for VMS support.
4287 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4289 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4290 This also avoids the problems with SC4.2 and unpatched SC5.
4291 [Andy Polyakov <appro@fy.chalmers.se>]
4293 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4294 These are required because of the typesafe stack would otherwise break
4295 existing code. If old code used a structure member which used to be STACK
4296 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4297 sk_num or sk_value it would produce an error because the num, data members
4298 are not present in STACK_OF. Now it just produces a warning. sk_set
4299 replaces the old method of assigning a value to sk_value
4300 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4301 that does this will no longer work (and should use sk_set instead) but
4302 this could be regarded as a "questionable" behaviour anyway.
4305 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4306 correctly handle encrypted S/MIME data.
4309 *) Change type of various DES function arguments from des_cblock
4310 (which means, in function argument declarations, pointer to char)
4311 to des_cblock * (meaning pointer to array with 8 char elements),
4312 which allows the compiler to do more typechecking; it was like
4313 that back in SSLeay, but with lots of ugly casts.
4315 Introduce new type const_des_cblock.
4318 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4319 problems: find RecipientInfo structure that matches recipient certificate
4320 and initialise the ASN1 structures properly based on passed cipher.
4323 *) Belatedly make the BN tests actually check the results.
4326 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4327 to and from BNs: it was completely broken. New compilation option
4328 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4329 key elements as negative integers.
4332 *) Reorganize and speed up MD5.
4333 [Andy Polyakov <appro@fy.chalmers.se>]
4336 [Richard Levitte <richard@levitte.org>]
4338 *) New option -out to asn1parse to allow the parsed structure to be
4339 output to a file. This is most useful when combined with the -strparse
4340 option to examine the output of things like OCTET STRINGS.
4343 *) Make SSL library a little more fool-proof by not requiring any longer
4344 that SSL_set_{accept,connect}_state be called before
4345 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4346 in many applications because usually everything *appeared* to work as
4347 intended anyway -- now it really works as intended).
4350 *) Move openssl.cnf out of lib/.
4353 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4354 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4355 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4356 [Ralf S. Engelschall]
4358 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4359 handle PKCS#7 enveloped data properly.
4360 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4362 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4363 copying pointers. The cert_st handling is changed by this in
4364 various ways (and thus what used to be known as ctx->default_cert
4365 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4366 any longer when s->cert does not give us what we need).
4367 ssl_cert_instantiate becomes obsolete by this change.
4368 As soon as we've got the new code right (possibly it already is?),
4369 we have solved a couple of bugs of the earlier code where s->cert
4370 was used as if it could not have been shared with other SSL structures.
4372 Note that using the SSL API in certain dirty ways now will result
4373 in different behaviour than observed with earlier library versions:
4374 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4375 does not influence s as it used to.
4377 In order to clean up things more thoroughly, inside SSL_SESSION
4378 we don't use CERT any longer, but a new structure SESS_CERT
4379 that holds per-session data (if available); currently, this is
4380 the peer's certificate chain and, for clients, the server's certificate
4381 and temporary key. CERT holds only those values that can have
4382 meaningful defaults in an SSL_CTX.
4385 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4386 from the internal representation. Various PKCS#7 fixes: remove some
4387 evil casts and set the enc_dig_alg field properly based on the signing
4391 *) Allow PKCS#12 password to be set from the command line or the
4392 environment. Let 'ca' get its config file name from the environment
4393 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4397 *) Allow certificate policies extension to use an IA5STRING for the
4398 organization field. This is contrary to the PKIX definition but
4399 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4403 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4404 without disallowing inline assembler and the like for non-pedantic builds.
4407 *) Support Borland C++ builder.
4408 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4413 *) SHA-1 cleanups and performance enhancements.
4414 [Andy Polyakov <appro@fy.chalmers.se>]
4416 *) Sparc v8plus assembler for the bignum library.
4417 [Andy Polyakov <appro@fy.chalmers.se>]
4419 *) Accept any -xxx and +xxx compiler options in Configure.
4422 *) Update HPUX configuration.
4425 *) Add missing sk_<type>_unshift() function to safestack.h
4426 [Ralf S. Engelschall]
4428 *) New function SSL_CTX_use_certificate_chain_file that sets the
4429 "extra_cert"s in addition to the certificate. (This makes sense
4430 only for "PEM" format files, as chains as a whole are not
4434 *) Support verify_depth from the SSL API.
4435 x509_vfy.c had what can be considered an off-by-one-error:
4436 Its depth (which was not part of the external interface)
4437 was actually counting the number of certificates in a chain;
4438 now it really counts the depth.
4441 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4442 instead of X509err, which often resulted in confusing error
4443 messages since the error codes are not globally unique
4444 (e.g. an alleged error in ssl3_accept when a certificate
4445 didn't match the private key).
4447 *) New function SSL_CTX_set_session_id_context that allows to set a default
4448 value (so that you don't need SSL_set_session_id_context for each
4449 connection using the SSL_CTX).
4452 *) OAEP decoding bug fix.
4455 *) Support INSTALL_PREFIX for package builders, as proposed by
4459 *) New Configure options "threads" and "no-threads". For systems
4460 where the proper compiler options are known (currently Solaris
4461 and Linux), "threads" is the default.
4464 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4467 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4468 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4469 such as /usr/local/bin.
4472 *) "make linux-shared" to build shared libraries.
4473 [Niels Poppe <niels@netbox.org>]
4475 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4478 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4479 extension adding in x509 utility.
4482 *) Remove NOPROTO sections and error code comments.
4485 *) Partial rewrite of the DEF file generator to now parse the ANSI
4489 *) New Configure options --prefix=DIR and --openssldir=DIR.
4492 *) Complete rewrite of the error code script(s). It is all now handled
4493 by one script at the top level which handles error code gathering,
4494 header rewriting and C source file generation. It should be much better
4495 than the old method: it now uses a modified version of Ulf's parser to
4496 read the ANSI prototypes in all header files (thus the old K&R definitions
4497 aren't needed for error creation any more) and do a better job of
4498 translating function codes into names. The old 'ASN1 error code imbedded
4499 in a comment' is no longer necessary and it doesn't use .err files which
4500 have now been deleted. Also the error code call doesn't have to appear all
4501 on one line (which resulted in some large lines...).
4504 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4507 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4508 0 (which usually indicates a closed connection), but continue reading.
4511 *) Fix some race conditions.
4514 *) Add support for CRL distribution points extension. Add Certificate
4515 Policies and CRL distribution points documentation.
4518 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4521 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4522 8 of keying material. Merlin has also confirmed interop with this fix
4523 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4524 [Merlin Hughes <merlin@baltimore.ie>]
4526 *) Fix lots of warnings.
4527 [Richard Levitte <levitte@stacken.kth.se>]
4529 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4530 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4531 [Richard Levitte <levitte@stacken.kth.se>]
4533 *) Fix problems with sizeof(long) == 8.
4534 [Andy Polyakov <appro@fy.chalmers.se>]
4536 *) Change functions to ANSI C.
4539 *) Fix typos in error codes.
4540 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4542 *) Remove defunct assembler files from Configure.
4545 *) SPARC v8 assembler BIGNUM implementation.
4546 [Andy Polyakov <appro@fy.chalmers.se>]
4548 *) Support for Certificate Policies extension: both print and set.
4549 Various additions to support the r2i method this uses.
4552 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4553 return a const string when you are expecting an allocated buffer.
4556 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4557 types DirectoryString and DisplayText.
4560 *) Add code to allow r2i extensions to access the configuration database,
4561 add an LHASH database driver and add several ctx helper functions.
4564 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4565 fail when they extended the size of a BIGNUM.
4568 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4569 support typesafe stack.
4572 *) Fix typo in SSL_[gs]et_options().
4573 [Nils Frostberg <nils@medcom.se>]
4575 *) Delete various functions and files that belonged to the (now obsolete)
4576 old X509V3 handling code.
4579 *) New Configure option "rsaref".
4582 *) Don't auto-generate pem.h.
4585 *) Introduce type-safe ASN.1 SETs.
4588 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4589 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4591 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4592 that links with OpenSSL (well at least cause lots of warnings), but fear
4593 not: the conversion is trivial, and it eliminates loads of evil casts. A
4594 few STACKed things have been converted already. Feel free to convert more.
4595 In the fullness of time, I'll do away with the STACK type altogether.
4598 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4599 specified in <certfile> by updating the entry in the index.txt file.
4600 This way one no longer has to edit the index.txt file manually for
4601 revoking a certificate. The -revoke option does the gory details now.
4602 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4604 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4605 `-text' option at all and this way the `-noout -text' combination was
4606 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4607 [Ralf S. Engelschall]
4609 *) Make sure a corresponding plain text error message exists for the
4610 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4611 verify callback function determined that a certificate was revoked.
4612 [Ralf S. Engelschall]
4614 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4615 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4616 all available cipers including rc5, which was forgotten until now.
4617 In order to let the testing shell script know which algorithms
4618 are available, a new (up to now undocumented) command
4619 "openssl list-cipher-commands" is used.
4622 *) Bugfix: s_client occasionally would sleep in select() when
4623 it should have checked SSL_pending() first.
4626 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4627 the raw DSA values prior to ASN.1 encoding.
4630 *) Tweaks to Configure
4631 [Niels Poppe <niels@netbox.org>]
4633 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4637 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4640 *) New config option to avoid instructions that are illegal on the 80386.
4641 The default code is faster, but requires at least a 486.
4644 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4645 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4646 same as SSL2_VERSION anyway.
4649 *) New "-showcerts" option for s_client.
4652 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4653 application. Various cleanups and fixes.
4656 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4657 modify error routines to work internally. Add error codes and PBE init
4658 to library startup routines.
4661 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4662 packing functions to asn1 and evp. Changed function names and error
4663 codes along the way.
4666 *) PKCS12 integration: and so it begins... First of several patches to
4667 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4668 objects to objects.h
4671 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4672 and display support for Thawte strong extranet extension.
4675 *) Add LinuxPPC support.
4676 [Jeff Dubrule <igor@pobox.org>]
4678 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4679 bn_div_words in alpha.s.
4680 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4682 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4683 OAEP isn't supported when OpenSSL is built with RSAref.
4684 [Ulf Moeller <ulf@fitug.de>]
4686 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4687 so they no longer are missing under -DNOPROTO.
4688 [Soren S. Jorvang <soren@t.dk>]
4691 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4693 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4694 doesn't work when the session is reused. Coming soon!
4697 *) Fix a security hole, that allows sessions to be reused in the wrong
4698 context thus bypassing client cert protection! All software that uses
4699 client certs and session caches in multiple contexts NEEDS PATCHING to
4700 allow session reuse! A fuller solution is in the works.
4701 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4703 *) Some more source tree cleanups (removed obsolete files
4704 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4705 permission on "config" script to be executable) and a fix for the INSTALL
4707 [Ulf Moeller <ulf@fitug.de>]
4709 *) Remove some legacy and erroneous uses of malloc, free instead of
4711 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4713 *) Make rsa_oaep_test return non-zero on error.
4714 [Ulf Moeller <ulf@fitug.de>]
4716 *) Add support for native Solaris shared libraries. Configure
4717 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4718 if someone would make that last step automatic.
4719 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4721 *) ctx_size was not built with the right compiler during "make links". Fixed.
4724 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4725 except NULL ciphers". This means the default cipher list will no longer
4726 enable NULL ciphers. They need to be specifically enabled e.g. with
4727 the string "DEFAULT:eNULL".
4730 *) Fix to RSA private encryption routines: if p < q then it would
4731 occasionally produce an invalid result. This will only happen with
4732 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4735 *) Be less restrictive and allow also `perl util/perlpath.pl
4736 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4737 because this way one can also use an interpreter named `perl5' (which is
4738 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4739 installed as `perl').
4740 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4742 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
4743 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
4745 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
4746 advapi32.lib to Win32 build and change the pem test comparision
4747 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
4748 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
4749 and crypto/des/ede_cbcm_enc.c.
4752 *) DES quad checksum was broken on big-endian architectures. Fixed.
4755 *) Comment out two functions in bio.h that aren't implemented. Fix up the
4756 Win32 test batch file so it (might) work again. The Win32 test batch file
4757 is horrible: I feel ill....
4760 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
4761 in e_os.h. Audit of header files to check ANSI and non ANSI
4762 sections: 10 functions were absent from non ANSI section and not exported
4763 from Windows DLLs. Fixed up libeay.num for new functions.
4766 *) Make `openssl version' output lines consistent.
4767 [Ralf S. Engelschall]
4769 *) Fix Win32 symbol export lists for BIO functions: Added
4770 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
4771 to ms/libeay{16,32}.def.
4772 [Ralf S. Engelschall]
4774 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
4775 fine under Unix and passes some trivial tests I've now added. But the
4776 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
4777 added to make sure no one expects that this stuff really works in the
4778 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
4779 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
4781 [Ralf S. Engelschall]
4783 *) Fix the generation of two part addresses in perl.
4784 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
4786 *) Add config entry for Linux on MIPS.
4787 [John Tobey <jtobey@channel1.com>]
4789 *) Make links whenever Configure is run, unless we are on Windoze.
4792 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
4793 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
4797 *) Add a useful kludge to allow package maintainers to specify compiler and
4798 other platforms details on the command line without having to patch the
4799 Configure script everytime: One now can use ``perl Configure
4800 <id>:<details>'', i.e. platform ids are allowed to have details appended
4801 to them (seperated by colons). This is treated as there would be a static
4802 pre-configured entry in Configure's %table under key <id> with value
4803 <details> and ``perl Configure <id>'' is called. So, when you want to
4804 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
4805 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
4806 now, which overrides the FreeBSD-elf entry on-the-fly.
4807 [Ralf S. Engelschall]
4809 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
4812 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
4813 on the `perl Configure ...' command line. This way one can compile
4814 OpenSSL libraries with Position Independent Code (PIC) which is needed
4815 for linking it into DSOs.
4816 [Ralf S. Engelschall]
4818 *) Remarkably, export ciphers were totally broken and no-one had noticed!
4822 *) Cleaned up the LICENSE document: The official contact for any license
4823 questions now is the OpenSSL core team under openssl-core@openssl.org.
4824 And add a paragraph about the dual-license situation to make sure people
4825 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
4826 to the OpenSSL toolkit.
4827 [Ralf S. Engelschall]
4829 *) General source tree makefile cleanups: Made `making xxx in yyy...'
4830 display consistent in the source tree and replaced `/bin/rm' by `rm'.
4831 Additonally cleaned up the `make links' target: Remove unnecessary
4832 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
4833 to speed processing and no longer clutter the display with confusing
4834 stuff. Instead only the actually done links are displayed.
4835 [Ralf S. Engelschall]
4837 *) Permit null encryption ciphersuites, used for authentication only. It used
4838 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
4839 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
4843 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
4844 signed attributes when verifying signatures (this would break them),
4845 the detached data encoding was wrong and public keys obtained using
4846 X509_get_pubkey() weren't freed.
4849 *) Add text documentation for the BUFFER functions. Also added a work around
4850 to a Win95 console bug. This was triggered by the password read stuff: the
4851 last character typed gets carried over to the next fread(). If you were
4852 generating a new cert request using 'req' for example then the last
4853 character of the passphrase would be CR which would then enter the first
4857 *) Added the new `Includes OpenSSL Cryptography Software' button as
4858 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
4859 button and can be used by applications based on OpenSSL to show the
4860 relationship to the OpenSSL project.
4861 [Ralf S. Engelschall]
4863 *) Remove confusing variables in function signatures in files
4864 ssl/ssl_lib.c and ssl/ssl.h.
4865 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4867 *) Don't install bss_file.c under PREFIX/include/
4868 [Lennart Bong <lob@kulthea.stacken.kth.se>]
4870 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
4871 functions that return function pointers and has support for NT specific
4872 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
4873 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
4874 unsigned to signed types: this was killing the Win32 compile.
4877 *) Add new certificate file to stack functions,
4878 SSL_add_dir_cert_subjects_to_stack() and
4879 SSL_add_file_cert_subjects_to_stack(). These largely supplant
4880 SSL_load_client_CA_file(), and can be used to add multiple certs easily
4881 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
4882 This means that Apache-SSL and similar packages don't have to mess around
4883 to add as many CAs as they want to the preferred list.
4886 *) Experiment with doxygen documentation. Currently only partially applied to
4888 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
4889 openssl.doxy as the configuration file.
4892 *) Get rid of remaining C++-style comments which strict C compilers hate.
4893 [Ralf S. Engelschall, pointed out by Carlos Amengual]
4895 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
4896 compiled in by default: it has problems with large keys.
4899 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
4900 DH private keys and/or callback functions which directly correspond to
4901 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
4902 is needed for applications which have to configure certificates on a
4903 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
4905 For the RSA certificate situation is makes no difference, but
4906 for the DSA certificate situation this fixes the "no shared cipher"
4907 problem where the OpenSSL cipher selection procedure failed because the
4908 temporary keys were not overtaken from the context and the API provided
4909 no way to reconfigure them.
4910 The new functions now let applications reconfigure the stuff and they
4911 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
4912 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
4913 non-public-API function ssl_cert_instantiate() is used as a helper
4914 function and also to reduce code redundancy inside ssl_rsa.c.
4915 [Ralf S. Engelschall]
4917 *) Move s_server -dcert and -dkey options out of the undocumented feature
4918 area because they are useful for the DSA situation and should be
4919 recognized by the users.
4920 [Ralf S. Engelschall]
4922 *) Fix the cipher decision scheme for export ciphers: the export bits are
4923 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
4924 SSL_EXP_MASK. So, the original variable has to be used instead of the
4925 already masked variable.
4926 [Richard Levitte <levitte@stacken.kth.se>]
4928 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
4929 [Richard Levitte <levitte@stacken.kth.se>]
4931 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
4932 from `int' to `unsigned int' because it's a length and initialized by
4933 EVP_DigestFinal() which expects an `unsigned int *'.
4934 [Richard Levitte <levitte@stacken.kth.se>]
4936 *) Don't hard-code path to Perl interpreter on shebang line of Configure
4937 script. Instead use the usual Shell->Perl transition trick.
4938 [Ralf S. Engelschall]
4940 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
4941 (in addition to RSA certificates) to match the behaviour of `openssl dsa
4942 -noout -modulus' as it's already the case for `openssl rsa -noout
4943 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
4944 currently the public key is printed (a decision which was already done by
4945 `openssl dsa -modulus' in the past) which serves a similar purpose.
4946 Additionally the NO_RSA no longer completely removes the whole -modulus
4947 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
4949 [Ralf S. Engelschall]
4951 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
4952 BIO. See the source (crypto/evp/bio_ok.c) for more info.
4953 [Arne Ansper <arne@ats.cyber.ee>]
4955 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
4956 to be added. Now both 'req' and 'ca' can use new objects defined in the
4960 *) Add cool BIO that does syslog (or event log on NT).
4961 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
4963 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
4964 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
4965 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
4966 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
4969 *) Add preliminary config info for new extension code.
4972 *) Make RSA_NO_PADDING really use no padding.
4973 [Ulf Moeller <ulf@fitug.de>]
4975 *) Generate errors when private/public key check is done.
4978 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
4979 for some CRL extensions and new objects added.
4982 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
4983 key usage extension and fuller support for authority key id.
4986 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
4987 padding method for RSA, which is recommended for new applications in PKCS
4988 #1 v2.0 (RFC 2437, October 1998).
4989 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
4990 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
4991 against Bleichbacher's attack on RSA.
4992 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
4995 *) Updates to the new SSL compression code
4996 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
4998 *) Fix so that the version number in the master secret, when passed
4999 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5000 (because the server will not accept higher), that the version number
5001 is 0x03,0x01, not 0x03,0x00
5002 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5004 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5005 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5006 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5009 *) Support for RAW extensions where an arbitrary extension can be
5010 created by including its DER encoding. See apps/openssl.cnf for
5014 *) Make sure latest Perl versions don't interpret some generated C array
5015 code as Perl array code in the crypto/err/err_genc.pl script.
5016 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5018 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5019 not many people have the assembler. Various Win32 compilation fixes and
5020 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5024 *) Modify configure script 'Configure' to automatically create crypto/date.h
5025 file under Win32 and also build pem.h from pem.org. New script
5026 util/mkfiles.pl to create the MINFO file on environments that can't do a
5027 'make files': perl util/mkfiles.pl >MINFO should work.
5030 *) Major rework of DES function declarations, in the pursuit of correctness
5031 and purity. As a result, many evil casts evaporated, and some weirdness,
5032 too. You may find this causes warnings in your code. Zapping your evil
5033 casts will probably fix them. Mostly.
5036 *) Fix for a typo in asn1.h. Bug fix to object creation script
5037 obj_dat.pl. It considered a zero in an object definition to mean
5038 "end of object": none of the objects in objects.h have any zeros
5039 so it wasn't spotted.
5040 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5042 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5043 Masking (CBCM). In the absence of test vectors, the best I have been able
5044 to do is check that the decrypt undoes the encrypt, so far. Send me test
5045 vectors if you have them.
5048 *) Correct calculation of key length for export ciphers (too much space was
5049 allocated for null ciphers). This has not been tested!
5052 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5053 message is now correct (it understands "crypto" and "ssl" on its
5054 command line). There is also now an "update" option. This will update
5055 the util/ssleay.num and util/libeay.num files with any new functions.
5057 perl util/mkdef.pl crypto ssl update
5058 it will update them.
5061 *) Overhauled the Perl interface (perl/*):
5062 - ported BN stuff to OpenSSL's different BN library
5063 - made the perl/ source tree CVS-aware
5064 - renamed the package from SSLeay to OpenSSL (the files still contain
5065 their history because I've copied them in the repository)
5066 - removed obsolete files (the test scripts will be replaced
5067 by better Test::Harness variants in the future)
5068 [Ralf S. Engelschall]
5070 *) First cut for a very conservative source tree cleanup:
5071 1. merge various obsolete readme texts into doc/ssleay.txt
5072 where we collect the old documents and readme texts.
5073 2. remove the first part of files where I'm already sure that we no
5074 longer need them because of three reasons: either they are just temporary
5075 files which were left by Eric or they are preserved original files where
5076 I've verified that the diff is also available in the CVS via "cvs diff
5077 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5078 the crypto/md/ stuff).
5079 [Ralf S. Engelschall]
5081 *) More extension code. Incomplete support for subject and issuer alt
5082 name, issuer and authority key id. Change the i2v function parameters
5083 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5084 what that's for :-) Fix to ASN1 macro which messed up
5085 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5088 *) Preliminary support for ENUMERATED type. This is largely copied from the
5092 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5093 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5095 *) Make sure `make rehash' target really finds the `openssl' program.
5096 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5098 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5099 like to hear about it if this slows down other processors.
5102 *) Add CygWin32 platform information to Configure script.
5103 [Alan Batie <batie@aahz.jf.intel.com>]
5105 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5106 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5108 *) New program nseq to manipulate netscape certificate sequences
5111 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5115 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5116 but the BN code had some problems that would cause failures when
5117 doing certificate verification and some other functions.
5118 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5120 *) Add ASN1 and PEM code to support netscape certificate sequences.
5123 *) Add ASN1 and PEM code to support netscape certificate sequences.
5126 *) Add several PKIX and private extended key usage OIDs.
5129 *) Modify the 'ca' program to handle the new extension code. Modify
5130 openssl.cnf for new extension format, add comments.
5133 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5134 and add a sample to openssl.cnf so req -x509 now adds appropriate
5138 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5139 error code, add initial support to X509_print() and x509 application.
5142 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5143 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5144 stuff is currently isolated and isn't even compiled yet.
5147 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5148 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5149 Removed the versions check from X509 routines when loading extensions:
5150 this allows certain broken certificates that don't set the version
5151 properly to be processed.
5154 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5155 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5156 can still be regenerated with "make depend".
5159 *) Spelling mistake in C version of CAST-128.
5160 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5162 *) Changes to the error generation code. The perl script err-code.pl
5163 now reads in the old error codes and retains the old numbers, only
5164 adding new ones if necessary. It also only changes the .err files if new
5165 codes are added. The makefiles have been modified to only insert errors
5166 when needed (to avoid needlessly modifying header files). This is done
5167 by only inserting errors if the .err file is newer than the auto generated
5168 C file. To rebuild all the error codes from scratch (the old behaviour)
5169 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5170 or delete all the .err files.
5173 *) CAST-128 was incorrectly implemented for short keys. The C version has
5174 been fixed, but is untested. The assembler versions are also fixed, but
5175 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5176 to regenerate it if needed.
5177 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5178 Hagino <itojun@kame.net>]
5180 *) File was opened incorrectly in randfile.c.
5181 [Ulf Möller <ulf@fitug.de>]
5183 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5184 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5185 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5186 al: it's just almost always a UTCTime. Note this patch adds new error
5187 codes so do a "make errors" if there are problems.
5190 *) Correct Linux 1 recognition in config.
5191 [Ulf Möller <ulf@fitug.de>]
5193 *) Remove pointless MD5 hash when using DSA keys in ca.
5194 [Anonymous <nobody@replay.com>]
5196 *) Generate an error if given an empty string as a cert directory. Also
5197 generate an error if handed NULL (previously returned 0 to indicate an
5198 error, but didn't set one).
5199 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5201 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5204 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5205 parameters. This was causing a warning which killed off the Win32 compile.
5208 *) Remove C++ style comments from crypto/bn/bn_local.h.
5209 [Neil Costigan <neil.costigan@celocom.com>]
5211 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5212 based on a text string, looking up short and long names and finally
5213 "dot" format. The "dot" format stuff didn't work. Added new function
5214 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5215 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5216 OID is not part of the table.
5219 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5220 X509_LOOKUP_by_alias().
5223 *) Sort openssl functions by name.
5226 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5227 encryption from sample DSA keys (in case anyone is interested the password
5231 *) Make _all_ *_free functions accept a NULL pointer.
5232 [Frans Heymans <fheymans@isaserver.be>]
5234 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5236 [Anonymous <nobody@replay.com>]
5238 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5239 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5241 *) Don't blow it for numeric -newkey arguments to apps/req.
5242 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5244 *) Temp key "for export" tests were wrong in s3_srvr.c.
5245 [Anonymous <nobody@replay.com>]
5247 *) Add prototype for temp key callback functions
5248 SSL_CTX_set_tmp_{rsa,dh}_callback().
5251 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5252 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5255 *) X509_name_add_entry() freed the wrong thing after an error.
5256 [Arne Ansper <arne@ats.cyber.ee>]
5258 *) rsa_eay.c would attempt to free a NULL context.
5259 [Arne Ansper <arne@ats.cyber.ee>]
5261 *) BIO_s_socket() had a broken should_retry() on Windoze.
5262 [Arne Ansper <arne@ats.cyber.ee>]
5264 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5265 [Arne Ansper <arne@ats.cyber.ee>]
5267 *) Make sure the already existing X509_STORE->depth variable is initialized
5268 in X509_STORE_new(), but document the fact that this variable is still
5269 unused in the certificate verification process.
5270 [Ralf S. Engelschall]
5272 *) Fix the various library and apps files to free up pkeys obtained from
5273 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5276 *) Fix reference counting in X509_PUBKEY_get(). This makes
5277 demos/maurice/example2.c work, amongst others, probably.
5278 [Steve Henson and Ben Laurie]
5280 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5281 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5282 are no longer created. This way we have a single and consistent command
5283 line interface `openssl <command>', similar to `cvs <command>'.
5284 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5286 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5287 BIT STRING wrapper always have zero unused bits.
5290 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5293 *) Make the top-level INSTALL documentation easier to understand.
5296 *) Makefiles updated to exit if an error occurs in a sub-directory
5297 make (including if user presses ^C) [Paul Sutton]
5299 *) Make Montgomery context stuff explicit in RSA data structure.
5302 *) Fix build order of pem and err to allow for generated pem.h.
5305 *) Fix renumbering bug in X509_NAME_delete_entry().
5308 *) Enhanced the err-ins.pl script so it makes the error library number
5309 global and can add a library name. This is needed for external ASN1 and
5310 other error libraries.
5313 *) Fixed sk_insert which never worked properly.
5316 *) Fix ASN1 macros so they can handle indefinite length construted
5317 EXPLICIT tags. Some non standard certificates use these: they can now
5321 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5322 into a single doc/ssleay.txt bundle. This way the information is still
5323 preserved but no longer messes up this directory. Now it's new room for
5324 the new set of documenation files.
5325 [Ralf S. Engelschall]
5327 *) SETs were incorrectly DER encoded. This was a major pain, because they
5328 shared code with SEQUENCEs, which aren't coded the same. This means that
5329 almost everything to do with SETs or SEQUENCEs has either changed name or
5330 number of arguments.
5331 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5333 *) Fix test data to work with the above.
5336 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5337 was already fixed by Eric for 0.9.1 it seems.
5338 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5340 *) Autodetect FreeBSD3.
5343 *) Fix various bugs in Configure. This affects the following platforms:
5347 unixware-2.0-pentium
5351 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5352 before they are needed.
5355 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5359 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5361 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5362 changed SSLeay to OpenSSL in version strings.
5363 [Ralf S. Engelschall]
5365 *) Some fixups to the top-level documents.
5368 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5369 because the symlink to include/ was missing.
5370 [Ralf S. Engelschall]
5372 *) Incorporated the popular no-RSA/DSA-only patches
5373 which allow to compile a RSA-free SSLeay.
5374 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5376 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5377 when "ssleay" is still not found.
5378 [Ralf S. Engelschall]
5380 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5381 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5383 *) Updated the README file.
5384 [Ralf S. Engelschall]
5386 *) Added various .cvsignore files in the CVS repository subdirs
5387 to make a "cvs update" really silent.
5388 [Ralf S. Engelschall]
5390 *) Recompiled the error-definition header files and added
5391 missing symbols to the Win32 linker tables.
5392 [Ralf S. Engelschall]
5394 *) Cleaned up the top-level documents;
5395 o new files: CHANGES and LICENSE
5396 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5397 o merged COPYRIGHT into LICENSE
5398 o removed obsolete TODO file
5399 o renamed MICROSOFT to INSTALL.W32
5400 [Ralf S. Engelschall]
5402 *) Removed dummy files from the 0.9.1b source tree:
5403 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5404 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5405 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5406 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5407 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5408 [Ralf S. Engelschall]
5410 *) Added various platform portability fixes.
5413 *) The Genesis of the OpenSSL rpject:
5414 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5415 Young and Tim J. Hudson created while they were working for C2Net until
5417 [The OpenSSL Project]
5420 Changes between 0.9.0b and 0.9.1b [not released]
5422 *) Updated a few CA certificates under certs/
5425 *) Changed some BIGNUM api stuff.
5428 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5429 DGUX x86, Linux Alpha, etc.
5432 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5433 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5437 *) Add -strparse option to asn1pars program which parses nested
5439 [Dr Stephen Henson <shenson@bigfoot.com>]
5441 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5444 *) DSA fix for "ca" program.
5447 *) Added "-genkey" option to "dsaparam" program.
5450 *) Added RIPE MD160 (rmd160) message digest.
5453 *) Added -a (all) option to "ssleay version" command.
5456 *) Added PLATFORM define which is the id given to Configure.
5459 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5462 *) Extended the ASN.1 parser routines.
5465 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5468 *) Added a BN_CTX to the BN library.
5471 *) Fixed the weak key values in DES library
5474 *) Changed API in EVP library for cipher aliases.
5477 *) Added support for RC2/64bit cipher.
5480 *) Converted the lhash library to the crypto/mem.c functions.
5483 *) Added more recognized ASN.1 object ids.
5486 *) Added more RSA padding checks for SSL/TLS.
5489 *) Added BIO proxy/filter functionality.
5492 *) Added extra_certs to SSL_CTX which can be used
5493 send extra CA certificates to the client in the CA cert chain sending
5494 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5497 *) Now Fortezza is denied in the authentication phase because
5498 this is key exchange mechanism is not supported by SSLeay at all.
5501 *) Additional PKCS1 checks.
5504 *) Support the string "TLSv1" for all TLS v1 ciphers.
5507 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5508 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5511 *) Fixed a few memory leaks.
5514 *) Fixed various code and comment typos.
5517 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5518 bytes sent in the client random.
5519 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob